Marketing Metrics

Decision Tree Extension

See slide 127-128. In this extension on the previous example, imagine that confidence was already fairly high and the change in the marketing mix would lead to only a small increase in the probability of the best outcome.

Growth

See slide 131

Landing Page Experiment

See slide 133-135

Computing Customer Equity: A Simple Approach

Customer equity is the sum of customer lifetime values for all current and expected customers we currently have and all the customers so to compute customer equity we could make the simple assumption that no customers defect. In other words, we assume the retention rate is 100%, and thus the total number of customers remains constant. This slide shows a simple way to compute customer equity under these assumptions. If a company wanted to assume a constant growth rate, like 5%, instead of 100% retention, they would use 105% for retention in the CE model. See slide 130.

Deciles

Each of ten equal groups into which a population can be divided according to the distribution of values of a particular variable. ex. "the lowest income decile of the population" Each of the nine values of the random variable that divide a population into ten such groups.

EBITDA

Earnings Before Interest, Taxes, Depreciation, and Amortization This is important because it shows you how your operations are doing, how well your business is doing

Intangibles

Goodwill Trade names

Goodwill

Goodwill arises when one company buys another company and pays more than the book value. Goodwill reflects the synergy of a company or the value of brands, talent, or strategic positioning of a company that goes beyond the assets listed on the balance sheet.

Inventory Turns

Inventory Turns = Annual Product Revenues ($)/ Average Inventory ($) See example on slide 128-129

How to compute the Return on Investment (ROI) of a business investment

ROI (%) = Net Profit ($) / Investment ($) ROI (%) = (Gross Margin - Marketing Investment) /Marketing Investment See example on Slides 92 + 93, 94 + 95!

Customer Equity

The total of the discounted lifetime values summed over of all of the firm's current and potential customers.

Trade names

Trade names can refer to specific brands. This category would arise on a balance sheet when a company purchases the rights to use a particular brand. For example, when Hostess Company went bankrupt, it sold the rights to another company to use the name Twinkies. That other company might show the value of the price they paid for the name Twinkies as an asset on their balance sheet.

"Transpromo"

Transactional printing combined with promotional marketing A transpromo is one example of customization. In this marketing tactic, the company sends out promotions with other transaction related materials, such as a monthly bill. Because the company knows a lot about the customer, as indicated by the detail on the bill, the company is able to insert offers that are tailored to the customers' needs.

Trial, Repeat

Trial rate (%) = (First-time triers per period)/total population Repeat Rate (%) = percentage of population who bought last time that will buy again this time. The rate of trial and the rate of repeat are two important metrics used to forecast the sales of a new product.

Satisfaction and Loyalty

Under what conditions will satisfaction not be related to loyalty? See Saporito, "Staying Power"

Integrated Customer Identification

Using common customer identifiers at every customer contact point, and accessing a common database, to capture maximum customer information and integrate maximum customer knowledge in every interaction.

Estimating Demand Curves

van Westendorp Price Sensitivity Meter Concept Test with Likelihood Scales Conjoint Analysis Discrete Choice Modeling

Amortization

is the "depreciation" of an intangible asset (e.g., a patent, a trade name).

From Conjoint to Market Share

"Conjoint simulators" use conjoint or DCM methods to estimate importance weights and then use First-Choice or Share of Preference models to estimate likely market share of hypothetical new products. In more advanced conjoint software, such as the Sawtooth software, modules are included that allow the user to estimate the market share for a new product as part of the conjoint analysis. Market share is estimated using one of several rules, including the first choice rule or the share of preference rule.

What do we Learn from Balance Sheets?

-The net worth of the business. -The existence of any important assets (e.g., cash). -The risk exposure or financial health of the business (debt/equity ratio). -Balance sheets provide several types of useful information, including the book value of the business, the existence of important assets, and the risk exposure or financial health of the business. For example, the balance sheet might reveal that a company owns several valuable pieces of land, or several valuable buildings. It might also show that the business has large loans taken out against those buildings and that land. Or a business with similar assets might not have as much debt, indicating that the company could be purchased and then loans could probably be acquired against those assets in order to generate cash.

Metric

-a measuring system that quantifies a trend, dynamic, or characteristic -descriptive

Deductive vs. InductiveMarket Research

B2C: Inductive Approach: Think "statistical inference." Collect large samples, infer population averages. B2B: Deductive Approach: Think "Sherlock Holmes." A few insightful observations are connected via a model to form a conclusion. Uses reasoning.

PURL

Personalized URL ex. slide 158, 159, 160

Customer Differentiation at Fidelity Investments

Fidelity cut off phone privileges for 30,000 investors (0.5% of customers). Avg. phone call costs $13, average automated call costs $1. Some BZ's call every hour. Only 34 BZ's have complained. See slide 147.

Accumulated Depreciation

Generally, assets are listed on the balance sheet based on the amount that was actually paid for the asset. For example, if you paid $50,000 for a truck, under equipment we would see a truck listed at $50,000. However, as the truck ages, it begins to wear out and its value may be less that $50,000. The accumulated depreciation for that truck would show how much value we should subtract from that truck to get a more meaningful estimate of the current value of that asset.

Possible Biases in NPS due to Unique Needs or Familiarity

How likely are you to recommend Keystone to a friend? How likely are you to recommend Keystone to a friend who likes to ski? Bacon's hypothesis: Providers of unique benefits will see lower NPS after controlling for satisfaction. How likely are you to recommend McDonalds? Bacon's hypothesis: Providers of familiar products will see lower NPS after controlling for satisfaction. Consider: How likely... to a ... who asks for your advice After using the NPS in several consulting experiences, I have come to wonder if it should be interpreted differently in different industries. For example, if a company provides fairly unique benefits, customers may be less likely to recommend the company because there are so few friends and colleagues who would be interested in those unique benefits. In addition, if a company offers a product or service that is already familiar to most consumers, customers may be less likely to recommend the company because most friends and colleagues will already be familiar with the company and have already formed an opinion. In either of the situations mentioned on this slide, the NPS may be low even though the product or service quality is very good.

Pricing B2B Innovations

Opportunity Analysis (Bacon & Butler) Induction v. Deduction The four questions Implications for Pricing In this next set of slides, I describe a method that my father, Dr. Frank Bacon, and his consulting partner Dr. Tom Butler, used in pricing decisions for business to business and technology oriented projects. As we will see, the model uses a customer-oriented approach to determining price.

MGC's

Most Growable Customers Goal: Customer Growth (Stars)

MVC's

Most Valuable Customers Goal: Customer Retention (Cash Cows)

Attitudes about TV Ads

Of marketers surveyed, 86% feel that measuring the effectiveness of their media strategies was the biggest challenge they faced last year. 62% believe TV ads are less effective than they were 2 years ago. 65% believe Internet usage measurement is more useful than TV viewing measurement. The three bullets on this slide indicate that marketers have some frustration with measuring the effectiveness of traditional media and wish it were as easy as measuring the effectiveness of Internet marketing.

Complex CLV Models

Need to know how to lay it out in excel!!! See slide 122!

Expectancy Confirmation

Non-satisfaction = When expectations are quite low, and expectations are met (confirmation). One interesting variation in the expectancy disconfirmation paradigm is when we have very low expectations and those expectations are met. In the previous line, we might have thought this situation would lead to satisfaction, but because the expectations are so low, a different term is sometimes used in the situation: non-satisfaction. In other words, when you expect bad service and you get bad service you were not necessarily dissatisfied but neither are you satisfied; you have some state of non-satisfaction.

Economies of Scale Examples

To what degree do the following industries have economies of scale? Why? Hair stylists Beer brewers Automobile manufacturers Grocers (e.g., King Soopers) Encyclopedia online Cell phone network services Churches Some of the industries listed on this slide have fairly low economies of scale and some have very high economies of scale. For each of these, I would encourage you to think of the bullets we have discussed on earlier slides reflecting the factors that lead to economies of scale. Think about how those factors might affect these industries shown on the slide. For example, if a hairstylist ramps up their volume considerably, would they change to a new production technology? Probably not. They may achieve some labor efficiencies, and there may be some wider allocation of fix cost related to management or advertising, but otherwise the fundamental production process does not change much with volume, and therefore the economies of scale for hairstyling would not be large. This explains why the industry is still very fragmented and we see many small hairstylist shops in business today. The few large chains are probably capitalizing on economies related to advertising, branding, and management efficiencies. b. Medium c. Large d. (bargaining power,

Dashboards

A relatively small collection of interconnected key performance metrics and underlying performance drivers that reflects both short and long-term interests to be viewed in common throughout the organization. In marketing at least, dashboards tend to show metrics in graphical form, showing "a reduced set of the vital measures in a form that is easy for the operator to interpret and use." (Farris, 2006, p. 33)1 See slide 161

Payback with ongoing investment

payback period= initial investment / (contribution * (units/period) - ongoing investment/period) In many marketing situations, there is an initial investment for the creative for a project, such as making a television ad or a print ad. There are also additional investments in paying for the media, such as TV airtime and the number of inserts in various vehicles or print media. This equation helps model these decisions. Example on slide 89!

Current assets

Cash Cash equivalents Short term investments Accounts receivable Inventory Work in process Supplies Prepaids *Current assets are used up or could be converted to cash within one year. Generally, current assets can easily be converted to cash within one year, although this one year time limit period is only approximate and there are exceptions. Most of the items listed under current assets are fairly straightforward.

How Long a Life?

Considerations include: 1. How long are customers in the market? -Dental services v. obstetrics (baby deliverers). 2. How long are customers retained? -Short-life customers will have low retention rates at some point. -At what point are 95% of the customers gone? 3. How high is the discount rate? -Less than 20% or so may mean a lot of value comes from later years. -High retention (>80%) and low discount (<20%) means later years can add value. 4) Length of marketing plans -CLV should be at least as long as retention expenditures for customers. -Will business change radically in the next few years, leading to loss of most customers as we now know them? (Netflix?) -Can customers be sold with the business? 5) High risk in later years may warrant higher discount rates, and/or a shorter life -Netflix? 6) Politics: Make your life as long as possible to maximize CLV! -ROI on acquisition spending will look better. In the simple customer lifetime value equation, the implicit assumption is made that a customer may have an infinite life. While we might all want to live forever, that's probably not a realistic assumption. In the more complex model of customer lifetime value that we are about to talk about, you will need to make explicit assumptions about the length of the customer lifetime. This slide and the next show some issues to consider when you decide on the length to use in your calculations. In some industries, customers are in the market for a long time, such as dentistry. Most people in the United States will use dental services for their entire life. Obstetrics is different. Women who seek the help of a doctor who delivers babies may only need the services of that doctor for a very short time in their lives. Therefore, when estimating the customer lifetime value of a dental customer, we would use a very long lifetime, but when estimating the lifetime value of an obstetrics customer, we would use a much shorter lifetime. Customer retention rate is also an issue. If the customer retention rate is low, most of the customers will be gone in a short period of time. This means you will only need to use a few years to get a good estimate of the customer lifetime value. The rate you use to discount those cash flows can also affect how many years you should use in your lifetime value calculations. For the discount rate we typically use, a fairly long lifetime would be reasonable. However, if the discount rate were quite high, say well over 20%, then cash flows far in the future would be discounted so much that they wouldn't be worth much in today's dollars. Under those circumstances, it wouldn't make much sense to model customer lifetime value with such a long lifetime. The length of your marketing plans might also affect the time you use to compute customer lifetime value. If you plan on spending retention dollars on customers for at least ten years after their first purchase, then your model of CLV should probably go at least ten years. If there is a high level of risk in your business, such that your business might be gone in a few years, then you might want to use a higher discount rate and a shorter lifetime in your CLV calculations. Office politics may also play a role. In marketing, we often try to get bigger budgets. Generally, if you sum cash flows over a longer period of time you will generate a higher estimate of CLV. More valuable customers, in turn, could be used to justify more marketing spending to get those customers. To fatten your budget, you might want to include more years than you might have otherwise.

How to compute the contribution margin

Contribution per unit ($) = Selling Price ($) - Variable cost per unit ($) Contribution Margin (%) = (Selling Price - Variable Cost ($)) / Selling price per unit ($) Contribution = "contribution to fixed costs" or "contribution to overhead" In computing gross profit margin, some fixed costs may be included in costs of goods sold, so gross profit margin may not equal contribution margin.

Objective and Task and ROI

Develop a model of how advertising affects sales. Identify key objectives to produce sales. Changes in beliefs, affect, behavior Determine costs for each task. Estimate ROI for each task. Complete tasks with acceptable ROI At a minimum, examine how well tasks were achieved. The first bullet on this slide suggests we start with a model of how advertising affects sales. Response hierarchy models are commonly used to describe how advertising affects sales. On the next slide we show some classic response hierarchies.

Keys to Online Advertising Success:

Great Creative Great Placement Attention to the Numbers Testing, Testing, Testing

Diminishing marginal returns

Diminishing marginal returns = for each additional (marginal) dollar spent, the ROI gets smaller and smaller. See slide 111. This slide shows the graphic of a typical marketing response curve. Most marketing expenditures will see diminishing marginal returns the more you spend. In other words, you get a very good bang for your buck for the first expenditures, but as you spend more and more, the bang for your buck gets smaller and smaller. In this example, notice that if you spend $1 million, the ROI is 200%, but if you spend $5 million, the ROI is only 6%. If you made decisions only based on ROI you might find yourself making many small investments, and that may not be the best way to invest your marketing money.

Discrete Choice Modeling

Discrete choice modeling (DCM) ask consumers to make choices among hypothetical objects (like full profile conjoint descriptions) DCM can be more mathematically complex than conjoint, but some feel may give slightly more accurate results. I have glossed over some of the details on how to estimate coefficients (importance weights) with discrete choice modeling. The mathematics can actually be much more complex and are beyond the scope of this class. Still, the intuitive explanation I have offered should help you to understand what you need to know about discrete choice modeling.

Online Testing at P&G

Crest Whitestrips sold only online for 8 months Test price point of $44 Sale promoted in selected magazines Sold well - used online sales figures to help sell through to retailers P&G believes in the Internet for sampling and testing, but not much for advertising. One method of conducting pricing research is to roll out a product online first. In this particular example, Crest was concerned that the price point would be too high to sell in a grocery store and the grocery retailers would not stock the product. P&G's online roll out was very successful and this allowed them to convince retailers to carry the product in their stores.

What goes on a balance sheet?

Current Assets Fixed Assets Current Liabilities Long-term Liabilities Net Worth

Types of Liabilities

Current Liabilities (due within a year) Long-term Liabilities

Types of Assets

Current assets Fixed assets (non-current) Intangibles

*How to compute debt/equity ratio

Debt/Equity Ratio = (Total Liabilities)/(Owners Equity)

Value Skew

High value skew = High percentage of revenue from small percentage of customers. Low value skew = all customers have about the same profitability. You have probably heard of the 80-20 rule. In some forms, it is said that 80% of our profits come from 20% of our customers. This is an example of value skew. If half of our profits came from half of our customers, meaning each customer were equally valuable, we would have no value skew.

Metrics for Ad Vehicles

If your objective is to sell ad space, track -Cost per visit (subscribers and non-subscribers) -Page views per visit -Visits per week -Repeat visit rate -Also note demographics by page Depending on the objectives of your website, you should choose different metrics to track the performance. This slide shows metrics that would be relevant if your objective was to sell ad space on your website. In this case, we would want lots of visitors, and we would want them to spend a long time on the website and to come back often. It would also be nice to know the demographics of the visitors to help advertisers better select the best vehicles for ad placement.

Selecting Projects Based on ROI

If you could estimate the ROI of several different projects, you could then select the projects that pass a certain rate. You would also need to consider your entire marketing budget and the size of the projects to determine how many of the projects you could fund. ROI may be only somewhat correlated highly with project size. NPV is more dependent on project size. See slide 96!

