MIS 2

Rent The Runway: Fulfillment as a Service

-ROI for fashion rental -nww logistics arm is another revenue form for the firm

Hasting and Quikster (netflix)

-tried to charge $8 and $ separately for internet and DVD; faced backlash -also rebranded DVD service quikster, making it more difficult to get product since internet and DVD service were separate -hastings was a laughing stock and lost 300,000 customers -their market share fell 12 billion

why facebook

-understanding facebook helps better understanding how Web 2.0 works for consumers, companies, and advertisers

Facebook's social graph

-global mapping of users, organizations, and how they are connected is called the social graph -coined by m.z. -you are the center

Disruptive Technologies: Understanding Giant Killers and Tactics to Avoid Extinction: Introduction

-kodak went from 90% market share to being crushed, all because of disruptive tech

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: good at managing cash

-pay suppliers as account payable (due in future) -negative cost conversion cycle (CCC), as they get cash from customers right away -this allows them to have an additional pool of cash and keep inventory turns high

The San Diego City Schools: Enterprise Resource Planning Return on Investment: Personnel

-processes complex in HR as well -in process of implementing SearchSoft, an applicant tracking system, hoped to complete by 2002

Big Skinny: Sponsored Search

-sponsored advertisers lets advertisers place their listings within search engines' results for specific keyword, allows advertisers to specify targeted keywords -BigSkinny would need to specify their keywords, and would need to pay for these, which is typically on a per click basis -highest bids on thin wallet

What job needs to be done?

-start w/ hypothesis-> get out and test (usually wrong) -customer segments problems/needs -What is c.s. trying to get done -is it a problem or need -what functional or social jobs are getting done? -emotional jobs -basic functions

Sunil Nagaraj

-university of north carolinna 2004 graduate -interned in program management at microsoft -first experience with tech start-up with ZeeWise, a developed software for franchised establishments -began harvard business school mid-2007, graduated 2009 -him and colleague Jack, agreed to begin a tech start-up and invested in 5000 each

ad trackability

-unlike magazines and tv, internet ads can be tracked instantly -online tools allow advertisers to calculate ROI, test creativity, and adjust ads quickly -tools allow you to see performance of site by region, category, date, and other criteria

Customer validation: be sure your startup vision isn't a hallucination

-you have PMF 1. you get ready to sell 2. get out and sell 3. develop positioning (how can I sell better?) 4. verify/repeat (are you ready to market?) -go back and pivot -physical or web/mobile

COUNT

-returns # of rows that match a specific criteria 1. SELECT COUNT(columnn_name) FROM table_name WHERE condition

The San Diego City Schools: Enterprise Resource Planning Return on Investment ERP System Benefits

1. cost savings 2. productivity improvements 3. soft benefits -annual benefits -to generate ROI, would have to base on 5-year useful life

DELETE FROM

DELETE FROM table_name WHERE condition;

INSERT INTO

INSERT INTOA table_name (column1, column2) VALUES (value1, value2) -don't need to specify for values added to all columns

MIN

SELECT MIN(column_name) FROM table_name WHERE condition

cost per action

advertiser pays each time a person does something (sale, lead, etc)

sharing economy: future outlook: established players get collaborative

-alphabet: 1/4 billion in uber, 1 billion in lyft -gm .5 billion in lyft -volkswagen $300 mil in gett -apple $1 billion in uber of china -RTR: vogue, lucky, glamour -walgreens and tech rabbit for drug store deliveries -IBM and deliv -AW hotel (NYC) and desks near me -marriott and hostmaker -GM cancelled partnership w/ turo -toyota discount for uber drivers -avis acquired zipcar

Strategic Concerns for Platform Builders' Asset Strength, Free Riders, and Security

-available on samsung, vizio, xbox, etc; developers pick up slack -weakened assets: ca undermine NE and SC (if can access on other sits, why go to FB at all) -revenue sharing: expensive data, too much data= free rider problem; security issues (FB's colassal walled garden)

Brand Strength and Best In Class Customer Service (netflix)

-brands are built through customer experience -top in american customer satisfaction index -early entry and effective execution

Business Model Canvas: Channels

-describes how a company communicates and reaches their customer channels in order to deliver a value proposition -communication, distribution, and sales channels

more ad formats and payment schemes (google)

-display (image) ads: banners, small rectangular buttons -rich media ads: animation, encourage user interaction -interstitials: ads that run before a user is able to reach his or her intended destination -the internet advertising bureau sets standards for ads -cpm: cost per thousand impressions -cpc: cost per click -affiliate programs: vendors share a % of revenue with websites that direct customers to their site and make a purchase -amazon is the largest, 4-15% -they might become an exclusive advertiser on a site

The San Diego City Schools: Enterprise Resource Planning Return on Investment: Employee Benefits

-employee benefits required significant manual work and was really complex -benefits department was administered by consolidated ominous budget reconciliation act program

how to get SMART

-establish policies and expectations for employees: short, simple, and clear 1. representation 2. responsibility 3. respect -know consequences, but not micromanaged -training is necessary -employees should be aware of IS and threats -online engagement is a valuable asset: linked tx > adobe -know what works and what doesn't (expressive, but protect brand)

google trendspotting

-google lets you see aggregate trends in which users search for, which yields powerful insights -allow anyone to explore search trends; breaking out the analysis by region, date, and other criteria

ad networks and competitive advantage (google)

-google's ad network is distribution play -attracts more advertisers -attract content providers -which attracts more advertisers -network effects!!! -benefit from more advertisers = economies of scale and a better return on Investment

Rent The Runway: Subscription: Bigger Than Rental

-have expanded their offerings -flexible, but subscribers get more -began with testing -you can buy now as well -different lengths of rentals -benefits from relationships with customer -revenue from subscription grows 150% a year, growth margins 50%

a crowded field of rivals and other challenges (netflix)

-highly fragmented market -some say they take heads on FAANG stocks, but hard to say since they're all different (streaming is secondary to all except netflix) -startups are no threat, but firms that have success elsewhere could be -apple, google play, amazon prime, youtube red, hulu, verizon -buts based business is doomed if tech infrastructure is not solid -ISPs not always happy with netflix's bandwidth usage (net neutrality debate)