Estimating Retention Rates

If you have a good database: -Cohort and Incubate (Farris et al.) -Go back into your database and track a cohort of customers from 5 or 10 years ago. -Typically aggregate data ("avg customer") -Apply Survival Analysis -Examine start dates and end (defect) dates of individual customers. -Model likelihood of defecting for individual customers or groups of customers. There are several methods that can be used to estimate customer retention rates. If you have a good database, you can look into your database and examine individuals who became customers some years ago and then track what happened to them over time. This method, called cohort and incubate, is what was used in the Tuscan lifestyles case. With a very good database, you might be able to apply survival analysis. For survival analysis to work, you must know exactly when customers left the company. This would have been difficult in the Tuscan Lifestyles case because if a customer stopped purchasing for a while, we can't be sure that they will never purchase again. For organizations that have a member database, they may be able to see when someone stopped paying membership dues, and this would be a clear indication that they had left the organization. In a detailed survival analysis, it is possible to identify specific events or changes in customer lifestyle that may increase the probability him or her defecting to a competitor.

Metrics for Customer Service Sites

If your objective is to offer cost-effective customer service, track Reduction in use of call center Reduction in use of human e-mail response Visits to customer service pages (e.g., FAQ's) Customer satisfaction with service We discussed metrics for advertising sites before. If we now consider websites where the objective is to offer customer service, rather than tracking some kind of increased revenue, we might instead look to track reductions in costs in other ways of serving customers. The savings from the reduction in costs of these other customer service methods can be used to compute an ROI of the customer service website components.

Metrics for Direct Sales Sites

If your objective is to sell products or services, track Cost per acquisition Visits per week Conversion rates Avg revenue per sale Customer retention rates Here are the metrics that we have discussed before that are commonly used to track the performance of a website that is focused on selling products or services.

Retention Rates from Survey Data

Implement a customer survey. On your survey, ask: "In what year did you first become a customer of our company? ______" Divide #customers in year n by number of customers in year n+1 to estimate the retention rate for that time period. Step 1: Smooth the Data (Slide 137) Step 2: Use Ratios to Estimate Retention Curve (Slide 138) By adding this question to a customer survey (e.g., your annual customer satisfaction survey), it is possible to estimate customer retention rates. If customer retention rates are very low, we would expect to see very few survey respondents who have been our customer for over five years; if customer retention rates are very high, we might see survey respondents who have been our customers for 20 or 30 years. This intuitive observation is the logic behind the next few slides. The exact math behind the calculations to follow is beyond what I would expect students to master for this course. Still, I wanted to include these for possible future reference. I am actively engaged in research and refining models like these. If you are ever involved with a company that wants to estimate retention rates in this way, please give me a call.

Advertising Metrics: Impressions

Impressions (exposures, opportunities to see [OTS]) Number of opportunities for people to see or hear the ad if they were paying attention. They don't have to be paying attention to count as an impression. We discussed impressions before, and I include the slide again for completeness. In some marketing activities, impressions are measured with a multiplier. For example, in magazine advertising, the publisher may know what the subscription rate is, and may also estimate the pass-around rate. The publisher could argue that the number of impressions delivered is actually larger than the subscription base because the subscribers show the magazine to one or more of their friends.

Incremental Sales: Multiplicative Model

Incremental Sales = sales multiplier from advertising x sales multiplier from trade promotion x sales multiplier from consumer promotion x sales multiplier from other promotions Which makes more sense - additive model or multiplicative model? Multiplicative terms are like interactions in regression.

Incremental Sales: Additive Model

Incremental Sales = added sales from advertising + added sales from trade promotion (discounts and other deals to retailer) + added sales from consumer promotion (coupons, in-store displays, sweepstakes) + added sales from other promotions (special marketing events?) In considering the impact of any promotion, we could think of that impact as being additive, that is, the effects of each promotion add together to drive new sales. Alternatively, you could think of each promotional effort as having a multiplicative effect (see you next slide), as if every promotion interacted with every other promotion to drive sales. In practice, we sometimes estimate the effect of promotions both ways, and see which model fits our data better. Some promotions have additive effects, and other promotions have multiplicative effects.

Categorizing Customers by Value

MVC's: Most Valuable Customers MGC's: Most Growable Customers BZ's: Below Zero Peppers and Rogers created these somewhat provocative categories for customers as we differentiate based on the value they provide to our organization. Most controversy among these categories come from those in the below zero, or BZ category. These are customers that we never expect to make money on. Once identified, we need to think about strategies regarding how we can change our marketing plans to make these customers profitable, or how we can get these customers to switch to one of our competitors. While it may seem sad to lose customers, or fire customers, keep in mind we are losing money on these customers. We would be more profitable if they would defect and lose money for our competitors instead. In Peppers and Rogers' original model, they consider other benefits beyond direct margins from customers. For example, they consider the value of referrals or the value of positive word-of-mouth. In Amazon's model, they might consider the value of the customer reviews that Amazon customers might write and post. Amazon might be willing to lose money on these customers if they felt the added value of the reviews the customers post made up for the loss in margin.

Setting Advertising Budgets

Percentage of sales E.g., ad budget is 3% of last year's gross sales. Caution: sales will cause advertising. Competitive parity Match competitors, or match relative to share of market. All you can afford Use whatever's left after paying the last bill! Objective and task (considered best by many) Figure out what needs to be done, identify costs for each task. Knowing how ad budgets are usually set is a good starting point to a discussion of whether we are spending the right amount or should be spending more on advertising. There are four commonly used methods for setting the advertising budget. The objective and task approach is probably the most theoretically sound, and so we will spend some time exploring ways to do this.

Prepaids

Prepaid refers to amounts that we have paid in advance. For example, a business might pay its rent for six months at one time. When doing so, it is as if the company owns the rights to use the building for six months, and so in a sense it owns something of value. ex. if you paid 6 months of rent in advance you own that without owing anything on it for 6 months

More Lessons...

Present data graphically to facilitate discussion. If you hand them 20 pages of tables, you've lost interest and may lose control of the meeting. With a handful of graphics, attention will quickly focus on the biggest fire to put out. A good way to prioritize! Graphics help immensely in presenting data. If you present loads of tables instead, people easily get lost, and they get distracted from the story that you want to be able to tell.

RFM

Recency, Frequency, Monetary Value These three variables are often associated with response rates in direct marketing (database marketing). Usage: -Score each customer based on RFM values, from highest to lowest (consider regression to develop your model) -Sort customers from highest to lowest -Contact customers in the deciles with high enough response rates to accomplish financial goals Example on Slide 116

Possible Simulated Test Market Procedures

Recruit participants (mall intercept or online panel). Self-administered questionnaire. Advertising stimuli. Mini-store shopping experience. Post-exposure questionnaire. Receive trial package (fulfillment study). Phone follow-up and offer to buy more; assess repeat and repeat frequency. Simulated test markets are often used to estimate some parameters, especially trial and repeat, in ATAR models. Simulated test markets are not full-blown test markets, rather they are somewhat artificial and more like survey research.

ROI of Product Line Expansion

See example on slide 59-61. As mentioned on the previous slide, it would be wise to estimate the ROI of any expansion of your product line before rolling out new products. On this page, an example is shown where we can calculate the ROI of this opportunity. In these product line expansion problems, an important first step is to compute the change in profit due to the new product introduction. That is, rather than compute the entire net present value with the old product assortment, and then compute the entire net present value with the new product assortment, we can start by computing the difference in profit between the two each year and then just plugging this difference into the worksheet to follow. This page shows the net present value and the ROI of the new product expansion decision. Note that the $800,000 per period is plugged in here as the change in profit comparing the old product line with the new product line.

Hierarchy of Effects ROI

See example on slide 73-74

How to compute the net present value (NPV) of a series of cash flows

See example on slides 99 + 101! NPV = the present (discounted) value of future cash inflows minus the present value of the investment and any associated future cash outflows. Net Present Value Principle: A dollar today is worth more than a dollar a year from now. Present (discounted) value = nominal (face) value * discount factor. = nominal $ x (1/ (1+i)^p) i = discount rate, p = period (e.g., year) On this slide I introduce the concept of net present value. This concept allows us to take into account that a dollar today is worth more than a dollar a year from now, or ten years from now. When we analyze short-term projects, such as a project that lasts less than one year, net present value will not be a very important concept. However, when we analyze marketing activities that last more than one year, net present value can make the calculations more meaningful. For example, running promotions for a specific marketing event may last less than one year and so would not require net present value calculations. However, computing customer lifetime value generally involves estimates of cash flows going many years into the future, and therefore net present value is commonly used in those calculations. Conceptually, we can see that $100 today would probably be more valuable to you that $100 ten years from now. After all, if you had $100 today, you could invest that money and you would probably have more than $100 ten years from now. For example, if you could invest that money at a 7% annual return, you could approximately double your money in ten years, giving you $200 in ten years from the $100 you had right now. Therefore, $200 ten years in the future is about the same as $100 right now. To calculate the conversion from future dollars to present dollars, we use a discount factor. The equation for the discount factor is shown on this slide and is equal to one over (one plus the discount rate raised to the number of periods). In most of our calculations, the number of periods will be expressed as the number of years. Also, in most of our calculations, I will use a discount rate of 10%, but I caution you that 10% may not be the best discount rate to use in your job, and on the next slide I will talk about how you could get a more meaningful estimate of the discount rate.

"Adjective Checklist" Shows Associations

See slide 100-101

Political Conjoint

See slide 101-103. The next few slides show how conjoint analysis could be applied to things other than products. In this example, conjoint is used to survey voters about what issues are important to them in a state election. Full profile conjoint is used, and the results are broken out by different demographics to gain deeper understanding into what is important to voters by segment.

Strength and Favorability of Association

See slide 102

Using Conjoint to Capture Brand Equity

See slide 104

Customer Lifetime Value at King Soopers

The average family of four spends about $8,000/yr on groceries. King's analysis: Assume a customer stays with us 10 years. The average customer is therefore worth $80,000. What's wrong with this analysis? The president of King Soopers visited my class many years ago and gave a presentation where he described this somewhat inaccurate approach to estimating customer lifetime value. He used this estimate in training employees to encourage them not to be shy about giving away free grocery products if any customer complained about something being spoiled or damaged. For example, if a customer complained about mold on a loaf of bread, an employee should not hesitate to offer the customer a free loaf, or maybe even two, in exchange. His thought was that the cost of a couple loaves of bread is far less than the customer lifetime value that might be lost if the customer switched grocery stores. He may have used a simple model here to make it easy to explain to all employees.

Book Value

The book value of a firm is the value of its tangible assets (assets you can touch, taste, hold) minus its liabilities. What is the book value of the company shown in the simple balance sheet earlier? What does it mean if a company's market capitalization* is greater than book value? What if mkt cap is less than book value? If you had $100,000 in accounts receivable, but have not been paid yet you can sell that to another company for less, maybe $75,000 and then all your accounts receivable get paid to that company instead and they can make money off of it and you get cash for those liabilities. Book Value = Retained Earnings - Intangible Assets

Financial Statements

There are three commonly used financial statements: balance sheets, income statements, and statements of cash flows. Of these, the most important statement to marketers is probably the income statement. I think the balance sheet is important to understand, and may be of use to students personally, so we will also spend some time going over balance sheets. The statement of cash flows is used primarily by whoever in the organization is responsible for managing cash, which is usually not a marketing person.

Example of Deductive Hypotheses

There is a need for improved automatic flight control of all types of commercial aircraft to permit all-weather "hands-off" landing at all major airports BECAUSE: 1) The technology exists to do it 2) The costs due to delays, wasted fuel, and lost revenues would easily pay for it 3) Most major airlines would like to retrofit their aircraft in the near future if such a system were available This example comes from research my father did over two decades ago on the use of autopilot in commercial aircraft. His client at the time was producing what we might call autopilot for military aircraft, but wanted to expand their market. These hypotheses were developed to help the company focus its research efforts and determine if there was a viable market opening for this product in this new market. Notice that a different expert might address each of these hypotheses. The first hypothesis might be addressed by talking to someone within the company who was a technical expert on the technology. The second question might be addressed by talking to a small handful of people who work at airlines and asking them about how much money they felt they currently waste in delays and waste of fuel by not being able to land aircraft in difficult weather. The third hypothesis might be addressed by talking to chief mechanics at the airlines (or facilities managers) that had deep knowledge of practices and policies around aircraft upgrades. Interestingly, in this particular example, this third bullet turned out to be the biggest challenge to the product rollout.

Relevant Costs in Decision Analysis

Decision analyses should include costs that are affected/changed by the decision (= relevant costs). "Sunk costs" are monies that have already been spent, and are therefore generally irrelevant to decision analyses. Manufacturing example: previous R&D $. Advertising example: previous ad creation $. In analyzing the attractiveness of marketing actions, it is important to distinguish between relevant costs and irrelevant costs. Costs that could be changed based on the decision we make are relevant costs. Costs that would not change based on our decisions are irrelevant costs. These costs are often sunk costs, meaning these are monies we have already spent or have already committed to spend, regardless of what decisions we make going forward.

Price Discrimination and the Law

Robinson-Patman Act: Price discrimination is illegal if it threatens competition. Primary line injury (predatory pricing): pricing low to squeeze out a competitor at the supplier level. Secondary line competitive injury: charging competing buyers different prices in order to drive one out of business. Price discrimination laws generally apply at the B2B level, not B2C. The Robinson-Patman Act makes some forms of price discrimination illegal. Note that these often relate to competing buyers, which means this usually relates to business-to-business marketing and not business to consumer marketing.

Chordiant's Recommendation Advisor

Uses current and historical customer data to estimate CLV for customers. When a customer calls in asking for product or service changes (e.g., cell phones), call center staff see a screen that shows how deep a deal can be offered to that customer with an acceptable ROI. Staffer's commission structure integrated with ROI as well. This example shows how innovative technology can be used to better manage customer relationships. Customer service agents in a call center received detailed information on each caller as they talked to that caller on the phone. In the background, Chordiant's software would run and display information on the screen in front of the employee describing deals the employee could make with the caller and still give the company a good ROI. If the employee could strike a deal with the customer that made better money for the company, the employee received a bonus related to the extra margin that the he or she was able to negotiate.

Problems with Dichotomizing Measures

Using the percentage in the top two boxes ("top 2 box score") discards information that would be reflected in the mean. Means are more reliable than top 2 box scores. Other things being equal, tracking studies using means will required smaller sample sizes than tracking studies using top 2 box. Top 2 box is still very popular in industry. Another problem with the NPS is that it dichotomizes what might otherwise be treated as a continuous measure. It has become very popular in marketing report scores like the "top two box score". For example, on a five-point Likert question, instead of reporting the average on the question, a research report might only report the percentage of respondents who answered with a four or a five. Because this takes a five-point scale and converts it essentially into a two-point scale (top two boxes versus bottom three boxes), we can say that the scale has been dichotomized. When ever we dichotomize a scale, we throw away information, and the resulting scores will generally be less reliable than if we had used the average on the full scale. The practice of reporting top two box is very popular in marketing, but I caution you that it is not the most accurate way to report survey data.

Liabilities

What you owe to others.

Decision calculus

-A set of numerical procedures for processing data and judgments to assist managerial decision making. -Decision calculus involves using metrics and mathematical models to make better decisions. -Decision calculus uses some management judgment and some mathematical modeling to provide solutions that are better than management judgment alone. -See slide 29 for example.

Income Statements (P&L Statement)

-A.k.a., Profit and Loss (P&L) statement, or Statement of Operations. -Income statements describe the financial activity that took place during a specific time period (e.g., a month, a quarter, a year). -The Income Statement tells whether or not the firm is making a profit. -Income statements are also referred to as profit and loss statements or statements of operations. Rather than looking at the business condition at a single point in time, income statements reflect all of the financial activities that took place in a company over a segment of time. It is not possible to determine from a balance sheet whether a business is profitable or not, the income statement is the statement we need to look at to determine if a business is currently making money from its operations. See example on Slide 53!

Balance Sheets

-shows an organization's financial status at one point in time -The balance sheet shows an organization's financial condition at a single moment in time. For example, the statement might show the organization's condition at 4:00pm on a Monday afternoon. The financial condition could be different the following Monday at 4:00pm. -A key aspect of the balance sheet is that the two sides of the sheet MUST balance. That is, the assets must equal the liabilities plus the net worth. As organizations conduct transactions, these transactions must be accounted for in a way so that the balance sheet always remains balanced. Assets | Liabilities + Net Worth See example on slide 38! This balance sheet is sorted by liquidity... the ones at the top are easier to turn into cash! THE MOST LIQUID ITEMS SHOULD BE AT THE TOP! *Retained earnings is the same as owners equity for the sake of this class Example on slides 45 + 46! Go practice! Hint: You are solving for the owners equity (retained earnings) by determining the difference between the assets and liabilities. Practice on slides 48 + 49!