Rent the Runway: Customer Evaluation

-initially, college students, though this has evolved -customer lifetime value; predict future values, often use NPV, hard to predict future -many start with rental and move to subscription

Sources of data

-internal operation systems -external systems (can be purchased): social networking/user generated, transactional data

crowdsourcing (social media)

-jeff howe -taking the job of a designated employee to the crowd -goldcorp used crowdsourcing to find gold in their property.. 80% helped; grew from $100 mil to $9 bil -waze for navigation -9/10 top 10 brands use crowdsourcing: mcdonalds, netflix, threadless. com, marketocracy (75% return), waltham, MA (innocentive), amazon mechanical turk, topcoder, FB (languages), kaggle

Common Startup Mistakes

-lets price on cost (now strategy) -don't think internal economics... think customer insight -price on value (buyer's value, don't usually think this way.. convince them)

crowdsourcing and code contests: the netflix prize and beyond

-netflix prize: 1 million to team to improve cinetmatch ratings 10% -successful so they have done more

Customer experience, complexity, pricing, and brand strength (netflix)

-netflix: one pricing scheme -want to be your entertainment choice -customer satisfaction beating peers

Disruptive Technologies: Understanding Giant Killers and Tactics to Avoid Extinction: How it Works

-no physical aspect, no banks... transactions are recorded on the blockchain; transactions are performed by miners, who get incentive from earning more "tokens" (peer-produced), ledger records transactions, and no one can transfer the asset without the private key... stored in encrypted wallet -tech is open source and rock solid; decentralized, passwords are impossible to guess, verification- no currency spent in two places at once, not a single use block chain

How Big is R it Really?

-opportunity -market opportunity analysis -identify customer/market need -size the market -competitors -growth potential

How do I know if I have the right customers?

-pass/fail signals and experiments (test interests, how many?) -business to business -web/mobile; thousands (not dependent online)

understanding channels: from atoms to bits

-physical/physical: food, household goods, automobiles, planes -physical/bits: stocks, insurance, software -bits/physical: amazon, consumer electronics -virtual/virtual: facebook, google, twitter (greater customization)

SUM

-returns sum of a numeric column; SELECT SUM(column_name) FROM table_name WHERE condition

key aspects of rise of social media

-rise of social media has coincided w/ computing -peer producing and crowdsourcing are leveraged -collaborative consumption: share resources -"strategy and the interent" hypergrowth w/o $ and advertising (network effects)

Better Advertisement Everywhere, What's Up Next?

-social televison? has netflix ceo on his board

Scale from selection: The Long Tail (netflix)

-traditional: 3,000 DVDs -netflix: 125,000 titles -long-tail: large # of unavailable through conventional stores. selection attracts customers and the internet allows a larger selection inventory -eliminates geographic constraints

Design For Action: Track 5: Testing With Users

WHY: refine your solution and your understanding HOW: be aware of your prototype, the context in which you are testing, how you interact with the user, and how you observe and capture to get the right feedback Rules: -host -player: create prototype experience -observers procedure: 1. let your user experience the prototype 2. have them talk about their experience 3. actively observe 4. follow up questions

Design For Action: Track 3: Driven Prototyping

WHY: the goal of creating a prototype is to gain insight from how the user interacts; can reveal necessary features that you had not thought of HOW: set up a format for users to create something which leads you to understanding their thinking by asking the user to do something; generative, but apart from your biases

Tableau

↳tableau is a data visualization tool and a business intelligence standard ↳consists of dimensional and measurement fields ↳can filter by different things (for example, sort by sum of sales, sort by country, etc)

Big Skinny: A/B testing

-A/B testing used to organize website -alternative to guessing -needed software engineer to fully use A/B testing

Triangulate: Business Model and Key Metrics

-August 2010: receiving revenue from selling FB credits who want more matches -considering monthly subscription service -monthly expenses= $50,000 -Nagaraj created metrics dashboard in april 2010, focusing on user base, acquisition, quality, virality, and coin usage -getting people to come back is biggest hurdle

Online Marketing at Big Skinny

-CEO Kiril Alexandrov -issue with website; once thought 4000 people had ordered wallets through online store, but majority were given away for free through glitch -originally sold BigSkinny wallets at at Harvard street market, but online marketing was lacking

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: acquisition

-broaden product offerings and absorb threats -diapers.com -wag.com -zappos -alexa -audible

Market type and revenue

-hockey stick curve: represents their revenue over time... new market -most go out of business; influence tipping point, can't innovate alone -existing: already customers and competitors; take share from competitors is a straight line -resegmented: hybrid between straight line and hockey stick (okay substitute with value until tipping point... differentiate)

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: Amazon's distruptive consumer hardware businesses: kindle

-media represents 20% revenue (atoms to bits) -kindle is about putting the store in the user's hands; expanded because of data gatherting -1st kindle e-reader was black and white and $399, with moore's law, it became 69 -it had staggering success, but at the end of the holiday season from its second year, it was top-selling in dollar volume and unit sales... 20% of amazon customers own -kindle fire: had color, 1/2 the price of an ipad, annd 1st with cloud -goal is not money from kindle, it arrives linked to an amazon account... allows you to buy books etc ($136 operating income, it sells more books than physical)

google and youtube

-paid 1.65 billion for youtube when it was a 20 month ad start up; 2nd largest search engine -FB streams as much as youtube, insta, and etc -YT is moving to TV shows, music, movies -YT has bigger primetime audience than US's top 10 TV shows, according to Google -YT expensive to store content -YT suffers from exploiters, fake news, exploiters, etc; uses AI to remove over 8 million videos a month

why study google

-they are one of the most influential and impactful firms in the modern era -hard to find a place where no company is impacted by google

the hierarchy of engagement

-user engagement power products, and the best products take that fuel and propel forward -3 levels to evaluate non-transactional consumer companies 1. growing engaged users 2. retaining users 3. self- perpetuating -the farther up hierarchy, products become better; well positioned for growth

mobile apps and the challenge of google search

-users only spend 14% of their time on the web -specific apps can take away the need and use of google -deep liking allows ads to launch their app -mobile search ad revenue has gone dow, but actual ad revenue has increased