Current Liabilities (due within a year)

Accounts Payable Notes Payable Wages Payable Unearned Revenues

Integrated Marketing Communications (IMC)

Achieving clear, consistent communication with maximum impact via a seamless combination of a variety of methods (e.g., public relations, broadcast, call center, online, etc.).

Price Premium

(Brand A Price ($) - Benchmark Price ($)) Price Premium (%) = Benchmark Price ($) Price Premium is computed using actual prices in the market place. Theoretical Price Premium is the price premium determined through conjoint analysis. Price Premiums are likely to exist under which forms of competition? When a brand has a price premium, that brand generally sells for a percentage higher than a typical product in the category. One of the benefits of a strong brand is that you might be able to hold a price premium and thus increase your margins. Price premiums are not possible in purely competitive markets because there is too little differentiation to allow one competitor to ask a higher price and still sell substantial volumes of product.

Multi-period ROI (MP ROI)

-Compute the NPV of all gross margins. -Compute the NPV of all marketing investments. -Compute ROI using NPV's. MP ROI = ((NPV * Gross Margin) - (NPV * Marketing Investment)) / (NPV * Marketing Investment) When we looked at ROI before, we were looking at projects that lasted less that one year. We can compute ROI for projects that last longer than one year by taking the net present value of the profits and the investment and then plugging those discounted values into the ROI equation. The multi-period ROI solution to the problem we used on some previous slides is described on the next slide. In practice, when determining which ROI equation to use, recognize that if the project is short-term (less than one year) don't worry about NPV. If the project lasts several years, do use the NPV adjustments.

Depreciation

-Depreciation is a term used in accounting to spread the cost of an asset over the span of several years. -In simple words we can say that depreciation is the reduction in the value of an asset due to usage, passage of time, wear and tear, technological outdating or obsolescence, depletion, inadequacy, rot, rust, decay or other such factors. -The use of depreciation in accounting helps to match costs to their relevant time period. -Depreciation is the accounting concept that we use to spread the cost of an asset over the many years that the asset may be used. See visual on slide 55! Its better to think of a purchase that depreciates as an asset that will last 5 years and then you depreciate those assets over the 5 years instead of at once.

Statement of Cash Flows

-Describes sources and uses of cash. -Important for financial managers who manage cash. -Important for setting budgets and planning. -Not critically important for marketing managers.

Apply Survival Analysis

-Examine start dates and end (defect) dates of individual customers. -Model likelihood of defecting for individual customers or groups of customers.

Cohort and Incubate

-Go back into your database and track a cohort of customers from 5 or 10 years ago. -Typically aggregate data ("avg customer")

Sarbanes-Oxley (SOX)

-Inspired by Enron and similar cases. -Act of US congress enacted in 2002. -Top officers of the firm (e.g., CEO, CFO) must vouch for the accuracy of financial statements. -If statements are later found to be fraudulent, top officers face civil and criminal penalties. -Whistle blowers are protected. -Rather than spend the resources to comply, some firms have delisted themselves from US stock exchanges. -Although most marketing managers do not have to be financial experts, if you rise high enough in the organization, you will need more financial expertise. According to this law, the firm's top of officers must vouch for the accuracy of financial statements. If fraud is later discovered in the financial statements, these officers will not be able to claim in court that they simply did not know of the fraud.

Interpreting ROI and IRR

-MP ROI has already been discounted into present dollars, so any positive return is attractive. -To be attractive, IRR should exceed your WACC (discount rate). -Note than because the discount rate is already built into the multi-period ROI, and therefore it is already in present dollars, any positive return is an attractive investment. In looking at IRR however, the result must exceed the discount rate in order for the investment to be attractive.

Mathematical Models

-Method of simulating real-life situations with mathematical equations to forecast their future behavior -Help explain how important variables (metrics) interact to affect outcomes of interest. -Metrics are descriptive, models can be prescriptive.

Pro Forma Statements

-Pro Forma statements are estimates of future statements (balance sheets, P&L's, cash flows). -Pro Forma's are helpful in showing "finance types" how marketing plans will likely affect the bottom line. -Pro Forma's can also be "what if" scenarios, exploring how the bottom line might be affected by different decisions. -It is very common for marketing managers to make projections about how profitable their operations will be in future periods. These projections are sometimes called pro forma statements. For example, a marketing team might suggest changing marketing tactics in a way that increases or decreases marketing spending. They would show projection on how sales would change, and therefore how operating margins would change. Such projections could be used to justify increases in marketing spending. See example on slide 64...

Decision Stages in a Conjoint Study

1) Determine Relevant Attributes and Levels (and full or fractional factorial) 2) Choose Stimulus Representation (Full-profile? Trade-off procedure? Adaptive conjoint? Discrete choice modeling?) 3) Choose Response Type (Ratings? Ranking?) 4) Choose Criterion (Liking? Purchase intent?) 5) Choose Method of Analysis (Regression?)

Points to Note About A-T-A-R Model

1. Each factor is subject to estimation. Estimates improve with each step in the development phase. 2. Inadequate profit forecast can be improved by changing factors (A, T, A, R). If profit forecast is inadequate, look at each factor and see which can be improved, and at what cost. The real power in the ATAR model comes from being able to do "what-if" analyses (sensitivity analyses). For example, you could explore how an increase in advertising could increase awareness which would in turn increase total sales. Or, you could explore how much increasing distribution, that is increasing availability, could increase sales. By exploring any of these differences compared to the cost of making the differences, you could estimate the ROI of any of these changes to your marketing plans.

Four questions to use to identify critical issues:

1. How is the process currently performed? 2. How much does the current method cost? 3. What is wrong with the present method (and what improvements are needed)? 4. What value would the improvements have? (-->Pricing implications!) One way to identify the critical issues in a business-to-business market is to use these four questions in your thought processes and interviews. Notice that the last question, related to the value of possible improvements, has clear pricing implications. Once you understand how much money a new product might save a business-to-business buyer, you have a much better idea of how much you might charge for that product.

CLV from Survey Data

1. In what year did you first visit our company? => r 2. About how many times a year do you visit our company? => f 3. How much money did you spend at our company today? => s $margin = f*s*(%margin) CLV = m(r/(1+i-r)+m

Payback Period

= The period of time required to recoup the funds expended in an investment = time required for an investment to achieve break-even. See slide 89!

CLV equation

=Margin * (Retention Rate / 1 + Discount Rate - Retention Rate) Example on slide 117

Simple CLV with Initial Margin

=Margin * (Retention Rate / 1 + Discount Rate - Retention Rate) + Margin Example on slide 118

Profit

=Revenue - Costs

Demonstration of Causality

A causal relationship is nearly impossible to prove, but a skilled researcher can offer very compelling evidence. The three conditions that must be demonstrated to show causality are: -Correlation between cause and effect -Appropriate sequence of events -Absence of other plausible explanations

Evaluating Taxonomies

A taxonomy is "a division into ordered groups or categories." In business it is common to hear of taxonomies such as the four interpersonal styles, the six pillars of quality, five types of leaders, or the five competitive forces. Goof taxonomies -mutually exclusive and exhaustive -resistant to further subdivisions or aggregation -should be helpful

Demographics of Site Visitors

Advertisers will pay more for certain demographics, but how do we know the demographics of anonymous visitors? Cookies and log-file approach: Guess demographics based on web behavior -Omniture, WebTrends, Coremetrics -Limited demographics, overestimates individuals because multiple computer use => multiple cookies Panel-based metrics: Consumers volunteer to complete long demographic survey and install software on home computer that tracks surfing -ComScore, Nielsen Online -Great demographics, but uses only home computer (not work) and so may underestimate traffic. There are several ways that we can estimate the demographics of our website visitors. One approach is to place cookies on the computers of our website visitors. Then, by tracking where they go on the web and observing their interests, we estimate what their demographics probably are. Several companies are involved in helping websites to use this approach, including Coremetrics. Another approach is to use a panel. In a panel, consumers volunteer to complete a demographic survey and allow themselves to be tracked traveling all over the Internet. Their surfing behavior is then used to estimate the demographics of various websites. Nielsen online is one example of a company that can conduct this analysis for you.

Share of Wallet

Amount we sell to customer/Total purchases by customer is a closely related phrase to, or perhaps another way to say, "share of customer." "Share of customer" is an important concept that Peppers and Rogers introduced. We often strive to increase revenues by increasing market share. The horizontal rectangle labeled traditional marketing in the figure shows that this company is serving a fairly narrow range of customer needs but serves a fairly large number of customers. This would be analogous to a bookstore opening several new locations in other parts of town to sell similar products to more and more people. Amazon started selling books, but learned that they could be profitable by selling more than books to the same people. The one-to-one marketing vertical rectangle shows this approach, where a company might have a smaller number of customers but they meet a wider variety of needs and thus achieve profitability. Of course, Amazon now has a box in this figure that looks closer to a square, where they have a very large number of customers and they meet a very wide variety of needs for those customers. To meet a wider variety of needs, you need to know more about your customers.

Full Profile Conjoint Analysis: Laptop Computers Example

Assume three attributes of laptop computers, with two levels each: Weight (6 pound or 10 pounds) Battery life (2 hours or 4 hours) Size of hard disk (20 or 40 Gigabyte) Task: Rank order the following combinations of these characteristics from 8 = most preferred to 1 = Least preferred In this example, imagine we want to consider how consumers would trade off the weight, battery life, and size of hard disk, for laptop computers. Notice that each of these three attributes has two levels. In this design, we would need 2×2×2, or 2^3, or 8 unique full profiles to describe all of the possible laptop computers. See slide 97-100

Importance Weights

Attribute importance weights can help in: 1. Defining the optimal combination of features for a product. 2. Finding the relative contribution of each feature to the overall product's value. 3. Using (1) and (2) to estimate market share for new products. 4. Developing market segments based on individual differences in feature importance weights. Methods for determining importance weights include Direct Measures, Preference Regression, and Conjoint Analysis. In the preference model (or compensatory model) just discussed, one key component is the w's, or what we can call the importance weights. Having good estimates of these importance weights is the key to understanding what consumers really care about in your product or service and how you might improve your product to best meet their needs and beat out the competition. In more complicated models, it is even possible to estimate the market share of a new product before you roll it out if you have good estimates of attribute importance weights. In the slides that follow, we will talk about several methods of estimating importance weights. Among the direct measures, we will talk about ranking, rating scales, and constant sum scales. Among the conjoint techniques, we will talk about trade-off analysis, full profile conjoint, and discrete choice modeling, otherwise known as choice based conjoint.

Banner Pricing Basics

Average CPM rates for different types of banner ads and for various targeting strategies vary greatly. Inexpensive text ads without any targeting can run for CPM rates of only cents with an average of less than $1. General purpose targeted banners are more expensive, and CPM rates range from several dollars to about $15, with an average of about $7. Highly targeted multimedia banners aimed at popular key words on high-traffic, specialized sites can fetch a CPM rate of as much as $30, with an average CPM close to $15.  I include this slide to give you a sense of typical cost per thousand rates for banner ads. Three categories of ads described on this slide range from $1 to $7 to an average of $15 CPM.

Cost per Acquisition

Average acquisition costs = Cost per acquisition (CPA) = (total acquisition spending)/(# new customers) Or in direct marketing, CPA = cost of acquisition mailing/response rate e.g., a $1.00 mailing with a 1% response rate means 1/.01 = $100 cost per acquisition The cost per acquisition represents how much money, on average; it costs to acquire one new customer. I usually compute cost per acquisition separately from customer lifetime value and then compare the two. If the cost per acquisition is higher than the customer lifetime value, then you don't have a viable business model.

Price per Unit

Average price per unit ($) = Revenue/units sold = price of SKU1* SKU1% sales + price of SKU2*SKU2% sales = average of price weighted by percentage sales A "statistical unit" is a set of SKU's with fixed weights. Changes in avg price per unit may be due to price changes or differences in sales within a bundle of SKU's; changes in price per statistical unit only happen when price changes on one or more SKU's. Example: Lift tickets, tuition We have discussed price per unit as if it were a single price for a single item. In reality, price per unit can be much more complicated. For example, in order to determine the average price per ounce of Coors beer, all of the different ways that Coors beer is sold would need to be considered, including six packs, cases, pony kegs, full-size kegs, etc. Each specific product is referred to as a SKU, or stock keeping unit. To determine the average price per unit, you would need to multiply the price of each SKU times the percentage of sales coming from that SKU, and then average these over the number of different SKUs for the product. One difficulty in understanding price when looking at the average price per unit is that the average price per unit may vary from quarter to quarter even if the prices marked on the various products do not change. The average price per unit could change if the percentage sales coming from different SKUs changed from quarter to quarter. For example, the average price per ounce of beer in a keg is much lower than the price of beer in a six pack. If a company sold relatively more beer in kegs in one quarter than another, the average price per unit would be lower in the quarter with more keg sales, even though none of the price tags had changed. To avoid the problem of apparent price changes when the actual price on each SKU does not change, the concept of a statistical unit is used. In computing the price per statistical unit, the weights of the SKUs are fixed. For example a company might simply take the average price per ounce of one six pack, the average price per ounce of one case, and the average price per ounce of one keg, and then take the average. If the average price per unit changes from one corner to the next, it must be that the price marked on at least one of those products has changed. Another common example in Colorado of a product that is sold in many different ways is ski lift tickets. To determine the average price per skier day, a resort would have to consider tickets purchased for a single day at the mountain, tickets purchased for multiple days at the mountain, tickets purchased at a discount at a retailer in Denver, etc.

Average Retention Costs

Average retention costs (ARC) = (total retention spending)/(# retained customers) Or in direct marketing, ARC = cost of retention mailing/response rate e.g., a $2 mailing with a 40% response rate means 2/.40 = $5 average retention cost The average retention costs reflect the cost of retention, on average, of a single customer. Retention costs can be important to track in customer lifetime value models to be sure that your retention costs are not getting too far out of line as the customer ages. Generally, retention costs are lower than acquisition costs, so investments in retention often have attractive ROI.

BZ's

Below Zero Goal: Create incentives to make them profitable or encourage them to switch (Dogs)

Behavioral Responses

Behavioral intentions Purchase intention: how likely would you be to buy... Switching intention: would you be willing to switch... Behavior Usage: have you every used/purchased ___? Frequency: how often do you... For new products or proposed changes in products, intentions are easier to measure than behavior. We can measure behavioral intentions or actual behavior. Actual behavior will generally be more predictive of future behavior than behavioral intentions. However, for new products, actual behavior will be difficult to measure because the product is not in the market yet, and therefore behavioral intentions may be the only measure we can use.

Comparing Ad Costs

Bigbuckinvestor.com is considering three different ad buys: Dollarsaway is offering a sponsorship deal; $300 for 6 months would buy a button on the site. The site gets about 100 visitors a day. Banner ads on slowbooks.com cost $5 CPM, and about 200 visitors a day will see the ads. 401myway.com is offering ad space at $6 per click. They estimate they can offer 50 ad views a day. All promotions are expected to have a CTR of .1%, and a CR of 2%. Which site offers the best CPA? Answer on slide 19.

Long-term Liabilities

Bonds Payable Mortgages

Bonds payable

Bonds payable are longer-term financial instruments that often run for several years. Mortgages are loans that are often attached to real estate and may have terms of 15 or 30 years.

*How to compute book value

Book Value = Retained Earnings - Intangible Assets The book value of a company is the value of its tangible assets minus its liabilities. Note that the book value does not include intangible assets, such as goodwill or trade names. In theory, if you purchased a company for its book value, you could sell all of its assets, pay off all of its liabilities, and break even. *If a company's book value ever exceeds its market value, it may be at risk of someone buying and liquidating the company, selling off all of its assets, paying off all of its loans and firing all of the employees.

Brand Equity

Brand Equity is the amount (or percentage) a consumer would pay for a brand over and above a fair value for the product assets. Price premium? What about discount brands? Consider also loyalty/retention => CLV Many of the points and even specific examples used in describing brand equity are taken from the Keller article assigned for this class.

How to compute the break-even volume

Break-Even Volume= Fixed Costs/ Contribution per Unit Break-Even Revenue= Break-Even Volume * Price per Unit The break-even volume is the number of units we need to sell in order to cover our fixed and variable costs. The break-even revenue is the revenue we would receive from selling the break-even volume. SEE EVENT EXAMPLE ON SLIDE 79! Conceptually, the break-even volume is the number of units you must produce and sell in order to cover your variable and fixed cost combined. To conduct a meaningful break-even analysis, you must first distinguish between variable costs and fixed costs. -Variable costs are directly related to the number of units produced. That is, if you produce one more unit, you must spend X more dollars in variable cost. -Fixed costs do not change with production volume. In many examples the fixed costs are the costs of buying a factory or a piece of equipment. In marketing examples, the fixed costs might be the cost of developing an advertisement, developing the creative for a brochure, or renting a facility for an event. See slide 70. The contribution margin refers to the profit you make on selling each unit.

Issues in the Choice of Decision Tools

Break-even and Payback don't explicitly consider the time value of money. NPV shows actual "money made," and thus may favor larger projects. (Project size is not standardized.) NPV is sensitive to choice of discount rate. IRR assumes one rate of return across all periods. IRR assumes reinvestment at IRR rate. IRR can give multiple answers if there are several positive and negative cash flows. ROI is highest on initial investments.

Financial Decisions

Breakeven Payback ROS ROI NPV IRR MP ROI ROMI

The A-T-A-R Model: Definitions

Buying Unit: Person, household, or department/buying center (in B2B). Aware: Has heard about the new product (measured with aided or unaided recall, expressed as percentage). Available: If the buyer wants to try the product, the effort to find it will be successful (expressed as a percentage). Trial: Purchase or consumption of the product (expressed as a percentage). Repeat: The product is bought at least once more (expressed as a percentage of triers). See slide 115, 116, 117, 118

CPM, CPP

CPM = Cost per thousand (1000 = mille in Latin) = (cost of advertising)/(1000 impressions) Cost per point (CPP) = Cost per rating point = (cost of advertising)/GRPs We have seen cost per thousand before, and the same concept can be applied to traditional advertising. Another relevant concept is cost per point, which reflects the average cost per gross rating point. Some media (television, radio, magazines, etc.) and some vehicles (specific TV shows or specific magazines, etc.) may offer more attractive cost per point than others. These cost differences affect media purchase decisions.

Click Through Rate (CTR)

CTR = percentage of viewers who click on an ad, = clickthroughs (#)/Impressions (#) CTR may count the same individual twice (counts clicks, not people) Beware that some clicks may be worth more than others. The click through rate is the ratio of click throughs to the total number impressions. Just because an ad gets a high percentage of click throughs, it may not be an effective ad. We would also need to consider what behavior happens after the click through. In a later slide, we will talk about effective yield as being a better metric for evaluating the effectiveness of a banner ad.

Cannibalization

Cannibalization Rate (%) = Sales Lost from Existing Products/ Sales of New Products Sales can be measured in units or dollars. Another important concept in estimating new product sales is to estimate the cannibalization rate. The equation for this rate is shown on the slide. Note that if the cannibalization rate where 100%, the company would experience no new net sales with the introduction of a new product.

Change in Production Technologies

Changes in production technologies with scale generally involve increasing fixed costs and reducing variable costs Fixed costs do not vary as production varies (e.g., machines, equipment, computer systems) Variable costs vary with production (direct labor, materials, packaging) As production volumes increase, it will generally make sense for companies to buy machines that can make more parts at a time or that can make parts more efficiently. Generally, this will involve spending more on those machines, in other words increasing fixed costs, while reducing labor which means reducing variable costs.

Personalized Multimedia Example

Client seeks better ROI than standard postcard requesting donations. Client has donor list of 14,000 former and active donors. List is segmented by prior donation level, former v. active donor, and geography. Copy tailored to segments, personalized with donor info. E.g., "Your gift of $50 on Sept. 14, 2006, personally impacted a child or family in a difficult situation." In this example, Measurable Marketing was invited to work with a client who had been using a standard postcard as part of a donor mailing in the past. Measurable Marketing suggested a campaign that used digital printing, personalized mailings, and personalized URLs.

Diseconomies of Scale

Communication costs Analogous to "coordination losses" on large teams Duplication of effort Related to miscommunications? Office politics More likely with multiple management layers? Top-heavy companies Assumes managers do not add value This slide shows factors that might increase cost per unit as volume increases. Although these effects are generally smaller than economies of scale, it is important to be aware of them. Communication costs are interesting because you may have experienced this on a group project. On a group project, the time you spend trying to set up meetings, trying to reach agreement within the group, or organizing your activities, are all sometimes called "coordination losses." These activities are not directly related to getting the work done, so they represent inefficiencies in the work process.

Making Research Valuable

Connect study results to ROI. -Start with a comprehensive model linking studies to attitudes (e.g., satisfaction), to intent (e.g., to purchase), and to financials. -Don't just confirm what you are already confident is true --Beware of overconfidence! --Debunk myths! Design studies to maximize information gain -Test several alternatives Design the research to support decision and action -Begin with the end in mind -Mock-up charts and tables before collecting data On this slide I summarize and extend some of the points made on earlier slides. To achieve high ROI in your marketing research, be sure that your research results will connect in some way to financials. If you I have no idea how your marketing research results might be connected to financial results, you may be wasting your research money. I have seen many research studies that seem to be designed to confirm what management already knew. Because there is very little information gain in the studies, they generally do not have attractive ROI. Instead, don't be afraid to test new ideas or to debunk myths (commonly held beliefs that are not based on data). It is good to go into your research projects with some idea of what decisions will be affected by your research outcomes. If no decisions will be affected, it's not clear how your research will have a positive ROI. Think about the decisions that need to be made, and the data that will inform those decisions, and then work backwards to how you will design the study to deliver the data in the format that will be necessary to make the best decisions.

When to Skim or Penetrate

Consider skimming when: Customers are not price sensitive. Economies of scale may not be available. Company lacks resources to grow quickly. Marketing resources Production capacity Consider penetration when: Reverse of skimming conditions and: Competition may enter soon Product may catch on quickly with large market The skimming strategy implies that the company will sell lower volumes. This may be helpful if the company does not have the resources or production capacity to quickly ramp up and sell large volumes of product. If the company had the resources, and especially if the company was concerned that competitors might soon enter, the company might be better off using a penetration strategy to capture as large a market share as possible early on.

Costs

Cost per Impression = Ad costs/impressions CPM = Ad costs/1000 impressions Cost per Click = Ad costs/clicks Cost per order = Ad costs/order Cost per Acquisition = cost of acquiring a new customer ≈ cost per order. Several terms can be used to describe the cost of online advertising. Of these, CPM is one of the most commonly used, and reflects the Cost Per 1000 (Mille is thousand in Latin) impressions. Cost per click is also commonly used, especially in pay per click advertising such as paid search advertising.

Three C's of Pricing

Cost-Based Pricing -Cost-Plus Pricing -Breakeven Pricing -Target Profit -Pricing -Markup Pricing Competition-Based Methods -Going-Rate Pricing -Price Premium Customer-Based Methods -Perceived-Value Pricing This slide introduces the three C's of pricing: cost-based methods, competition based methods, and customer based methods. While different companies seem to emphasize different methods, it is recommended that a company consider all three methods when setting price. I like the three-legged stool analogy. Just as removing one leg from a three legged stool will make the stool unstable, ignoring one of the C's when you set pricing strategy may lead to bad pricing decisions. Some marketing purists recommend that you begin with customer base methods, and then look at the other methods to see if the best price for the customer can lead to a profit or can be successful in your competitive environment.

CRM

Customer relationship management is a broadly recognized, widely-implemented strategy for managing and nurturing a company's interactions with customers and sales prospects. It involves using technology to organize, automate, and synchronize business processes—principally sales related activities, but also those for marketing, customer service, and technical support. The overall goals are to find, attract, and win new customers, nurture and retain those the company already has, entice former customers back into the fold, and reduce the costs of marketing and customer service. See slide 157

Price Discrimination Considerations

Do different segments have different elasticities? Are segments separable? Will price tailoring have an attractive ROI? Is the discrimination tolerable? Different product or service features Time of day, day of week, season, advance purchase Quantity of purchase Geography (international - texts?) Demographics: age (children, students, seniors), gender, profession (teacher, soldier) Other If a company wants to charge different consumers different prices, it is important to be able to separate the segments. For example, some publishers sell international versions of their textbooks online. These books are usually cheaper versions of the standard textbooks and are designed for international students who may be on very tight budgets. However, some US students may be able to purchase the same books online and save money. Because the buyer segments are not easily separable, the publishers may lose money by offering this cheaper version of the textbook. Various demographic variables are often used as a basis for price tailoring. It is important that these variables are acceptable to consumers, and this acceptability largely depends on cultural norms. For example, we might readily accept that on ladies night, women would get a discount and men would not. However, if the same bar offered "tall people night," and offered discounts to tall people, customers might be offended.

Economies of Scale: A closer look

Economies of scale exist when the cost per unit produced is lower (more economical) at high volumes of production (large scale) than it is at low volumes of production. Causes of economies of scale: Procurement bargaining power (Wal-Mart) Wider allocation of fixed costs, esp. advertising (GM vs. Chrysler) Experience curve Change in production technologies The four bullets on the bottom half of this slide describe the conditions that could lead to an industry having high economies of scale. To obtain procurement bargaining power, a large retailer may be buying in such large quantities that it can press its suppliers for lower prices. Therefore, the retailer's (e.g., Walmart's) costs will be lower as it grows larger and larger. In general, larger companies will be able to spread fixed cost over more units sold. This spreading of fixed costs extends to advertising expenditures. For example, because General Motors sells many more cars than Chrysler, if the advertising budgets were exactly the same, the advertising dollars spent per car would be lower at General Motors than at Chrysler. On the slides that follow, I will drill down a bit deeper into the experience curve effect and how changes in production technologies can bring about economies of scale.

A Healthcare Quality Metric

Established by National Institute for Health and Clinical Excellence (NICE, England and Wales). Qaly = Quality-adjusted life year Qaly = Improved life quality * years, e.g., improved from .5 to .7 (0-1 scale) for 15 years = .2*15 = 3 Qaly's. If treatment costs $15,000, cost per Qaly = $5,000. Drugs that exceed $45,000 per Qaly may not be approved. (Dept of Transport uses related cut off for safely improvements: $2.2 million per life saved.) Increasingly, innovative metrics are being developed for all kinds of services. This slide shows a metric used in the UK to make policy decisions about which medical treatments to fund and which treatments not to fund. As our technology advances, we will likely discover an increasing number of very expensive medical treatments that can help improve the quality of life somewhat. The model shown on the slide is on attempt to develop a decision rule for which treatments to use.

An A-T-A-R Model of New Product Sales

Estimate sales of new products using estimates of Awareness Trial Availability Repeat

Advertising Response Metrics

Even if the hierarchy of effects model is not perfectly valid, it may still be helpful to form objectives around these outcomes: Beliefs: Cognitive components of attitude. Awareness, knowledge Affective responses: feelings, emotions, evaluations, liking (subjective responses). Liking of image Behavioral responses: What consumers do May include behavioral intentions Although the hierarchy of effects model is not perfectly valid, we do know that there are several important intermediate effects between a consumer seeing our advertisement and the consumer purchasing our product. We can't be sure in which order these effects occur, but they are all important and in our metrics class, so we discuss how to measure each of these. These intermediate effects can be described in three broad categories, including beliefs, affective responses, and behavioral responses. These three broad categories come from long established theories in psychology and are the same categories we saw in the hierarchy of effects model.

Problems with Difference Measures

Every measure has some error. Taking the difference between two measures, each with some error, often compounds the error. Example: Suppose measure A has a mean of 5 and varies from 4 to 6 (+/- 1), while measure B has a mean of 4 and varies from 3 to 5 (+/- 1). It should be clear that A - B will average 1, but will range from -1 to 3 (+/- 2). Another problem with a score like the NPS is that it computes two different measures (the percentage of promoters and the percentage of detractors) and then computes the difference between the two. Both of the original metrics will contain some measurement error and sampling error. When we create a new measure that is the difference between these two measures, the new measure may have even more errors then the original two measures. Therefore, most people who study measurement carefully (psychometricians) recommend against using difference measures.

Museum Hierarchy

Example slide 68-69

Museum Pricing

Example slide 70-71

Cost per Acquisition (Order)

Example: What is the CPA if the banner ads cost $2.00 per thousand, the click through rate is .05% (.0005), and the conversion rate is .1% (.001)? Suppose CLV is $50,000. What's the ROI? Answer on slide 17.

Inventory Accounting

FIFO: First In, First Out: First unit inventoried is the first unit expensed upon sale. LIFO: Last In, First Out: Last unit inventoried is the first unit expensed upon sale. See slide 130 Products may be purchased at several points in time to replenish our inventory. Different prices may have been paid for the same products over time, as the price fluctuated in the marketplace. When a product is eventually sold, there is an interesting question of how much we should say we paid for it (the actual cost of goods sold). This slide shows to commonly used methods for identifying the cost of goods sold from inventory.

Driver Alert Sensor

Fifty-five percent of fatal traffic accidents in the U.S. involve straying out of lane. System senses stripes on road, alerts driver if car is drifting out of lane. Consider also automatically guided vehicles (AGV) in industry. Many advanced sensor based technologies are emerging that would make cars safer. In the early stages of the product lifecycle, these new technologies might be expensive and may only be of interest to professional drivers (trucking companies). By conducting in-depth interviews, you could determine the current cost per year of accidents related to drivers slipping out of their lane or dozing off at the wheel. The benefit of saving these costs could then be directly related to the price you would charge to a trucking fleet to retrofit their vehicles with this new technology.

Simulation Models

First-Choice Model: Estimate the first choice for each segment. Build up to total market share. Share-of-Preference Model: Take ratio of product A's utility over the utility of all brands in the market.

Fixed Cost Allocations

Fixed cost allocation means taking a fixed cost (e.g., factory lease) and spreading it out over several product lines (e.g., based on square feet used). Utilities, factory lease, insurance, management, etc. may be allocated. Allocation is somewhat subjective, and different methods of allocation may lead to different "costs" per unit. In many of our analyses, we have treated product costs as being primarily variable costs. However, many firms allocate fixed costs to the cost of goods sold. This makes the cost of goods sold reflect something closer to the total cost of production, rather than just the variable costs. However, the allocation of fixed cost is somewhat subjective.

Gross Rating Points (GRPs)

GRPs = Reach (%)* Average Frequency (#)*100 = Ratings Points*Average Frequency GRPs may exceed 100. TRPs = Target Rating Points; GRPs delivered to a target audience.

ACSI and Market Cap?

Fornell (University of Michigan) and his coauthors argued that a hypothetical, back-tested portfolio of stocks chosen based on their performance in ACSI outperformed the New York Stock Exchange (the Dow), the NASDAQ and the S&P 500, a finding that has since been supported by other researchers.

One-to-One Marketing

Four Implementation Steps: Identify your customers Names, addresses, habits, preferences, across all contact points, media, products, locations. Differentiate your customers Based on Value and on Needs Interact with your customers Push for the cost-efficiency (i.e., I-Tech) that will enable relationship marketing Customize some aspect of your organization's behavior toward your customers Mass customize a product or tailor a service. "Treat Different Customers Differently" At the dawn of the Internet marketing age, consultants Peppers and Rogers wrote a series of books on what they called one-to-one marketing. Even when I read current articles and books, I can see that many authors are borrowing from their model. For the next few slides, I want to describe this classic model for you, drawing on the original sources. One-to-one marketing involves getting to know your customers so well that you can begin to have personal relationships with each one of them, instead of treating them like one homogeneous group. Such personalization is typically only possible with state-of-the-art technology.

Internal Rate of Return (IRR)

IRR = The discount rate that results in a net present value of zero for a series of future cash flows after accounting for the initial investment. If you used the same numbers as the example on the previous slide, you could raise or lower the discount rate and see how the sum of discounted cash flows changed. When you change the discount rate to 14.6, the sum of discounted cash flows becomes zero. This is the calculation used to determine the internal rate of return. Generally, when projects have an IRR that is higher than the discount rate, the investment is attractive. In other words, the rate of return on the project is larger than the interest we would have to pay on a loan, and therefore if we borrowed money and took on the project we would make money.