Big Skinny: Interactive Content

-wanted interactive engagements on website to enhance engagement on website -helped users select wallets for family and friends

Disintermediation, digital distribution, and gathering customer data (netflix)

1. don't need to share revenues to third parties 2. studio -> consumers means studios keep valuable data

Business intellgence systems

Use data created by other systems to provide reporting and analysis for orangizational decision making -the ability to pull data from across and organization and make decisions is key to business success

SQL aliases

give a table or column a temporary name 1. SELECT column_name AS alias_name FROM table_name 2. SELECT column_name FROM table_name AS alias_name

it's a multiscreen world: getting netlfix everywhere

-at first, only on PC -didn't want to have their own hardware -went with software platform to install a multitude of platforms -partnership helped with distribution -cable networks hesitant as well

creativity and the role of the leader: bringing the process to bear= carefully

-can creativity scale? -vibrant, ongoing collaboration and free idea flow tend to dry up as a business adds people and projects -how do you get lift out of adding layers instead of adding weight -the classic response to increased scale in an operation is increased reliance on process/ continually improving it -when organization focus on process improvement too much, it hampers long-term innovation -things can keep getting better, but you can still be doing the wrong thing you need to approach the problem from different angles 2. map the phases of creative work -process management is appropriate in some phases of creative work but not others -six stigma has destroyed innovation -the leader's job is to map out the stages of innovation and recognize the different processes, skill sets, and technology to suppot that each requires -to predict an idea is to have an idea -you must accept that the discovery in pharmaceutical innovation is inherently muddleheaded and models like the six sigma are working toward reducing variability and achieving greater conformance to a desirable norm -efficient models make good sense for the middle and end stages of the innovation process, when the game has moved from discovery to control and reliability -know where you are in the game -appreciate the different types of creative types among your people and realize that some are better are certain phases than others 3. manage the commercialization handoff -few people have equal capabilities in idea generation and idea commercialization -innovation reaches a point where it will best be served by people who know how to take it to market -projects often lose steam at the handoff to management -instead of trying to teach inventors to spot market opportunities for their discoveries, kayffman links postdoctoral scientists to commercializers -these opposing models highlight the tension that always exists in the management of creatives, whther to round out their individual skill sets or allow them to run with their unique strengths 4. provide paths through the bureaucracy -linked the life of an idea in a large corporation to a bill going before congress -these powerful constituencies inside the company collectively beat things in a shape that more closely conforms to the existing business model rather than the opportunity in to market -managers need to recognize what this process does to ideas and deliberately contain it -the need to create in a culture in which creativity can thrive, repeatedly returning to the image of a gardeners who prepares the creative soil and nurtures the seedlings of ideas 5. create a filtering mechanism -gardens do have weeds; managers must not only water and fertilize, but also kill off the stuff that holds no potential -at what point and by who should this determination be made? -a personal commitment to success and profession ramifications can be served -unless the people sitting in judgment represent a variety of disciplines, functions, and viewpoints, they are unlikely to make wise decisions your ideas in the real world either live or die out there as opposed to in committee, where death to creativity occurs

Challenges of Going Global: Low ARPUs, Legitimate Rivals, Unreachable Users

-crowdsourcing localization effort asked users to look at FB phrases and offer translations to their language -FB wannabes everywhere -less than 50 million people have global income enough to interest major advertisers -have to adjust to advertise in these areas, such as FB lite -z wants china

streaming and the data asset (netflix)

-data is what netflix has that rivals like cable do not have -more accurate recommendations, designs, determine appropriate amount of content, shape decisions -streaming data > DVD data... fed to collaborative filtering -75-80% of what users watch comes from recommendations -A/B testing to test features -content: use data to see if renewal worthy, use actor, genre, etc when determining if worth streaming -helps with original content decisons -80% success rates -even determines trailers that you see

Big Skinny: Online Distributors

-established sales presences on amazon, buy.com, and eBay -these retail platforms were robust and could handle a large volume of interactions

creativity and the role of the leader: pulling to together

-it sounded just like the kind of leadership we wearing our creative worker hats would appreciate having -how do you get a management layer made up of real humans will aspire to do it -the person bringing it all together is the producer; he or she must exercise leadership in highly ambiguous context, where there is no clear yardstick for how good the product is and there are no clear rules for who getd to control the output -these producers who operate at the center of the storm without being the focus of attention and are proactive with a diverse group of experts without being overcontrolling

How to grow customers, keep them, and attract even more

-key is customer relationships 1) physical channel (awareness -> interest -> consideration -> purchase) 2) web/mobile channel (acquire -> activate -> [keep] up-sell -> next sell -> cross sell -> referrals -create viral loops

secret of google'e successes

-pull instead of push advertising -pricing method (cpc instead of cpm) -selling strategy (auction instead of fixed prices) -at networks (adsense)

The San Diego City Schools: Enterprise Resource Planning Return on Investment: System Implementation

-shortest time was 10-24 months

Rent The Runaway: Entrepreneurs Expanding an Industry by Blending Tech with Fashion: Intro

-the founders, Jennifer Hyman and Jennifer Fleiss, recognized a disconnect in the fashion industry, in which a majority of designer brand's target market could not be reached because they could not afford the brand -now, the have over 9 million customers -their focus is on selection, convenience, and quality -"closet on the cloud"... rental and subscription based -rentals are 10% of the retail price -stylists are available for support -100% of inventory is overturned -complex business model help face-off competition -brand, scale, partnerships, data -often implement new features to improve experience and efficiency

LIKE

-used in WHERE clause to search for a specified pattern; -%,=0,1, multiple characters; _ single character -'a%' any values that start with a -'%a' any values that end with a -'% or %' or in any position -'_%r' r in second position SELECT column1, column2 FROM table_name WHERE column LIKE pattern

U.d advertising spending

tv is decreasing and digital is inncreasing -paid search dominates internet spending -2014 revenues= $66 billion