Iconic Rote Learning & Mere Exposure

Iconic Rote Learning = Through repetition, people learn images, words, jingles, etc. (no reinforcement required). Mere Exposure = The more a person is exposed to a neutral stimulus, the higher the probability the person will like the stimulus. There are two concepts that can be helpful in understanding how increasing awareness happens and how it can lead to a positive effect of outcomes. First, iconic rote learning is a form of learning that occurs through repetition without any reinforcement. That is, consumers do not need to have a strong incentive to learn about your message and they don't receive any reward, such as a coupon, for responding to your message. Instead, simply through observing many repetitions of the same images, those images become etched in the consumers' minds. The same effect happens for words, phrases, songs, or any other content that is repeated in these messages. Mere exposure is the effect where a person develops positive feelings towards a neutral stimulus simply by being exposed to that stimulus several times. A consumer might begin to memorize your brand after seeing it many times in many advertisements and then the consumer may actually develop a positive feeling towards your brand merely because he or she is more familiar with the image. These two concepts provide some evidence that awareness by itself can lead to some liking for your product or brand.

Establishing Your Performance Measures

Identify all the objectives of your site (e.g., sell products, sell advertising, build the brand, reduce call center traffic, etc.) Prioritize your objectives Identify meaningful metrics of each objective Establish meaningful goals for metrics Track metrics and compare with goals This slide summarizes the strategy of using web metrics that we have discussed up to this point. It is important to identify the specific objectives of the site before identifying which metrics will be most meaningful and then implementing plans to improve those metrics.

Common Online Metrics

Impressions - Something (e.g., a banner ad) is served to a surfer. Pageviews - a page is served to a surfer. Hits - total number of objects (ads, images, text, etc.) served to surfers. Hits may be relevant to designing your IT capacity, but it's not a great measure of consumer behavior on a website. Visitors - number of people who visit the site. Several metrics are commonly used when tracking online behaviors. Impressions are a metric that marketers are concerned with because display advertising is often priced based on the number of impressions delivered. Page views can also be important because they reveal how deep someone went into a website. A hit could be related to any text, ad, or image being served, and because of this ambiguity, hits are not a very useful measure for marketing people. The number of visitors is also important because it reveals something similar to the number of customers visiting a traditional store. See slide 6.

Managing CLV at Aetna

In 2010, plans to lose 600,000 to 650,000 members. Sickly people are not usually profitable. Raising premiums may not work because sickly people are more desperate for coverage. Scaling back prescription drug coverage might help because sickly people value this more and healthy people won't notice.

Information Theory

Information Theory says that information is generated to the degree that prior uncertainty exists, and the research resolves that uncertainty. Information theory is important to consider again here because the value of the research is related to the amount of information that we can get out of the research. A somewhat technical definition of information gain is shown on the slide. See slide 131. A more technical definition of information gain, measured as entropy, is shown here. Measures of information gain are actually very important in some sciences, including computer science and adaptive testing. The calculations shown on the slide are beyond the intended scope of this class. Instead, I would hope you consider these two important takeaways: Information gain is more likely when you have more possible outcomes in an experiment, and The information gain is greater when those outcomes are equally likely (as opposed to one being extremely likely and one being extremely unlikely). Therefore, when you are setting up experiments or market research, try to include several viable alternatives and not just two.

Modeling Sales Take-Aways

Keep good records! Include any relevant data related to sales (e.g., weather). Information is power! A better understanding of the factors that affect sales may help with marketing tactics (e.g., knowing how weather affects attendance may guide changes in facilities). Statistically controlled sales has smaller variance than raw sales. Smaller variance will make it easier to detect differences in sales due to changes in promotion. To develop a good sales model you must keep good historical data. If possible, collect data on the smallest time period available, such as sales per day as opposed to sales per quarter. Sales per day data will give you about 360 observations per year, while quarterly data will only give you four data points per year. Because the daily data reflects a much larger sample size, the daily data will give you a more insightful forecast.

Accounting Principles (GAAP rules)

Key GAAP (Generally Accepted Accounting Principles) Rules: -Accrual: recognize revenue when transaction happens (not necessarily when the cash flows, e.g., principle payments are not an expense). -Matching: Match revenues and expenses for the period. -Capitalization: Expenditures that generate benefits over many time periods are depreciated or amortized. Expenditures that only generate immediate benefits are "expensed."

Fixed assets (non-current)

Land Buildings Equipment Less accumulated depreciation

The Expectancy Disconfirmation Paradigm

Learn chart on slide 7. Turning now to satisfaction, I want to introduce a classic model used in marketing to define satisfaction. In marketing, we think of satisfaction as a function of performance relative to expectations. Before consuming a product or a service, we generally have an expectation of what that experience will be like. Then, after the product or service experience, we can reflect on whether the actual performance was the same as we expected, or whether it was better or worse than expected. If the performance was below what we expected, the expectations were confirmed but in a negative way and this leads to a feeling of dissatisfaction. If our expectations were on target with performance, then our expectations were confirmed, and we will generally be satisfied. Finally, if the actual performance exceeded what we expected, our expectations are confirmed in a positive way. This leads to feelings of satisfaction. It is important to note then that performance is not the same thing as satisfaction. The performance might be very high, but if the expectations were even higher, the customer might be dissatisfied.

Scale Choices

Length: 7 +/- 2 Odd or even? Use what they used last time. Consider using an unbalanced scale Neutral point is not in the middle E.g., not satisfied, a little sat, somewhat sat, very sat, extremely sat. Intermediate points don't have to be labeled, but reporting can be difficult without labels. What does a 4.2 average on a 5 point scale mean? Somewhat satisfied? On the slide I offer some general guidance on designing metrics from survey questions. First, you might wonder how many points should be in a scale. Is a 100 point scale more accurate than a seven point scale? If you ever use a 100 point scale, you will find that many respondents do not use all of the skill points. In fact, responses may cluster around points like 60, 70, or 80. From many studies over the years, we have learned that a scale length of seven points, plus or minus two points is a pretty good length. In other words, scales of blank five, six, seven, eight, or nine, are all pretty good. There is some controversy over whether we should use odd or even scales. For example is a six point scale better than a five point scale? Noticed that in a six point scale, there is no middle point. This means that even on agree-disagree question format, the respondent must either slightly agree or slightly disagree; there is no neutral point. Some believe that nudging a respondent to take a position offers more valid data. I personally like including a neutral point because I believe sometimes people have neutral opinions. In studies I have seen comparing odd and even scales, it appears that either scale can give you valid information and there is no strong practical reason to favor one or the other.. My best practical advice if you were working for a company on a project would be to use whatever scale they used last time. That is, if they used a five point scale last year, use a five point scale again this year. That way, it will be easy to compare this year's responses to last year's responses. The potential benefit of this comparison should outweigh the potential benefit of switching to a scale format that might seem more valid. In terms of labeling scales, you may not have to label all the points on a scale. For example, you might label one as extremely dissatisfied and five as extremely satisfied, with no other labels. The data you get from this question should be valid, but it might be difficult to explain your results. Example, using the scale, if the average response is close to four, how will you describe this to management given that four had no label on the survey?

Affective Responses

Likert scales often used (agree/disagree) Affect = evaluation, valuing, liking (subjective) I like this brand This brand is a good value for the money This brand is high quality Quality is important to me Likeability Affective responses reflect how consumers feel about your brand or product. Because these are feelings, and there are no right or wrong feelings, affective responses are much more subjective.

Penetration

Market penetration (#cust prod purch/#cust in market) Brand penetration (#cust brand purch/#cust in market) Penetration share (Cust brand purch/cust prod purchase) Total number of active customers The slide shows several metrics that are used when we think about penetration. Market penetration is essentially the percentage of customers in a market who have purchased the product of interest. For example, when a product is new, we might identify a large potential market, but also see that only a small percentage has purchased the product early in the lifecycle. At this point, the penetration would be low. Brand penetration refers to the penetration of one brand of interest. Penetration share refers to the market share among the customers who have already purchased the product, recognizing that the penetration rate might still be low. The total number of active customers is the number of customers who have already purchased the product.

Apply Subscription Model Online?

Must have content not found for free on the web. Minneapolis Star Tribune experiment Special Vikings in depth coverage Three month subscription for $5.95 As one example of a paywall the Minneapolis Star Tribune has offered special coverage of the Minnesota Vikings football team available only to subscribers who pay an additional fee for this in-depth coverage.

Opportunity Analysis (OA) v. Traditional B2C Market Research

OA works particularly well in B2B markets. The nature of the customer need in B2B markets is generally functionally-based, and is therefore somewhat more logical. Inductive techniques may work well in B2C markets, but deductive techniques may be more useful in B2B markets. The Bacon and Butler approach to market research is completely different than what you might have learned in a typical market research class because this technique focuses on business to business products and not business to consumer products. One of the reasons these two market categories are so different is that in business to business markets buyers are typically focused on functionally based needs. In consumer markets, consumers may also have other, more subjective needs, such as social needs or emotional needs. Because the nature of the need is more functionally based in business to business markets, the need is more predictable and logical, and we can learn more from small carefully selected samples or what we will call deductive research techniques, whereas large samples are more useful in B to C markets, or what we will call inductive research techniques.

Depreciation calculations

One simple rule: Divide historical cost by number of time periods of use. E.g., a $40,000 asset with a life of 10 years depreciates $4,000 per year.

Additional Considerations

Outright sale or license? -License may reduce risk, limit rights Pay-per-Use -Good when usage varies across customers Subscriptions -May encourage increased usage over pay-per-use Leasing -Limits risk, helps with buyer's cash flow Most of our previous slides have suggested that in our pricing decisions, we only consider an outright sale of the product for a specific price. However, there are many other financial arrangements that could be used to capture the value of the benefits that we offer consumers. In your pricing decisions, I encourage you to use these other options creatively where effective. One example of a very effective pricing strategy change was in the early days of the Internet. Early on, AOL charged a fee per hour of time spent on the Internet. With this pricing strategy, consumers would log on, do what they had to do very quickly, and then quickly log off. Later, AOL switched to a subscription model, where you paid a set fee per month and could surf as much as you wanted during that month. As you might guess, there was a huge increase in Internet traffic with this pricing change. The boom in traffic was quite beneficial to AOL, because they were also making substantial revenues from advertising. By carefully considering all of the revenue flows, they made a wise decision about changing pricing policies for their online services.

*How to compute owners equity

Owner's Equity = Assets - Liabilities

How to compute the payback

Payback Period = Time required for an investment to achieve break-even. The payback period is closely related to our break-even analysis. If we know how many units we sell per week or month, we can calculate how many months it will take to achieve the break-even volume. This amount of time is the payback period. Practice on slide 84!

ROI of Disintermediation

Practice on slide 34-35. This slide shows an example of a company considering removing an intermediary, a process known as disintermediation. In this process, the company removing the intermediary must be careful to be sure that the benefits provided by the intermediary can still be offered to the next customer down the supply chain. This specific example makes a big assumption that the retailer will also lower price, and you should be aware that retailers will not always do so. This solution is very similar to other net present value and multi-period ROI problems that we have worked before. To get the change in profit, notice that the demand increased 30%, and the original sales were 1 million units per year. Therefore, the increase in unit sales would be 300,000. The profit to the manufacturer is still just $1 per unit, so the increase in profit to the manufacture is $300,000 per year. In this solution, be aware that only years zero through three are shown. To calculate the ROI over 10 years, I used a larger spreadsheet that is not pictured here.

Balance Sheet or Income Statement?

Practice on slide 62!

VoiceStream Wireless Example

Pushed alternative promotional concepts over the Internet among wireless users in new markets. Visual images with voice over used (standardized across respondents because digital) Discrete choice modeling used to determine importance weights. Fast, reliable results led to actuals 41% over forecast in the first month.

Rating Points, Reach, Frequency

Rating Point(#): percentage*100 of a defined population who receive at least one impression. (why not "see at least one impression"?) Reach (#): number of people who receive at least one impression. Reach (%): percentage of defined population who receive at least one impression. Average Frequency: average number of impressions per person reached. = impressions/reach

Return on Investment (ROI)

ROI, or return on investment, conceptually is the money we make on a marketing investment after covering the cost of the marketing investment, divided by the marketing investment. Return on investment is often expressed as a percentage. The investment is often conceptualized as the marketing dollars at risk. In other words, the marketing money that we spend is independent of any possible increase in sales. That is, if we had no additional sales, we would still lose this money. If we doubled sales, we would still spend exactly this amount of money. Unfortunately the term "return on investment" is used very casually and liberally in marketing and can be used to simply imply that our marketing efforts had some impact. I have heard people say things like "our return on investment was a $1million increase in sales." I don't see this as a measure of return on investment, but instead simply an increase in sales. In this class, I will use the term "return on investment" to specifically refer to these kinds of profit over investment calculations that result in percentages.

Return On Marketing Investment (ROMI)

ROMI = (Change in Revenue x Contribution %) - Change in Marketing Spending / Change in Marketing Spending If the project lasts more than one year, use MP ROMI, if it's less than one year, don't discount cash flows. Sometimes in marketing you might hear about return on marketing investment. We will compute this in the same way as other ROI calculations. Generally we look at the increase in profit (which could be the increase in revenue times the percentage contribution on that revenue) minus the investment (which could be the increase in marketing spending) divided by the investment.

How to compute the Return on Sales (ROS)

ROS (%) = Net Profit ($) / Sales Revenue

Direct Measures

Ranking offers good discrimination but ordinal data (can't do multi-attribute models) and is difficult with a large number of items. Itemized Scales (5-point or 7-point) scales are easy to fill out but offers only interval data (multi-attribute models are questionable) and may not lead to much discrimination or variance. Constant sum scales offer ratio-scale data but are tedious to fill out, especially for a large number of items.

More Lessons...

Ratchet up the sophistication over time. Excel works fine for beginning dashboards. Consider one tab for summary, separate tabs for "drilling down" into sub areas. Look online for ways to create "speed dials" in Excel if you want to dress up your dashboard. As you gain some experience, and hopefully some early success with your simple dashboards, you can increase the sophistication over time. See slide 170-173

Regression Analysis

Regression analysis can be thought of as a computer program for solving systems with several equations and several unknowns (variables). When the system has many more equations than variables, and no perfect solution, regression finds the best fitting solution. Some variables are generally found to be statistically significant (different than 0), while others may be found to be insignificant (≈ 0). Regression analysis can be thought of as a computer algorithm for finding the best fitting weights for a set of equations like those seen on the previous line. When using regression, we may find that some of the weights are not significantly different than zero. Such a solution would imply that some of the attributes are not important to that consumer in his or her making decisions.

Gross Profit Margin (GPM)

Relevant Pricing =(Sales-Cost of Sales) / Sales See example on slides 59 + 60!

Operating Profit Margin (OPM)

Relevant to Marketing Efficiency =EBIT/Sales See example on slides 59 + 60!

Social Marketing Tracking

Remember "Identification"? Ask visitors to enter various ID's (name, email, Facebook, Twitter, etc.) E.g., consider a sweepstakes. To enter, visitors must enter their ID's. Some time ago we discussed the importance of customer identification when we talked about 1 to 1 marketing. Customer identification can be particularly tricky if customers contact us in many different ways, including through Facebook, Twitter, direct email, etc. To achieve effective identification, we need to know our customers by all of their various names. This can be achieved by asking customers to fill out profiles, or to enter sweepstakes. When they complete the information forms, we collect all their identifiers from all of the web communities they participate in, and then we have a way to connect all of their identifiers.

Net Profit Margin (NPM)

Reveals overall financial performance =Net Income/ Sales See example on slides 59 + 60!

Market Share

Revenue market share Unit market share Relative market share Market (industry) concentration People often try to improve customer satisfaction in order to improve market share. There are several different measures of market share listed here. Revenue market share is based on the total sales in dollars of the competitors, while unit market share is based on the unit sales of all competitors. To understand the difference between these two metrics, consider the case of two competitors who sell exactly the same number of units, but one has a higher price. These competitors would have exactly the same unit market share, but the one with the higher price would have a higher revenue market share. Relative market share refers to one company's market share relative to the competition. For example, a 20% market share may not look large, but if that is the largest market share in the industry, that could be a great competitive advantage. Relative market share can be measured against the largest competitor in the industry, or the next largest competitor in the industry if the target under is the largest. Market, or perhaps more appropriately industry, concentration is a measure of where on the scale an industry falls between a few large competitors or many small competitors. One measure of industry concentration is the sum of the market shares of the top three competitors combined.