SELECT DISTINCT

used to return unique values in the result set. It filters out all duplicate values. Here, the result set lists each genre in the movies table exactly once. 1. SELECT DISTINCT specifies that the statement is going to be a query that returns unique values in the specified column(s) In access: SELECT COUNT(*) as DistinctCountries FROM (SELECT DISTINCT country FROM customers)

wildcard characters

used to substitute one or more character in a string; used with SQL like operator and together they find a specified pattern in a column % 0 or more characters _ single character [] single characters within brackets ^ any character not in brackets - range of characters

SQL joins

combines rows from two or more tables based on a related column between them 1. SELECT column_name INNER JOIN table2 ON table 1.columnn_name=table2.columnname; (all records matchig in both) 2. SELECT column_name FROM table1 LEFT JOIN table2 ON table1.column_name=table2.column_name; (all records Ieft, matchig right) 3. SELECT column_name FROM table 1 RIGHT JOIN table2 ON table1.column_name=table2.column_name; (all records right, matching left) 4. SELECT column_name FROM table1 FULL JOIN table1=table2 WHERE condition; (all records in left or right matching)

The San Diego City Schools: Enterprise Resource Planning Return on Investment: persistent risk

-ERP system had to be delivered effectively, but with school system, had been full on failures and partial failures

SELECT

-extracts data from a database -example: SELECT column1, column2 FROM table_name

Over engineering data

When a data model has so many variables that its solution will only work the specific data used to create it

SQL self-join

regular join, but table is joined with itself SELECT column_name FROM table1 T1, table2 T2 WHERE condition;

SQL BETWEEN

selects multiple values within given range SELECT column_name FROM table_name WHERE column_name BETWEEN value1 and value2

Design For Action: Track 6: Feedback Capture Grid

WHY: facilitate real life time capture of feedback and prototypes; used when presenter-critiquers is anticipated; give feedback about design team process, or capture a user's feedback about a prototype; grid is systematic HOW: 1. set off a blank page with four quadrants 2. draw a plus in the upper left quadrant, a delta, in the upper right, a question mark in the low left, and light bulb in the lower right -fill the four quadrants with your user's feedback accordingly

UPDATE

UPDATE table_name SET column1=value, column2=value, WHERE condition;

Understanding Mixtape: Discovering new insights via human engagement, Track 6: Short Story-and-capture

WHY? allows team members to come up to speed with what they saw and heard in the field, helps them draw more meaning from the experience (starts synthesis process), and lastly, capturing each detail of the fieldwork, you begin the space saturation process HOW? unpack observations, share user stories, team members use post-its to headline most important observations, which can be helpful in grouping and understanding what is going on with each user

step 5 of data mining

data evaluation: the data findings are compared to the business's objectives, determining if they should be shared across the entirety of the organization.

2. Business Model Canvas: Value Propositions

-building block describes the bundle of products or services that create value for a customer segment -the reason customers turn to one company over another; solves problem and fulfills need -examples: 1. newness: new offering 2. performance 3. customization 4. getting the job done 5. design 6. brand/status 7. proce 8. cost reduction 9. risk reduction 10. accessibility 11. convenience/usability

costs of netflix

-fixed costs were land, building a plant -marginal costs were associated with units produced, and for computers, it was effectively 0

The art of the minimum value product

-not a minimal product -based on interaction and understanding customer needs -"new market"; not features... time problem

why is facebook the current winner?

-they did not invent the web or social media, but their platform allows them to be the winners

Pricing your product

-what are my revenue streams? the strategy/the company/ uses to generate cash from each customer segment (multiple cs = multiple rs) -within my revenue streams, how to I price the product? -pricing tactics used to set price in each customer segment -answer questions: 1. what are customer's willing to pay for (guessing at first, but later know) 2. how do customers pay? 3. how much do customers pay?

partnerships are two way streets

-what defines a partner: shared economics (what they need you and why you need them; mutual success/failure, co-development/invention, common customers, BUT... not always the same size... you're a startup)

Types of analytics used in AI

1. Descriptive (least complex): simply describe what happened. Employed across all industries 2. Predictive: anticipate what will happen, mostly employed in data driven organizations as a key insight 3. Prescriptive (most complex): provide recommendations as to what to do, employed mostly in organizations that lead in industry predictive and prescriptive are machine learning focused

What is customer discovery

1. state hypothesis (out of building; customer need?) 2. test problem (out of building; is this a customer need) 3. test solution (do we have the solution?) 4. verify pivot (do we have high-value need? ready to start selling?) -answer is usually no; go through many times

Understanding the increase in ad spending (google)

3 factors drive online ad growth trends: 1. increased user time online 2. improved measurement accountability 3. targeting -people spend more time on their phones than watching tv -advertisers can reach consumers at work though the internet -it is hard to see what's working on other channels, but the internet enables you to see the number of impressions and allows you to more effectively target users

google and providing internet

-google providing fast, free internet access; supports O3B satellite network; project loon -purchased titan aerospace for solar powered internet access -fiber optic networks in Ghana and Uganda -projectFI in US offers low-cost mobile data plans -google is becoming its own mobile virtual network operator... vertically integrated -google lobbied government to make telecom carriers to be more open

Users, payers, and multi-sided markets

-google: who is the customer? anyone, but need software development -model is free, yet profitable -advertisers are paying -for each customer segment, different model (value proposition, customer relationships, also different revenue model)

will the users stick

-retention matters, especially long run -investors look for accuring benefits w/ engagement -mounting losses if the user leaves -accuring benefits: the more you use a product, the better It gets; company needs to use data to improve the product -mounting losses: the more you use the product, the more you lose by leaving; ex: more notes, more addicted -anonymity apps look good for level two

the sharing economy: introduction

-tech allows citizen suppliers (individual instead of suppliers) -microentrepreneurs -collaboratively consumed products; "sharing economy" -different in these services is that they're leveraged through the internet; much more efficient -some are smaller, some are enormous, like uber and airbnb -lyft backed by alphabet at gm (4.3 billion raised, valued at 1 bil) -44% us citizens have participated in the sharing economy, 22% sold goods -a lot of consolidation

Why get out of the building?

-test hypothesis of customer problem -look for insights: change canvas -start hypothesis, design experiments, execute tests, analyze data, determine if insights= hypothesize (WHY?)

Mobile Is Tougher, But Global Players have Big Mobile Platforms

-testing to see if it works is harder with mobile, since it is hard to test a small group -app updating is slower than web -we chat= competition: digital payments, shopping, gaming, banking, taxi-service

The San Diego City Schools: Enterprise Resource Planning Return on Investment: SCDS Human Resource ERP Module

-the current SCDS human resource management systems were mostly composed of 30 year old legacy systems that had only recently been migrated to the client/server platform; used MS access, powerbuilder, and other development tools -no integration among systems, no automatic sharing, so each area had to enter in their own data, often resulting in duplication of data -limited data transfers

Rent The Runway: Customer Engagement (Mobile and Social)

Mobile: biggest challenge is getting firms to try their experience -key to convenience -phone app is always accessible... 40% of traffic -however, changes/updates are easier on the web Social: mobile is a viral word of mouth machine -referrals mattter; 75% higher conversion rate from friend referrals -social proof: positive influence when someone finds out others are doing something -launched photo reviews on website since customers love sharing their looks (find women like me tool, post questions) -people are 200% more likely to rent if they see the photo -leveraged social media -blogs have 200% higher conversion rate than paid-media -90% new customers from word of mouth

null values

NULL values represent missing unknown data. By default, a table column can hold NULL values. If a column in a table is optional, we can insert a new record or update an existing record without adding a value to this column. This means that the field will be saved with a NULL value. SELECT LastName,FirstName,Address FROM Persons WHERE Address IS NULL SELECT LastName,FirstName,Address FROM Persons WHERE Address IS NOT NULL

The Hierarchy of Engagement: Growing Engaged Users

a lot of companies have growth early earn, but it comes from different forms -not about growth of users, but growth of users completing the core action -snap = snapping -twitter = tweeting -can be hard to predict if growth will be accompanied by a core action, but signs include -product with clear value exchange -compounding growth

Mobile Revenue on the Upswing (FB)

-1st Q 2016: 905 FB's MAU's visited w/ mobile devices -51% ad $ outside the U.S.

AVERAGE

-returns average value of a column SELECT AVG(column_name) FROM table_name WHERE condition

The Customer Development Process

1. "post" BMC to wall (hypothesis) 2. get out to talk/gain insight -CDP: discovery -> validation (product market fit) -> creation -> company building (pivot between discovery and validation) -first 2: search -second 2: execution

The San Diego City Schools: Enterprise Resource Planning Return on Investment: unique problems for school districts

1. limited budgets, especially on IT 2. under-qualified personnel 3. changing leadership

Design For Action: Track 4: Wizard-of-OZ prototyping

WHY: fake functionality (saves tine) HOW: determining what you want to test or explore; figure out how to for functionality ; leverage existing like twitter or skype and combine with your introduction

Understanding Mixtape: Discovering new insights via human engagement, Track 7: User Empathy Map

WHY: good design is grounded in understanding who you are designing for (empathy maps help synthesize observations) HOW: 1. unpack: create 4 quadrant layout, populate map by taking notes of these traits: say, do, feel, think (infer) 2. identify needs: we noted traits; verbs, not nouns 3. identify insights: grow from contradicting between user attribution; ask yourself why these occur, write them down

Design For Action: Track 1: Prototyping

WHAT: getting ideas from your head to physical form; allow you you to quickly and test different possibilities WHY: 1. empathy gaining: deepen understanding 2. exploration: build to think 3. testing: create prototypes 4. inspiration GOALS: -learn -solve-disagreements -start a conversation -fail quickly and cheaper -manage the solution building process

Design For Action: Track 1: Ideating

WHAT? ideate is the mode of design process when you generate radical design alternatives -"flaring" not "focus"; quantity of ideas, not quality is the much WHY? transition from identifying problems into exploring solutions -step beyond obvious; increase innovation potential -harness collective perspectives and strengths of teams -uncover unexpected areas and exploration -create volume and variety in innovation options -get obvious and go beyond -keep generating ideas and evaluating ideas separately

Design For Action: Track 4: Stoke

WHY: help teams loosen up and become mentally and physically active HOW: do an activity that get creatively and engagement going -category, category, die! -soundball: stand in a circle and throw and imaginary ball at each other -yes, let's: everyone stands around and someone makes imaginary offer

Big Skinny: other options

affiliate marketing and email marketing

step 6 of data mining

deployment: if the data is determined to be beneficial, if can be shared across the organization and used to implement the goals the business wanted to achieve.

Pivot: firing the plan, not people

pivot: customer discovery v customers validation -asks: what do you do when hypotheses don't equate reality -a change to business model components (iteration=minor, pivot=more, done by founders... keep up with constant speed)

Advertising and Social Networking: A Challenging Landscape but A Payoff (FB)

the bulk of their revenue is from advertising, which is extremely popular

Walmart leveraging data

-Walmart has created and leveraged data to achieve world class supply chain efficiencies (7th best in the country) and becoming the #1 retailer in the world. -responsible for 12% productivity gains in the economy -they developed their proprietary system Retail Link in 1991, which enables them to track records, inventory, reorder inventory, and schedule delivery efficiently -their inventory turnover ratio is equal to 11.11, which means they sell the equivalent to their entire inventory every 4-5 weeks -leading suppliers adopted RFID, which allows them to track inventory as it comes in, as well as boosting sales 20% since inventory was where it belonged -Scan + Go was developed to enable customers to scan items in their cart, offer coupons, and create a mobile shopping list -Data Cafe provides managers with central clearing house for data driven problem solving -leverage Hadoop based data to support data efforts through social media, as well as purchased Kosmix to deepen their social and big data expertise and Walmart Labs -they often share data with suppliers (17,000), which has increased sales by 19%; however, they usually keep their data usage and AI usage under wraps -they have implemented new technology in their stores, which increases in store pickup -however, there are problems too. They need to find new markets to boost profits, and be on the look out for being targeted by unions and criticism. Additionally, their suppliers face a catch-22 between accessing a large market, but also having to accent such low prices