Online Ad Testing Costs Labor

Salary + 1/3 Salary = total annual labor cost (Add 1/3 for insurance, taxes, etc.) Total annual labor cost/2080 hours = $/hour For interns or part-time employees, straight $/hour may be an accurate estimate. It will take some time and energy to set up your online ad testing. Here we see how you might estimate the cost of that setup time. The top of the slide shows estimates for someone who is paid on a salary basis, and the last bullet shows how we estimate labor cost for interns or part-time employees (employees without paid benefits). Example on slides 54-55

Baseline Forecast

Sales may be a function of Season or month Overall economy Sector economic performance Past general growth rate Forecast with moving average model Forecast with regression Sales = b1(month) + b2(econ) + b3(sector) + b4(growth) The challenge in understanding sales lift is to understand what the baseline sales would have been if your new promotion had not occurred. To estimate baseline sales, you generally need a good forecasting model. Many different variables can be included in your forecasting model to try to accurately predict what sales would be if you made no changes in your advertising. Regression analysis, or some cousin of regression analysis, is often used to estimate sales forecast models.

Fractional Factorial Conjoint

See slide 104. Instead of using a full factorial (all possible hypothetical products) design in full-profile conjoint, we could use a fractional factorial (only some hypothetical products). This picture shows a hypothetical analysis that might include three attributes each at three levels, which would require 3^3, or 27 different full profiles. However, we could drop out some of those full profiles and still get good estimates of the importance weights. In fact, if we were careful about which ones we dropped out, we could get away with as few as 9 full profiles instead of 27. By reducing the number of profiles, the conjoint study could be easier for respondents and less expensive. Knowing exactly which full profiles to drop is tricky, and certainly beyond the scope of this class. Generally, specialized software is used to analyze our designs and to suggest the few full profiles that we must keep in order to run a good fractional factorial design.

Brand Investment - Short Term, Long Term, Very Long Term?

See slide 105

Carpet Cleaner Example

See slide 105-106. The slide refers to a classic conjoint example from a Harvard Business Review article from the 1970s. The paper was written by Paul Green, who is one of the most cited authors in the area of conjoint analysis. Some interesting aspects of this design include the use of different package designs and the inclusion of brand names, along with price. With these variables included, it would be possible to determine the brand-related price premium associated with each brand. RANKING BASED CONJOINT EXAMPLE

Remembering Product Placement

See slide 107

Pulsing Advertising Increases Recall

See slide 108

Additional Conjoint Example

See slide 108.

Classic Brands Reinforced Over Time

See slide 109 Mr. Clean is on example of a brand icon that has been used for decades. In one study, more consumers said they could recognize Mr. Clean than could recognize the Vice President of the United States at that time.

Discrete Choice

See slide 109. In discrete choice conjoint, a consumer is shown several full profiles, but instead of being asked to rate or rank the profiles, the respondent is asked to identify which one he or she would purchase. Thus, in many ways, discrete choice is similar to ratings-based full profile conjoint, except that instead of ratings, the respondents are asked to identify their top choice. Some feel this format is more realistic because it is more similar to actual choices. Discrete Choice in Teacher Performance Compare teacher attributes and choose which teacher should get a raise and how strongly you prefer that teacher on a sliding scale Sawtooth Software's Adaptive Choice-Based Conjoint (ACBC) "Adaptive" in that the software serves different choices to different respondents based on their earlier responses to gain the most information with the fewest questions.

Choices can become Ranks

See slide 110: Example: Suppose A is chosen from A and B. Suppose C is chosen from A and C. Then C > A > B Ranks: C = 1, A = 2, B = 3. With ranks, utilities can be found with regression. On this slide, I give a simple example of how if we asked a consumer to make choices among three products, A, B, and C, we could deduce their rankings from their choices. And once we have rankings, we could proceed with a regression analysis just like the earlier examples.

Share of Preference Model

See slide 114. This slide shows the equation that could be used to estimate the market share of a new product using the share of preference model. The model involves computing the overall compensatory score for the new product of interest and dividing it by the sum of all of the compensatory scores of all of the existing products and the new product combined. By generating estimates of market share, market share simulators can allow product managers to design products to achieve market share goals. The sum of your score divided by the sum of all the scores in the market

M/A/R/C Research's DesignorSM

See slide 115. This slide shows an example of results from a simulator, showing how market share and total revenue may trade off for various product designs. With this knowledge, marketing managers could select the product design that would best meet whatever goals the manager wanted to reach.

Pay Per Click Ad Words

See slide 12. In this screenshot, we can see a Google search using the phrase "MS Marketing programs." I've circled the terms "sponsored links" to show you the sections of this page that reveal advertisers who have bid on the search term used. If we were to click on any of the links in the sponsored sections, that advertiser would have to pay some small amount of money for that click. The three links on the lower left side of the page are not sponsored links, and these websites rose to this high position in the search just based on the content of their sites. These websites have apparently done a good job at search engine optimization.

The Simple Method

See slide 120. This simple method can be applied in situations where we are expecting to spend a certain amount on a promotion, and are wondering whether additional research should be done to optimize this promotion. For example, imagine a company considering spending $1 million on advertising. Suppose the company was considering spending $75,000 on research to improve the advertising spend before rolling out the campaign. Suppose they expected that this research could improve the spend by 10%, essentially improving the effectiveness of the advertising by 10%. The improved effectiveness behaves like the improved margin in this example. Thus, the ROI is the improved effectiveness: 10% of $1 million, minus the investment of $75,000, divided by the investment of $75,000. In this example, the resulting expected ROI is 33%.

Simulated Test Markets Volume Projection

See slide 121-124

Example of Simple Method

See slide 121.

Decision Tree Approach

See slide 122. Estimate expected value of outcomes with and without research using decision and probability trees. Then compute ROI. The decision tree approach is a somewhat more complicated method of determining the ROI of a marketing plan, and I've tailored the approach to understanding the ROI of marketing research. The method begins by understanding the concept of expected value. The expected value of a project is the sum of the probabilities of each possible outcome multiplied by the expected profit of each of those outcomes. For example, if we had a 50% chance of earning $10,000, and a 50% chance of earning $5,000, the expected earnings would be $7,500.

A Basic Decision Tree

See slide 123. Here I compare two alternatives. On the left side we see an alternative of doing no testing for a project and on the right side we see an alternative for doing some testing with the project. In both situations, the best possible outcome is $2 million in profit and the worst possible outcome is no profit. Notice that the probabilities of success change with testing; without testing the probability of the best outcome is only 20%, but with testing the probability rises to 80%. To estimate the return on investment of the research spending, we must compare the expected values of the no-test and the testing alternatives. Using the equation from the previous slide, we can see that the expected value without testing is $0.4 million, and the expected value with testing is $1.6 million. After considering the investment required to achieve this change in expected profit, we can see that the ROI of the testing is expected to be 380%.

Decision Tree with Mgt Uncertainty

See slide 124. Sometimes, the value of marketing research includes an increase in the probability that management will actually proceed with the project. This can be illustrated with a decision tree that has two stages to it, as shown on the slide. In the first stage, management decides whether or not to go with the project, and in the second stage, after the rollout, we discover whether we get lucky and achieve the best outcome or the worst outcome. Notice that on each branch, the probabilities must add to 1.0, or 100%. The method of computing the expected value for these two stage models as shown at the bottom of the slide. In this example, the expected value of this project at this point in time is $1.75 million.

Distribution Metrics

See slide 125 Numeric distribution, as a percentage, is the easiest to understand, but could be misleading. If your product were only in half of the stores you would have 50% numeric distribution. However, if the stores you were in were the largest stores, you would have greater than 50% all commodity volume (ACV). Product category volume (PCV) is more specific, and relates to sales of just your product category. ACV and PCV probably give more meaningful estimates of product distribution then simply looking at numeric distribution.

Decision Tree: Testing

See slide 125. Extending the example from the previous slide, now suppose that we could spend $1 million in testing and achieve the changes shown on the slide, including an increase in management confidence and perhaps a change in the marketing mix that increases the probability of achieving the most attractive outcome. With these changes, the expected value of the entire project is now substantially higher, at $6.3 million.

Facings and Share of Shelf

See slide 126

Supply Chain Metrics

See slide 127 The two metrics on this page are often used to evaluate the performance of supply chain members. A high out-of-stock percentage would mean that your product is frequently not on the shelf when it should be, and this could clearly undermine your sales. One reason for a product being out of stock could be that deliveries were not made on time. The percentage of on-time delivery is one metric used to track how reliably deliveries are being made.

Marketing Experiments Webinar

See slide 132. Over the next few slides, I show an example of market research testing from a webinar that I attended. In this example, four different options, or experimental treatments, were used, including a control landing page and three different new pages (only new pages 2 and 3 are shown in these slides). The use of multiple landing pages increases the chances of finding one very good page and as you can see, the use of that new page led to an increase in over 200% in total leads. Although the presenters did not reveal their ROI, it's easy to imagine that it was very attractive for this project.

The Direct Marketing (ROI) View of the Adoption Process

See slide 14. On the left side of this slide we see the traditional hierarchy of affects process that is commonly assumed when using mass media advertising. The notion that awareness will eventually lead to sales has been somewhat of an act of faith in this type of advertising. In direct marketing a more measurable hierarchy is generally used. This hierarchy emerged in print direct marketing, such as catalogs or other forms of direct mail. As the Internet gained popularity, the same type of thinking has been applied to online. By tracking cost per thousand, click through rates, conversion rates, and retention rates, we can gain much more detailed insight into consumer behavior and can compute much more accurately what the ROI is on our marketing spending. As Internet marketing has become more popular, we have seen increased pressure on marketing people to demonstrate the ROI of more traditional forms of advertising.

Seal & Moody Exhibit 3

See slide 14. This slide shows a graph that indicates how much larger a sample you would need if you used various types of dichotomized scales. As you can see, all of the percentages are positive. The graph indicates that if you report top box with a sample size of 500, you would need 100% larger sample then if you had reported the mean. In other words, in order to have the same accuracy and reliability in your analysis, if you would rather report top box than average scores, you would need twice the sample size. This substantially adds to the costs of your marketing research. Again, I'm not a fan of dichotomizing scales, but these are popular in marketing and managers seem to find them easier to understand than averages.

Modeling Sales to Detect Changes

See slide 140

Media Mix Modeling(a.k.a., marketing mix modeling)

See slide 142

Levels of Customer Identification

See slide 142!

Avista MMM Dashboard

See slide 143-149

ACSI Score

See slide 15. The American Customer Satisfaction Index metric was developed at the University of Michigan. (Being a Michigan grad, I admit to some bias in favor of this measure.) You can see that the ACSI actually includes three different questions that are closely related to customer satisfaction. By taking the average of these three questions, we end up with a score that is actually more reliable than any one question taken alone. Professor Claus Fornell was one of the developers of the scale, and Professor Fornell is highly trained in psychometrics.

Concept Test

See slide 68. The slide shows the typical five-point scale that's used in a concept test. Because respondents often overrate their likelihood of purchase, one industry rule of thumb is to discount their ratings by looking at only the top two box scores and taking only half of this combined percentage (50% of top two box score).

ROI Tracking

See slide 15. This slide shows how we could compute cost per acquisition from cost per thousand, click through rates, and conversion rates. In doing these calculations, be very careful with percentages and remember to keep your decimal places straight when you do any calculations using percentages. As we have discussed before, it is very important that the CLV be larger than the cost per acquisition. Also, the ROI can be estimated based on the difference between the CLV and the CPA divided by the CPA.

Use Regression to Estimate Several Effects

See slide 152

Church Giving

See slide 152 Distribution of donations to a church typically reflects a value skew. This means that a small number of individuals are very important, financially, to the organization. It is not unusual for these highly valuable customers to have more influence than the average churchgoer in the politics of the church.

Pipeline Analysis

See slide 153

Drive Customers to Interact via Cost-Effective Media ("Customer Migration")

See slide 154

Hypothetical Website

See slide 154 In this hypothetical example, several different metrics are used to track where visitors are in the process of becoming subscribers. These metrics reflect different steps in the sales funnel for this website. By examining the conversion rates from one step to the next, and across years, you can gain insights about how different changes to the website may or may not have been beneficial.

More Interaction, Greater Share of Customer

See slide 155

Sales Response Function

See slide 155

ACSI Benchmarks

See slide 16. One of the advantages of using the ACSI is that you can compare your company's performance to industry benchmarks. We can also see how different industries compare to each other on this measure. Notice that cable and satellite TV scores near the bottom. The scores are updated every year, and cable is typically near the bottom. This may be because high switching costs and low competition in this industry do not push the providers to offer higher quality service.

Personalization added costs...

See slide 162

Personalization Raises Costs, Lowers ROI, Increases Profit

See slide 163

Dashboard Examples

See slide 163-167

Branding & ROI Perspectives Work Together

See slide 20. Although we might think of branding and online as being separate marketing efforts, the two actually work together. When brand awareness is increased, we will likely see higher click through rates. So, if we have achieved higher liking of the brand through traditional advertising, we may see higher conversion rates and higher retention rates. Thus, our off-line marketing efforts can improve the effectiveness of our online marketing efforts.

B2B v. B2C Adoption Process

See slide 21. Any of the terms we have been using are more commonly used in B2C advertising. Sometimes the same words are used in B2B advertising but sometimes different words are used to reflect the terminology around a more typical sales funnel. The differences emerge in part because personal selling is much more common in B2B markets.

Consumer Hierarchy

See slide 23. This slide shows a classic model that I've seen used by many companies to describe their customers. Ideally, we migrate people from the bottom of the pyramid towards the top. At the top of the pyramid, we can make greatest use of one-to-one marketing and database marketing. At the very top, the advocates may be the same as the promoters in the Net Promoter Score model. These customers help with our marketing efforts by providing positive word-of-mouth.

Measuring Web Traffic

See slide 24 This slide compares three different methods of measuring web traffic and estimating the total number of unique visitors on different websites. As you can see, the metrics don't always agree. Finding good measures of the number of unique visitors and the demographics of those visitors remains a difficult challenge.

Reasons to be in Long Term Relationships

See slide 24. In case you have not seen this classic slide before, I include it to show how retaining customers in long term relationships can lead to a variety of benefits for the company. By paying close attention to customer satisfaction and the customer experience, we should be able to increase our customer retention rates. While we may lose money on the first sale to a new customer, this slide illustrates how in later years we should be able to get increased benefits from price premiums (loyal customers may be willing to pay slightly more for our products rather than switch to a competitor), lower service costs (because the customers are already in our system and understand our processes), and increased sales volumes (perhaps from increased share of customer) from loyal customers.

Combining Ad Revenue with Subscriptions

See slide 25-26 for example.

Trends in Newspaper Ad Revenue

See slide 28. Print newspaper ad revenues have dropped dramatically in the first decade of the century. Online revenues have increased, but the net revenues are down from previous levels, threatening the financial viability of many newspapers.

Newspaper Advertising Trends: National v. Local

See slide 29. This slide shows somewhat more current data indicating that newspaper ad revenue may have bottomed out in about 2009 and since then has been fairly flat.

Pizza Market Share

See slide 3. This industry has fairly low concentration, or we might describe this industry as being highly fragmented because there are many small competitors.

Newspaperdeathwatch.com

See slide 30 This website tracks how newspapers are doing and includes a list of newspapers that have failed due to a drop in advertising revenue. The list includes the Rocky Mountain News, which was a Denver-based newspaper and a competitor of the Denver Post. Many online content providers, such as newspapers, are experimenting with different types of online paywalls. With a paywall, some content would be available for free to all visitors, but other content, the content behind the paywall, would only be available if the visitor paid a fee of some sort, such as a subscription fee.

Fixed Cost Allocation Example

See slide 31. A factory is divided into two halves; one half is set up to produce product A, and one half is set up to produce product B. The product A half only runs one shift a day, the product B half runs two shifts a day. What percentage of the factory costs should be allocated to each product line? This slide poses a classic cost allocation question. There are actually several ways of answering this question that would be discussed in a cost accounting class. In our class, I simply want to introduce the idea that fixed costs can be allocated, but exactly how those costs would be allocated is beyond the intended scope of this class.