Data mining tools

1. Rapid miner: open source software written by java. One of them best platforms for conducting predictive analysis , offering integrated environments for deep learning, text mining, and machine learning. Can be on premise or cloud based. Offers a balance of custom coding features with a user friendly interface 2. Orange: open source component based software written in python. Has painless pre-processing features and one of the best platforms for basic data mining analysis. User oriented approach, however limited data connections 3. Mahout: open source platform which focuses on unsupervised learning process. Creates machine learning algorithms for clustering, classification, and collaborative filtering. Users can implement their own algorithms, so requires more of a specialized background 4. Micro strategy: business intelligence and data analytics software that complements all data mining models. Wide array and gateway and drivers, so can connect to any enterprise resource and analyze its data. Analyzes in real time. By paring with data mining tools, advanced models can be created to be deployed across the organization and provide insights about the market

Reinforcement learning

an algorithm learns to perform a task by trying to maximize rewards received for its actions ↳ USE WHEN: there is a lot of training data, and you cannot clearly define the ideal end state ↳ the algorithm takes an action on the environment; it receives a reward if the action brings the machine close to maximizing total rewards. the algorithm maximizes the best series of actions by correcting itself over time

step 4 of data mining

data modeling: use mathematical models to find patterns in data

step 2 of data mining

data understanding: explore the properties of the data, ensuring that it will actually help you achieve said goals. data visualization tools (such as tableau) can be used

Neural networks

interconnected neural cells. With experience, networks can learn, as feedback strengthens or inhibits connections that produce certain results. Computer simulations of neural networks show analogous learning. computing systems that are inspired by, but not identical to, biological neural networks that constitute animal brains. Such systems "learn" to perform tasks by considering examples, generally without being programmed with task-specific rules. -referred to as the "black box", meaning the are relationships between the data, but these relationships are difficult to define in a traditional formula -appropriate data

wikis

- a wiki is anyone can edit directly within a web browser (hawaiin quick) -collaborative efforts; wide spread and proprietary -what you see is what you get; click on history to see past edits (roll back) -commercial or open source -starting wikis can be challenging, though snowball effects when user sees value, wikimasters filter content -the larger the community, the more accurate content (griefers update pages- sometimes with seconds) -firms like Disney Pixar, can use wikis to keep employees updated; sony uses for collaboration projects

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: fire tv

-$39 or $69 version w/ ultra HD: netflix, hulu, voice search -runs 180 apps versus 1200+ apps -apple tv, google chromecast, android tv= headstart, but no scale -70 . million homes have one

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: The Emperor of E-Commerce

-1st only books on internet; 400 sq foot warehouse in seattle -now: largest online retailer with 400+ warehouses worldwide -Logo: we carry everything from A to Z -HQ in seattle -market cap: walmart + best buy + nordstrom + macy's + kohls + jecpenny + sears - main ideals: selection, price, customer experience... these reinforce competitive advantage= brand = more customers = scale =vertical integration = saves -3rd part sellers = 2 sided network effects = competitive advantage -data!!! = each digital movement is logged; fine tue customer experience, demand predictions and squeeze costs

ad networks: distribution beyond search (google)

-20% ad revenue from firms google does not own -ad networks: different apps/websites run google's ad for a cut -HTML code -typically cost about .70 -adsense was first developed to target keyword on the website -65% top-200 ad-supported websites use adsense -$7 billion per year, consisting of 2 million publishers -bundles adsense w/ other content properties= google display network -rivals: yahoo and aol

The San Diego City Schools: Enterprise Resource Planning Return on Investment: background

-2002: 8th largest school district in US, second largest in California -142,000 students, 180 education facilities -25,000 employees -2001-01 operating budget was 1.1 billion -Alan Bersin: superintendent, had brought along reform to improve teaching and learning -reform efforts receiving national attention, but only 39% teachers reported respected in survey -one of main efforts was to modernize business practices -hired Veronica Fromann as head of business operations, who knew district lacked IT leadership

The San Diego City Schools: Enterprise Resource Planning Return on Investment: ERP and education industry

-2nd largest industry in the US, 800 billion in annual industry with 40% in k-12 -81 school districts in US served more than 50,000 students and only 2% went to IT spending -ERP allows to integrate and maintain different functions to reduce data fragmentation -systems were currently incompatible and hard to upload info -the ERP system would allow a common database of info so data could be shared by all; increased accountability and transparency -because of this, ERP was a difficult undertaking

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: Amazon game studios and video

-android, ios, fb, firestick, close to nintendo +, recruiting top talent -$300 mil for HBO titles on prime -$1 bil on video licensing -amazone studios -thursday night football -bought twitch gaming stream (SS unique viewers; less bandwidth than google, netflix, etc) -beef with hatchetee; delyayed shipments and people questioned their committment to customers -barnes and noble refuses amazon titles, target doesn't carry kindle -41% US book sales, 61% online sales -amazon: whole sale (has to agree to agency now bc of lawsuit; although apple still cahrgers 30% from in-app purchases, so not on kindle) -apple: agency

uber's wild ride: sharing economy success and lessons firm a fallen founder

-4 years in: 633 cities, 76 countries -2017: 7.5 billion . revenue, lost 4.5 billion -50,000 new jobs a month; background checks and new car models -SF: 74,191; NYC 90,766 (40 hours) -80% job satisfaction -says reducing DUIs and pollution; some studies say opposite -taking cab market and expanding -convenience and trust increase -stellar example of PMF -eliminates human dispatchers, capital cost of cars, auditing driver quality -main complaint: surge, pricing, but important for service -data drives growth growth and success; employees crunch #'s to optimize operations (weather, events, etc); "God view" map, heat maps; expands and attracts drivers -APIs on tripadvisor, opentable, united airlines; user health -network effects and moving early -some say worth less than valuation, some say next $100 billion company -doubling in size every 6 months -backed by google and amazon; competitors in some ways -regulation issues; fought san fran and took "cab" out of name; reclassify drivers and pay benefits and taxes; backed by consumers and protested by cabbies -partnership w/ cities like Boston -growing, but competition in china already had 90% of market; regulation strict; sold to didi after 1 billion in losses -ban in Dehli; growth of 40%; ola 3x as much operations -largest investor; softbank also invests in didi, grab, ola, 99

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: moving early and scale = data

-A/B testing is critical for shopping patterns and advertising -one of the best as far as completed check outs -A/B testing is key; buttons, colors, placements -can do several at once -P13N personalization effort allows for a/b testing, collaborative filtering, and cookies -35% of sales are from recommendations -advertising: 1.7 billion in 2017, feature rich, and good at retargeting -offer products offered by others next to their own (amazon marketplace) -amazonmarketplace= long tail with risks -2 sided market -can learn from sellers -afiiliate marketing program: others get $ for marketing and generating sales

Triangulate: B2B data-driven matching engine

-David Chen: July 2009, called to ask to work for triangulate -plugs-ins included captured URL history, page titles, number of tabs opened, and time stamps -decided to get data from APIs -Nagarij would always test out people to see if they worked; did equity arrangement -got meeting with VC firm; said call with eHarmony deal -spent some money on paid ads for "dating site" -150 individuals clicked on ads were sent to landing page based on behavior, 63% match -first-version of triangulate in october 2009 collected data using 8 live-stream connectors based on APIs -team began building version two of engine, which would consist of algorithm using historical data for successful couples

The San Diego City Schools: Enterprise Resource Planning Return on Investment: Hope for HR: An Integrated System

-HR would be costly, and needed school board approval, needed to present a compelling argument to secure approval -lay out ROI model

ACT I: David Becomes Goliath: Crafting Killer Assets for DVD-by-Mail Dominance (netflix)

-Hastings: biggest regret is taking firm public too early so people knew their info and how profitable they were -this drew in competitors like blockbuster and walmart, though netflix won out and continued to have revenue growth and profits -rivals could not match netflix's brand, scale, and data assets, all of which reinforced each other

Rent The Runway: Founding the Business: Are We on to Something?

-Hyman and Fleiss teamed up to explore product market fit, the idea that their solution would be one market actually wants and will grow -started with low tech tests, minimum viable product, a bare bones offering that allows a team to collect customers' feedback.... they wanted to know what works so they could pivot -one test was a popup rental shop, for college students -investigated willingness to pay online as well

Design For Action: Designing The Intervention

-I.D. began from studying users and creating a product brief, creating a design from there; at first, the different between the new and the old ways were not super different, but designers soon realized that regardless of how much designers understood users upfront, designs could not accurately predict user behavior, so, the solution was to go early to the user with a low-resolution prototype and slowly improve, step-by-step= ITERATING RAPID CYCLE PROTOTYPING -not only improve the product, but provided funding and organizational commitment (not fear if unknown) -in the old approach, winning was the exception not the rule; now, go to the executive in charge multiple times; becomes a gradual commitment (more interactions)

Design For Action: Conclusion

-Intercorp's success in boosting middle class peru depended on thoughful design of many artifacts: leading edge bank, innovative school system, and business adapted for frontier towns across Peru -equally important was mapping out steps necessary to engage all relevant parties -intervention is a multistep process consisting of many small steps; interactions w/ users is key

The San Diego City Schools: Enterprise Resource Planning Return on Investment

-Jeff Wiemann part of the team that would implement San Diego School's new ERP system -school system had requested finance, HR, and student information system software to "alleviate the district's overburdened and inadequate information technology" -they started with an HR system in order to mitigate risk since the HR system would require the least modification, giving IT practice -Wiemann knew HR expensive, but also vital. to meet the district's desire to meet new investment goals and operate more like a business to make more strategic decisions -the biggest challenge was building a business case that justified the cost of the HR system by showing the value of the return -the project was one of the largest investment's in the school's history, so he needed to build a really strong case by showing the benefits and risks

Big Skinny: A product that sells itself?

-Kiril started Big Skinny to offer world's thinnest wallets and solve 5 key problems men and women have with their wallets -proprietary material: tough and thin (nylon: water resistant and more durable) -fully packed with 40 cards and less than an inch thick -potential customers could see difference in thickness in person easier -customers were also big fans and would speak up for the wallet at street fairs -also sold wallets through retail stores and wholesale tradeshows -began billboard advertising, postcards, print ads in magazines

Rent the Runway: Operations and Logistics

-NJ distribution center is size of two football fields (largest dry cleaning facilities in the the U.S.) -new facility in Dallas -tech helps prioritize what needs to be done (screen tells staff if item needs to go right away) -color coding items -prioritizing shipping -cleaning: more people than tech -seamstresses for repair -so efficient, many items arriving at 8 am can be sent out by 12 hours -dresses typically turned 30 times -goals . of leaders is to try and expand infrastructure, not just profits; large center provides room for growth -privately held, but still profitable by EBITDA

Triangulate: Product

-Nagaraj felt more comfortable with busienss -33,00 users in california -wings pages were "home" "my ideal match" "my friends" and "my profile" -at first, they were having to look through a lot of pictures to weed out inappropriate content, more users helped -started focusing on product development efforts; boost engagement of users -pivotal tracker: A/B testing; estimates how many points worth of work can complete in a week... small changes over time, so need to be careful to keep track

Disruptive Technologies: Understanding Giant Killers and Tactics to Avoid Extinction: Examples

-Nasdap stock exchange -world's 45 leading financial institution formed R3CEV to build infrastructure -IBM's growing practice -Nationwide wants to build a blockchain -DuBeers, 1/3 world's diamond mining, development blockchain for customers to reduce fraud -spotify acquired mediachain labs to ensure musicians are paid -kodak developing to track rights for photographers -FB blockchain division by form head of paypal

The San Diego City Schools: Enterprise Resource Planning Return on Investment: the ERP Big Picture at SDCS