Channel Margins

See slide 32. Pricing is not just decisions about the price that the end consumer will see, but pricing decisions also affect pricing throughout the channel. Between the manufacture and the end consumer, there are many entities that provide benefits and tack on their own margins, raising the price slightly as the product moves closer to the consumer. On this slide I show how each player in the middle of the channel, otherwise known as intermediaries, may add their own margin. Looking at the some of the margins in the channel, it would be tempting to think we could cut out all of the intermediaries and offer the product to the consumer at a lower price. However, intermediaries usually provide value, and so cutting out one or more of these players (disintermediation) may reduce substantially the value that the consumer sees.

Software for Analyzing Log Files

See slide 35-36. To determine important metrics, track number of page views, or track number of visitors on certain pages, you will need web-tracking software. One of the most commonly used packages is Google Analytics.

Alternative Ad Formats

See slide 37.

Fixed and Variable Costs

See slide 37. As we dig deeper into cost based pricing methods, I wanted to remind you of the difference between fixed cost and variable cost as we have seen it on earlier slides.

Economies of Scale

See slide 38. This slide shows the relationship between production quantity and average cost per unit of a hypothetical product. The picture shows that the average cost per unit falls as the production volume (quantity) increases. Therefore, for this product, there is an economy of scale. As we will see in the slides that follow, economies of scale will be more available in some industries than others. In other words, not all industries have large economies of scale.

Smartphone OS Share

See slide 4. Contrast to pizza, the operating system industry for smart phones is highly concentrated. This diagram indicates that the combined market share of the top three competitors is about 95%. Industry dynamics, including barriers to entry, are very different in highly concentrated markets compare to highly fragmented markets.

The experience curve primarily relates to increased efficiencies in the use of labor. Simply put, as we gain more experience in doing something, we can do it faster and better.

See slide 41. This slide shows a hypothetical experience curve. Notice that big reductions in cost occur in the beginning as volumes increase, but later in the process reductions in cost are less frequent and less drastic. Recall that reductions occur in proportion to the doubling of production volume, so for companies that already have large production volumes, they would need much larger production volumes to see additional experience curve efficiencies.

Effective Yield: CTR isn't Everything

See slide 43 Click through rate is not the only metric to track. For sites interested in sales or some type of conversion to leads, you would want to look at your click through rate and your conversion rate. The product of these two metrics is known as the effective yield. Note that the ad at the bottom has a lower click through rate but a higher effective yield, indicating that this ad would be nearly twice as effective as the ad on the top with the higher click through rate.

Production Technology Example

See slide 43. This slide shows a different production technology on each side of the slide. The technology on the left has lower fixed cost but higher variable cost, and the reverse is true for the technology on the right. If the production volumes are low, the technology on the left will be more cost-effective, but if the production volumes are high, the technology on the right will be more effective. One point that became clear to me when I worked as an engineer was that the cost of producing a product is a function of how many units will be produced. When I would ask a engineer who specialized in estimating the cost of producing products how much my new product would cost to produce, he would always ask, "how many of these are you going to make in a year?" His estimates of cost per unit were always quite different depending on the volume of units to be produced in a year.

Sample Tracking Spreadsheet

See slide 44 Here we see a sample of the kinds of spreadsheets you might get when you run a substantial banner advertising campaign. Not all of your spreadsheets will show conversion rate or effective yield; they may, however, show other helpful metrics, such as average number of page views after clicking through. At the bottom of the slide, I show the sample sizes needed, in terms of impressions, to achieve various levels of accuracy in your click through rates. For example, to be able to estimate click through rate with an accuracy of ± .2%, you would need to show at least 10,000 impressions.

Test, Test, Test

See slide 45

Regression with Data from Banner Experiment

See slide 46

B-to-the-E Concept Test

See slide 69. Anheuser-Busch's new product Fruity-smelling beer Spiked with caffeine, herbal guarana, and ginseng This is an example of a hypothetical concept test for a product that Anheuser-Busch actually rolled out briefly some years ago.

Purchase v. Familiarity

See slide 72

A Simple Value Model

See slide 74. This slide suggests a continuum where product markets on the left side of the scale are more focused on functional or economic benefits, and the markets on the right side of the scale are more focused on emotional or social needs. B2B products tend to be on the left side of the scale and B2C tend to be on the right side of the scale, although there can be some overlap.

Competition-Based Methods:Consider Industry Structure

See slide 46. Having discussed cost-based methods, we now move on to talk about the second C, competition based methods. Industry structure is important to consider when reflecting on the competitive environment. In pure competition, industry concentration is usually low and there are many competitors. There also tends to be very low product differentiation and perhaps as a consequence, very low margins. This type of competition is sometimes referred to as a commodity market because some commodities, such as corn or pork bellies, are generally bought and sold as if there were no differences among them and price competition is severe. Our goal as marketers is generally to avoid playing in a commodity market, or purely competitive situation, because the margins are too low. We would like to find some way to differentiate our offering so we could at least be in a monopolistic competition situation. In monopolistic competition, there is some product differentiation and higher margins. In a monopoly, there is only one competitor and thus it could be said that the product differentiation is high. Usually, margins are also high although as we will see a few slides later, sometimes a sole competitor may choose to set price low in order to quickly penetrate the market. Oligopolies are still common but are a bit more difficult to understand. In an oligopoly, there is typically high industry concentration and few competitors. The differentiation may be high or low but is more typically low. Even with low differentiation, the margins are at least moderate and often high. High margins are possible in oligopolies, even with low product differentiation, because the oligopolists cooperate with each other and set prices as if they were one large monopoly. It is illegal in the United States for competitors to meet and set prices in a way that restricts competition and allows for competitors to make unusually high margins. However, even without meeting, many competitors recognize that there is a potential for oligopolistic behavior and hold their prices high. As long as there is no evidence of meetings or formal agreements, oligopolistic behavior can continue. Industry concentration is mentioned again on this slide. I want to remind you of the definition and examples of pizza and cell phone operating system industries on the following two slides.

Boston Consulting Group Matrix

See slide 5. The BCG matrix shows one classic way that a company can think about the various businesses that the company participates in. The horizontal axis in the BCG matrix is relative market share. Generally, we would rather be in markets where we have a high market share than in markets where we have a low market share. As markets mature (growth slows), if we have a large market share, the business might still be quite profitable and require little additional investment. This type of business, high market share and low growth, is sometimes referred to as a cash cow because it can provide excess cash that we can invest in other businesses or product lines.

Game Theory and Pricing

See slide 51. Game Theory helps explain why oligopolies exist, and why oligopolies are inherently unstable. On the left side of the slide I show the prisoners dilemma and on the right side I show the analogous oligopolist's dilemma. In the prisoners dilemma, we imagine two prisoners, A and B. Both are accused of a murder, but the police do not have enough evidence. If neither prisoner cooperates with the police, they may be convicted of a lesser charge, and serve a short prison sentence. If one of the prisoners cooperates with the police, that prisoner might go free, but the other prisoner might serve a longer sentence. However, if both cooperate with the police, both prisoners will serve long sentences. Thus, it is to the prisoners' joint advantage that neither talk to the police, but each will be tempted to talk to the police to try to be able to go free. The prisoners would be torn between cooperating with each other and competing with each other. The same behavior holds true for oligopolies, as shown on the right side of the slide. As long as both players keep their prices high, both will make good money. However, each might be tempted to "cheat", and lower their price, hoping to substantially increase their sales volume. However, if both lower their prices, a price war may ensue and the oligopoly may become a purely competitive situation, and neither competitor will see high margins. As pricing decisions may be made on a quarterly basis or even more frequently, the oligopolist's dilemma is played over many rounds, and the players will get a chance to understand each other's behavior. If both players can easily see what's going on, the oligopoly may be fairly stable and profitable. However, if players are able to cheat and not get caught, or if they can't easily see what their competitors are doing, the oligopoly may degenerate into a pure competition. The inherent instability in oligopolies is one reason why they tend to only exist when there are a small number of competitors.

Customer-Based Pricing:Demand Schedule and Demand Curve

See slide 53. Having reviewed cost-based methods and competition-based methods, we now move along to customer-based methods of pricing. In this method, it is very important to understand demand curves. The table on the left side of this page shows a demand schedule, and the figure on the right side of the page shows a demand curve. Usually, the demand curve line is downward sloping. That means, for most products, if you lower the price, you will sell more. There are some special cases where sales might actually increase if you raise the price. For example, with perfume, if you raised the price slightly, consumers might believe the perfume is of higher quality, and demand might actually increase. These cases, known as Giffen goods, are fairly rare, and the phenomenon only exists over a narrow range of prices.

Demand in Monopoly or Monopolistic Competition

See slide 54. This slide shows more detail of a demand curve. To understand the curve, we might imagine a group of people lined up according to what price they would be willing to pay for product. The people on the far left would be willing to pay the most for this product, and the people on the right would be willing to pay the least. Suppose there is a single price, X, for a product in this market. The person who would be willing to pay exactly X dollars would be paying his reservation price. Everyone standing to the left of this person would have been willing to pay more, so they would be very happy to pay just $X for this product. Because they did not have to pay as much as their reservation prices, they would probably consider the product to be a good value. The difference between each person's specific reservation price and the price they actually had to pay for the product is their consumer surplus. To marketers, consumer surplus represents value that we gave away to customers. Ideally, we would keep more of this value for our company. One way to achieve this ideal is by charging different prices to different customers.

Skimming v. Penetration

See slide 55. A skimming strategy means that a new market entrant charges a very high price at first and maybe drop the price slowly over time. A penetration strategy is the opposite; the competitor starts with a low price and hopes to sell a lot of products to quickly achieve market share and discourage other competitors from entering the market.

"Integrating" the Demand Curve

See slide 57. One way to charge different customers different prices is to offer slightly different versions of a product and sell each version at a different price point. The green boxes in this diagram represent additional revenue that the company would have earned by offering additional products at different prices. There is a multitude of ways to differentiate products and create a product line. I heard of a printer company that sold two printers, one was the fast printer and the other was the slow printer. The fast printer sold at a higher price. Interestingly, both printers came off the same production line and all printers started out as fast printers. In one additional production step, a worker went in and adjusted some fast printers so they would only print slowly, thus creating the slow printers. This example is interesting because the slow printers would actually cost more to produce, and yet we would charge a lower price. This example illustrates how basing prices just on cost might overlook important profit opportunities.

Four Classic Response Hierarchies

See slide 63.

van Westendorp Price Sensitivity Meter

See slide 66 & 67 & Reading.

B-to-B vs. B-to-C Market Research

See slide 75. The slide summarizes some of what I have been saying about business-to-business versus business-to-consumer market research. Because the business-to-business need is more functional, it is more logically predictable, that is, it would be easier to guess which product a buyer would want to purchase. In business-to-business markets, there may be many technical attributes to consider in making the purchase decision, but the decision rules that benefit buyers the most are fairly predictable once you have a deep understanding of the market. This can be obtained in business-to-business markets with a small carefully selected sample of industry experts. By conducting long, in-depth interviews with these experts, you can gain rich understanding of the buyer needs in a relatively short period of time (just a few weeks). Another reason why small samples work well in business-to-business markets is that there are often very few buyers anyway. If there are only 10 major buyers in a marketplace, it would be impossible to collect a sample of 300 on a survey, but collecting 300 on a survey would be fairly common in business-to-consumer markets.

Beliefs: Brand/Product Knowledge

See slide 77.

Opportunity Analysis (OA) Procedures (abbreviated)

See slide 78. After thinking through the four questions, you should be able to identify critical issues for a new product in a market. These critical issues can then be translated into hypotheses that you can use to focus your additional research efforts. These hypotheses are particularly powerful if they include a causal implication, that is, if they use the word "because" in them. Once you have a good set of hypotheses, you are ready to conduct additional interview-based research to address your hypotheses. The challenge in this third phase is to identify exactly the right expert to address each hypothesis.

Beliefs and Affective Responses Interact

See slide 79

GRP example

See slide 84

Reach and Overlap

See slide 85

Common Equation for the Multi-Attribute Model

See slide 86. This is an equation commonly used to understand and estimate a consumer's preference for a product. In my Consumer Behavior class, we referred to this approach as the linear compensatory model. The model assumes that if a product is very strong on one attribute that benefit might compensate for the product being weak on another attribute, and overall the product might still be attractive.

Advertising Response Curve

See slide 87

General Marketing Response Curve

See slide 88

Direct Measures: Ranking

See slide 88. Ranking is a fairly simple and common way to capture how consumers feel about the relative importance of attributes. One benefit of ranking is that it offers good discrimination; that is, you usually get a clear picture of what is the most important attribute, what is the next most important attribute, etc. The downside of rankings is that it gives you a type of data known as ordinal measures, which are somewhat imprecise. Also, if you have many attributes to consider, the ranking task gets very tedious. For example, lf you ask a consumer to rank 20 attributes, they might quit the survey at that point.

Advertising Effectiveness

See slide 89

Direct Measures: Itemized Scales

See slide 89. These types of scales are easy to use and common in market research. They offer the advantage of being easier to complete than rankings because respondents can score them one at a time. One downside of this measure is that consumers may get fatigued if you have many of these items and then rate all of the items as being very important. When this happens, you don't get much discrimination among the top attributes; many attributes will appear to be approximately equal in importance. Another, more subtle disadvantage of itemize rating scales is that this type of scale yields what is called interval data, which is better than ordinal data, but not as good as ratio data. Interval measures lend themselves to many statistical techniques, but can lead to trouble if used in compensatory preference models. Comparing this slide and the last slide, I would like to make a clear distinction in between ranking and rating. Rankings yield ordinal data, and ratings yield interval data. In common usage, some people use the term ranking and rating interchangeably, but these have very distinct definitions in measurement and marketing research.

Levels of Activity & Ad Payment

See slide 9. There are several different ways that people pay for advertising online. In a portal deal or sponsorship, you might pay a set amount to have a display ad featured on a website for a specific amount of time. Traditional banner ads are paid for based on impressions, usually using a cost per thousand, or CPM metric. In pay per click mode, the advertiser only has to pay if someone clicks on the advertisement or link. In this situation the advertiser may get some free exposure when visitors see the ad but don't click on it. In affiliate programs, the advertiser only pays if someone buys the product. For example, Amazon may allow vendors to display their products on the Amazon website, but these vendors may not have to pay Amazon until someone actually purchases one of their products.

Diversify Spending to Get Biggest Impact per Buck

See slide 90

Direct Measures: Constant Sum

See slide 90. The third type of direct measure is the constant sum scale. With this scale, consumers are asked to allocate points across many attributes so that the total of all the points is equal to some pre specified sum. The disadvantage of this type of scale is that it can be cumbersome for consumers to complete, especially if there are many attributes. When surveys are conducted online, many online survey software packages include a function that will compute a running sum for respondents as they complete the question, helping them with their math. One advantage of this type of scale is that it can give ratio scale measurements. That is, if a respondent gives 20 points to the first attribute and 10 points to the second, it is fair to say that first attribute is twice as important as the second attribute. In ordinal scales, these ratios don't hold; it is difficult to say quantitatively how much more important the attribute ranked 1 is than the attribute ranked 2. With interval scales, we get a better idea, but still an attribute rated a 6 is not clearly twice as important as an attribute rated 3.

Estimating Response Curves with City-Level Experiments

See slide 91

Which Campaign is better?

See slide 93-94

Trade-Off Matrix and Analysis

See slide 93. The matrix in the top left section of the slide shows the data collection format. The matrix on the lower left shows what the survey might look like after a single respondent completed it. As would be expected, this respondent ranked the maximum number of channels (250) and the lowest possible price ($100) as the top choice. The second and third choices, however, are less predictable and reveal something about the trade-offs that this respondent would make between number of channels and price. The matrix on the right side of the page shows how the data could be reorganized in an Excel spreadsheet to facilitate further analysis. Notice that these are exactly the same numbers but they have been rearranged to make analysis easier. To analyze the data, recognize that we have essentially 16 equations and 2 unknowns. The 2 unknowns, W1 and W2, are the importance weights for price and channels, respectively. By studying these equations and finding the best fitting W1 and W2, we can estimate the importance weights for the individual that completed the survey. With two equations and two unknowns, we can find exactly one solution for W1 and W2. With more than two equations, and given that human respondents are often a little bit inconsistent in their answers, instead of finding one perfect solution, we will find one solution that seems to fit the data better than any other solution, although the solution may still be less than perfect. The tool we use to find the importance weights with this data is regression analysis.

Trade-Off Analysis with Regression in Excel

See slide 95. This is the output you would see in Excel if you ran the regression using the data shown earlier. In this output, the column labeled "coefficients" shows the importance weights. The weight labeled intercept is generally not that important in conjoint analysis, but the weights associated with each attribute, price and channels in this example, are very important. We can see that both these coefficients for price and channels are statistically significant because the t-statistics have absolute values that are much greater than about 2.0 and the probability value (p-value) associated with each t-stat is much lower than .05. Therefore, in this case, we would conclude that both weights are important to this consumer. The r-square of .982 indicates that our solution to these equations fits the original data very well. R-square can range from 0 to 1.0, where 1.0 would be a perfect fit. The fact that the fit is close to 1.0 indicates an excellent fit. The solution is not perfect, but it is very good. By taking the ratio of the importance weights, it is possible to estimate directly how a consumer would trade off price and channels. It looks like this consumer would be willing to pay approximately $.47 for each additional channel. Other consumers may be willing to pay more or less per channel. With a large sample, we could estimate the reservation prices for channels for many consumers. Notice that the importance weight for price is negative. This happens because higher prices are less attractive. Other words, price is negatively correlated with overall preference if all other things were equal. One unusual thing about the use of regression in this context is that ranks were used as the dependent variable. Ranks yield ordinal measurement, and regression requires at least interval measurement, and therefore the assumptions behind the regression model are violated in this application. However, over the years, many researchers have found that the estimates given by regression in these situations are actually very close to the optimal estimates and therefore regression is still commonly used.

Full-Profile: More than two attributes at a time

See slide 96. In full profile conjoint, we ask consumers to consider several attributes simultaneously, not just one against another in two-at-a-time fashion as we saw in trade off analysis. The consumers are shown a complete description of a single hypothetical product, or "full profile" of a hypothetical product. Consumers may then be asked to sort the cards from their favorite to their least favorite, or they may be asked to rate the cards on some scale, for example a 100 point scale. (Shorter scales, such as 9 or 11-point scales are commonly used.) Because of these rating scales, this approach to conjoint is also sometimes called ratings-based Conjoint.

Keller's Dimensions of Brand Knowledge

See slide 98

Brand Awareness and Associations

See slide 99

Net Promoter Score (NPS)

See slides 9 & 10.

Benefits of Intermediaries

Selling: contacting potential customers. Assorting: creating sets of products from several sources. Sorting: breaking bulk Transporting: physically moving the product. Financing Grading (especially meat & produce) Marketing information and research This slide briefly lists the benefits that intermediaries offer. Most of these are straightforward, but I will touch on some that may be less familiar to you. Assorting is the process of pulling together products from several different places to create an assortment that would be attractive to the consumer. For example, a hardware store brings together tools, fasteners, paint, and many other things that would all be of interest to a consumer. Without this assorting, the consumer would have to contact all of these different providers individually to get the products he or she would need for a particular project. Sorting, or breaking bulk, is the process of buying products in very large quantity and then breaking up the quantity into the smaller sizes that consumers would typically like to purchase. For example, an intermediary might buy melons by the truckload, but most consumers only want to buy one melon at a time. Grading involves identifying the quality of different products, typically raw materials or produce. For example, eggs are graded as being grade A, grade AA, etc., and eggs are graded by size as well, such as large and extra-large. This process allows consumers to easily identify exactly the quality of product they are looking for and helps channel members to communicate clearly about what quality of product is being exchanged.

Share of Voice

Share of Voice (%) = Brand advertising/ Total Market Advertising Related to "competitive parity" method of setting the advertising budget. Share of voice is one approach in the "competitive parity" family of approaches to setting your advertising budget. Just as revenue market share is the ratio of your sales to the total sales in the market, share of voice is the ratio of your advertising to the total advertising in the market. To achieve competitive parity, you might want to set your share of voice equal to your market share.

Multiple Revenue Streams Examples

Ski resorts: Lift tickets, food, other retail Movie theaters: Ticket sales, snacks Amazon's Kindle (book sales) The slide shows examples of other types of businesses that have multiple revenue streams and may need to adjust their prices in some streams in order to maximize total revenue. For example, it is believed that Amazon sells its Kindle (its e-reader) at close to the manufacturing cost and therefore makes almost no profit on the Kindle. However, Amazon knows that if it sells more Kindles, it will sell more e-books, and apparently has done the math to figure out that the low pricing on its Kindle will maximize Amazon's revenue.

NPV & MP ROI Exercises

Slides 107-109

Implications of Diminishing Returns to Marketing Spending

Spending the entire budget on one method of promotion may take you past the point of diminishing returns. May need to spread budget to other promotional methods to achieve optimal ROI. Consider experimenting with several different methods in several different cities. Full and fractional factorial designs Because returns for spending will diminish at some point in any vehicle or medium, it makes sense for companies to diversify their advertising across several different media. For example, the Daniels College of Business spreads its spending across online, radio, print, and out-of-the-home (outdoor) advertising. Rather than buying radio advertising that would give one impression per city per year in United States, Daniels focuses its radio spending in select cities where we have drawn interested students in the past. Within those cities, we can spend enough to pass a threshold level of advertising and achieve some effective awareness.

Online Ad Payment Terms

Sponsorship: advertiser pays per time period or issue (e.g., $500 per month or $500 per issue) CPM: advertiser pays per 1000 impressions (e.g., $20 CPM) Pay-per-click (PPC): advertiser pays only for impressions that are clicked on (e.g., $3 per click). Search engine marketing (SEM) often uses pay-per-click (Google ad words). "Organic" search engine optimization (SEO) involves optimizing your site in the research engine rankings; no payments are made for "advertising." Commissions (e.g., affiliate programs): advertiser pays a percentage of sales generated from ad (e.g., 5%) Here is more detail about pay per click advertising. A distinction is made between search engine marketing and search engine optimization. Search engine marketing (SEM), also own as paid search, occurs when advertisers pay for certain words in a search engine. When someone searches with those terms, and then clicks on the advertiser's link, the advertiser has to pay a fee to the search engine. Search engine optimization (SEO), also known as organic search, occurs when websites are edited and fine-tuned to include certain keywords to make them more noticeable to search engines. A website that is optimized can achieve high rankings in a search, such as a Google search, without having to pay Google for that listing. Organic search is free to the extent that the website does not have to pay Google, however, the website would have to pay a talented developer or search engine agency to optimize the websites pages, so SEO is not generally totally free.

Lessons Learned from Assessment

Start with what's most important (e.g., most important learning goals). Don't start with all the measures you can get your hands on! Look for the simplest, "good enough" measures on the most important goals. Don't let the perfect be the enemy of the good. Even rough estimates will generate animated discussion on important topics. One thing I have learned is that it is important to identify your most important metrics upfront, and be sure to include those in your dashboard. If you start by collecting all the metrics you have, you will end up getting lost in the woods, and it will take you a long time to develop a useful system. Another thing I learned is that simple metrics are often good enough, and if you are waiting for the perfect metric, you may end up waiting too long. Just get something down that works pretty well, and you can refine the dashboard later.

Sunk Costs and Relevant Costs

Sunk costs don't change no matter what you decide. Relevant costs may or may not be incurred, depending on your decision. Sunk costs are not relevant to the analysis; they should not effect ROI. In these decision tree analyses, you must carefully consider which costs have already been committed to and which cost will change depending on your decision. If a particular cost would not change regardless of your decisions or the outcomes, then that cost is irrelevant to our decision tree analysis. In the sample problems online, I include a couple examples that illustrate when cost might be considered relevant or irrelevant.

Satisfaction Approach

Survey clients: -How satisfied were you with the market research you received? -Often uses satisfaction scales -Not really ROI Using the satisfaction approach to understand the value of marketing research, we might simply survey the people who received the research services and ask them how satisfied they were with the research. Typically, itemized rating scales, such as five or seven point satisfaction scales, are used for the surveys. Because these only capture satisfaction, this is not really a true ROI measure.

Willingness to Search

The "acid test" of brand loyalty? (Farris, p. 43) How likely would you be to settle for some other brand if Brand A were not available? Not at all likely (1)........Extremely likely (5) Some believe that the metric of "willingness to search" is the acid test of brand loyalty. As with the NPS score, I'm not convinced that any one metric is the answer to your problems, but several metrics combined might give you rich insights. This metric might be one to consider.

Debt/Equity Ratio

The Debt/Equity Ratio is a measure of financial leverage. Debt/Equity Ratio = (Total Liabilities)/(Owners Equity) Highly "leveraged" businesses have a high debt/equity ratio. These businesses typically have high interest payments, and are considered risky because if anything goes wrong and they can't make payments, they may quickly be in bankruptcy. Low leverage implies the firm may be able to borrow if necessary. The debt/equity ratio can be computed by quantities on the balance sheet. This ratio is estimated by the sum of all the liabilities divided by the total owners equity. A company with a very high debt/equity ratio can be considered highly leveraged. High leverage may be a risky position because it may mean that the payments on dept are high relative to the amount of assets available. If something changes for the business, such as a sudden drop in revenue, and they cannot make payments on their debt, they may not be able to borrow more money and they may go bankrupt. Companies with low leverage often have other opportunities to borrow money if they suddenly need cash, so they are in a less risky position.

Deceptive Pricing

The FTC (Federal Trade Commission) is authorized to take action against deceptive marketing practices, including Promoting a "sale price" that is really the everyday price Regular price may be determined by how long the price was offered (more than half the time) or how many units were sold. Check policies at the state level. Individuals may also sue companies for damages There are some interesting legal issues related to pricing. In particular, you have to be careful how you use the term "sale price." Companies have been taken to court before for having products on sale too often. Although these laws may vary, in some states, if you are on sale more often then you have your normal price, a court might consider your sale price to be deceptive.

Are Oligopolies Legal?

The Sherman Anti-Trust Act prohibits forming monopolies to restrict competition. "Price fixing," when competitors agree to set prices to act like a monopoly, is prosecuted under the Sherman Anti-Trust Act. Oligopolies are not legal. They essentially behave like monopolies, and like monopolies, they are regulated by the Sherman Anti-Trust act. The act of meeting with a competitor and agreeing on a price is known as "price-fixing" and is specifically prohibited by the Sherman Anti-Trust act. Proving that competitors had formal price-fixing agreements can be difficult, however, so many industries still operate as oligopolies.

NPV Discount Rate

The discount rate is the "cost of capital," which is the rate you would have to pay to get more money. Generally, we can get money by borrowing or by issuing stock (equity). The best estimate of the cost of capital is the weighted average of these two costs (WACC = weighted average cost of capital). Typically, the discount rate is set by the people in finance.

Experience Curve

The experience effect: the cost per unit decreases by 10% to 30% each time production volume doubles. -Specialization of labor -Operator skills improve. -Operators innovate. -Management reengineers processes. The experience curve primarily relates to increased efficiencies in the use of labor. Simply put, as we gain more experience in doing something, we can do it faster and better.

Conjoint Analysis

The multi-attribute choice models Attribute importances Direct measures Conjoint Analysis PC Example Carpet Cleaner Example Student Ratings Example Discrete Choice Modeling From Conjoint to Market Share

Customer Experience Management (CEM) Metrics

The service met my needs The service was easy to use The service was enjoyable Forrester combines these three to create CXi, their Customer Experience index. High CXi in 2012: USAA, Kohl's, Amazon.com; Low = telecom, health ins. Forrester research suggests these three measures as metrics of the customer experience. I thought that "easy to use" and "enjoyable" were interesting and different than satisfaction, loyalty, or likelihood to recommend, and might make useful metrics to include in your customer surveys .

Assets

Things of value that you own. Assets are recorded at "historical cost," not current market value. Assets may be depreciated over time. In that case, the balance sheet shows the historical cost and the accumulated depreciation (both are shown on asset side of balance sheet). Assets are generally recorded at the price that was paid for the asset in the past, not the current market value of the asset. This is done because it is often difficult to get an accurate evaluation of the current value of an asset. In order to reflect a more reasonable valuation of assets, balance sheets often show accumulated depreciation, which reflects the decrease in value for assets that may have worn out over time. Depreciation is also a useful mechanism for spreading the cost of an asset out over time, to allocate that cost to the time periods when the asset was actually used up.

Baseline Sales, Incremental Sales, and Lift

Total sales = baseline sales + incremental sales Lift (%) = Incremental sales/baseline sales Cost per incremental sales = marketing spend/ incremental sales Suppose lift were 10% on baseline sales of $1,000,000, and the incremental marketing was $40,000. Margin is 50% of sales. What is the ROI? See slide 137-138 Lift, as measured as a percentage, is the incremental sales from a promotion divided by the baseline sales. Before ROI became the hot term, lift was a hot term to describe how marketing analyst studied the effectiveness of marketing tactics.

Awareness

Unaided (without a cue), e.g., "What providers come to mind when you think of cell phone service?" Top of mind: % of customers naming a brand first in its category. Aided (with a cue), e.g., "Have you heard of brand A?" Consider distracters: Ask about fictitious brands. Ad awareness, can be aided or unaided, ask about ad, not brand. Beginning with beliefs, or cognitive outcomes, we can talk about awareness. The distinction is made between unaided measures of awareness and aided measures of awareness. Unaided awareness may underestimate the true awareness, and aided awareness may overestimate the true awareness. Advertisers are sometimes interested in ad awareness, which is different than brand awareness. A consumer may be aware of the brand, but not aware of a specific advertisement that was shown during a specific television program. Ad awareness can be an informative measure of the attention getting power of a particular advertisement.

Cost-Based Methods: Margin

Unit Margin ($) = Selling price - Cost per unit Margin (%) = (Selling price - Cost per unit) / Selling Price Markup (%) = (Selling price - Cost per unit) / Cost per unit Practice on slide 28-29: Exercise: a product costs $30 to produce and sells for $40. What is the unit margin What is the margin? What is the markup? The unit margin is essentially the same as contribution margin that we discussed regarding break even analysis. Markup percentage is a related concept but is computed differently. Markup percentage is commonly used in many retail settings. Look at the exercise at the bottom of the slide and the answers on the next page to be sure you understand the difference between percentage margin and percentage markup. Percentage margin is computed based on sales price, and markup is computed based on cost.

Common measures of "Brand Equity"

Y&R Brand Asset Valuator: Relevance, differentiation, esteem, knowledge Philips Electronics: Uniqueness, relevance, attractiveness, credibility Aaker: Differentiation, satisfaction/loyalty, quality, leadership, value, personality, associations, awareness, market share, price, distribution Moran: Market share, relative price, durability/loyalty Interbrand: Total earnings - earnings from tangible assets. The brand equity measures developed by the top two providers appear to use more indirect measures, and we see more direct measures used by providers lower on the page. Keller defined what is meant by direct and indirect measures.

Are Monopolies Legal?

Yes, but the Sherman Anti-Trust act prohibits companies from combining to create a monopoly if the resulting entity might inhibit free trade. Example: A large firm buys its only competitor and then restricts supply to increase prices. A monopoly is legal in the United States if you "grow your own" monopoly. That is, you must start with a patented technology, grow your company and maintain your monopoly over time. That is perfectly legal in the United States. What is not legal is when an existing company attempts to purchase all of its competitors so that it might achieve a monopolistic advantage. The Sherman Anti-Trust act prohibits these behaviors because they would restrain free-trade and would not be beneficial to the end consumer. Large acquisitions often require the approval of the federal trade commission (FTC), and acquisitions may not be approved if they would inhibit free trade.

Accounts receivable

money that is owed to you This asset refers to money that is owed to the company. Perhaps the company sold some products on terms that allow the buyer to pay cash in 30 days. These amounts would be accounts receivable, but are not yet cash. Supplies refer to items we might have for general use in the office or factory, such as copier paper, pencils, and so on.

Depletion

is the "depreciation" of a natural resource (e.g., a gold mine, an oil well).

Accounts payable

money you owe, but you haven't paid it yet accounts payable generally reflect amounts that we owe other companies for items we purchased on credit. Notes payable generally refer to money we borrowed from and owe to a bank. Unearned revenues are an interesting category. They are somewhat the reverse of prepaids on the asset side. When a company gets paid in advance for something they have not yet delivered, that would be unearned revenue. In a sense, the company now owes something of value to another company, so we should think of it as a liability.

How to compute the profit-based sales target

target volume = (fixed costs + target profit) / contribution per unit Practice on slide 82!