-SDCS management sought to develop an enterprise system that would modernize nearly every aspect of the district's business operations -the district staff had been planning to bring together an ERP system that would do this for more than 18 months, and then brought along Wiedmann, and Froman directed him to create a system that would achieve the district's long term goals -SIS system is holy grail of ERP systems, but most on the market only offered 50% of functionality required, which is why they decided to try HR first since he had hands on experience

triangulate

-Sunil Nagaraj: october 6th, 2010 -wings was a one-click, self-proclaimed dating service, automatically created user's profile based on behavior on social networking, like facebook -march 2010: triangulate had received 750,000 in funding -early october 2010, wings had 33,000 active users -top priorities at this point were to ensure that wings was a product users loved and validate the business model

streaming changes viewing habits and frees creative constraints (netflix)

-WWW. what you want, when you want, whatever screen -binge-watching -no constraints from TV -creative: no pre-determined time, no cliff-hangers, more complicated stories

ORDERBY

-ascending or descending (default ascending) 1. SELECT column1 FROM table_name ORDER BY column1, column ASC

The San Diego City Schools: Enterprise Resource Planning Return on Investment: ERP System Costs

-Wiemann estimated cost 15 million for small scale ERP to 125 million for full scale -software: software is usually less than 1/3 of implementation costs. licenses needed for operating systems, but for this not over 500,00- -implementation/consulting: IT resources for design, test, and building the system. SDCS would need outside consultants. at least for HR, wouldn't need to make too an insane amount of projects -infrastructure: hardware and network required to support the system. the cost depended on the number of users and locations that accessed the system. guessed $1mil for hardware, $10,000 a month for networking -training: 10% of total budget; depends on # of users being trained -ongoing costs: software maintenance and support fees, infrastructure maintenance and support fees, and on-site staff

What is the Business Model

-a business model describes the model of how an organization creates, delivers, and captures value -the starting of any meeting, discussion, workshop, etc is with understanding what a business model is; understanding the business model you're dealing with is the key to having these discussions and creating strategic alternatives -the 9 building blocks cover customers, offers, infrastructures, and financial viability; in the end, they fit together like a puzzle -the left side is about efficiency, and the right side is about value -KP to KA is production -VP is what your produce -CR to CS is customers

the battle unfolds (google)

-ad networks will mature.. they need new sources of revenue -holding of alphabet breaks growth oriented goals into small businesses... risky markets -brings in more revenue than microsoft, but taxed; microsoft has mature market too (no growth, so stock price doesn't move) -pushing stock forward requires billion dollar efforts (hard to find billion dollar markets) -semantic web: easier to automatically categorize info -google still has competition from innovation siri, alexa, bing, yahoo -google v apple in many markets -google's position outside of its core market is unclear -google's ad network has switching costs, but doesn't cut out competition -switching costs with search are incredibly low -although google has not committed any major anti-trust laws yet, FTC officials believe some monopolistic behavior... EU had accused as well

Big Skinny: Algorithmic Search

-algorithmic searches determines what is relevant to a user's query and which results comes up -ranking high in search is key for consumers seeing your product -even though site submit website, no guarantee that the site would appear in searches -retailers could create product feeds for listing products -google uses search engine optimization

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: Beyond Clicks: More Bricks

-being largest retailer is huge, though thinking of growth through physical retail -many students rent textbooks through amazon; amazon campus @ purdue 2015 -pickup, lower costs by 40%, purdue says $1mil in 1st two years -prime student -amazon books: not super efficient or selective... don't make much money -organized like website, only 4+ starts rating s -carry other amazon products, use to promote prime -pop up shops starting spring 2017 -amazon fresh: freshdirect copy -amazon has 45% of online market, though 90% of 3/4 trillion grocery business in person -click and collect in seattle -losing 2x as much to spoilage than most grocery retailers -grocery business: 4.5% in US -bought wholefoods for 13.7 billion (overlap with amazon customers) -sells amazon prime, 5-10% discount on groceries -in stores purchases= data -cloud based server farms fuel machine learning and could allow highly efficient stores

What do customers have to gain?

-benefits/desires through your company -gains;happy: money, effort, time -social: look good, status, increase power -expectations and beyond: quality, more of, less off -measure success v. failure: performance, cost -delight: features, performance, quality -increase adoption: cost, quality, fun, lower risk -easier: services, lower TCO, flatter learning curve

Taking Big Skinny Online

-bigskinny.net -site's goal was to enhance brand and propel online direct sales -wallets typically sold on impulse and value -how could the site cause buyers to buy the product online

Disruptive Technologies: Understanding Giant Killers and Tactics to Avoid Extinction: From Bit to Blockchain and Beyond, A Disruptive Innovation for Money and More?

-bitcoin: not backed by gold or government; appeared out of thin air, extremely volatile; cryptocurrency, digital currency, digital money -many think bitcoin could spend upend money as we know it -many focus on the blockchain, a highly-secure, decentralized, distributed transaction recording and verification mechanism... all sorts of uses, including stock sales and digital document signing; used by many governments, venture capitalist have invested 1.3 billion

sharing economy: Airbnb: hey stranger, why don't you stay at my my place

-brian chesky and joe gebbia and nathan belcharczyk -poster child for sharing economy -backed my sequoia capital; raised 4.4 billion, valued @ 31 billion, profit =100 mil= (2017) -guest to listing: 11 to 1 -200 mil have stayed w/ so far -81,000 cities, 192 countries -value on supply and demand side (large cut on both) -trust=essential; anonymity on airbnb; verified identities; hq photography -still negative incidents occur; trashed homes, weird incidents (24/7 phone service, payment guarantees) -use tech to monitor transactions and scame -data mining = 15 perabytes a day; 150+ data scientists -in some places, people are breaking the law; health/safety concerns; some legal issues, though many attempting to advocate -competition: homeaway, mariotte, tripadvisor, accor hotels, hyatt

Scale from the Distribution Network (netflix)

-built scale from 58 highly automated distribution centers -could deliver movies overnight to 97% of the population -scale reinforced their brand -located close to USPS facilities and pre-sort mail before dropping off at USPS

WHERE

-can be combined with AND, OR, and NOT operators -it filters records based on more than one condition examples: 1. SELECT column1, column2 FROM table_name WHERE condition; used to extract only records with a specified condition -operators include =, >, <>, <, <=, >=, like, in -can also be used with UPDATE and DELETE 2. SELECT column1 FROM table_name WHERE_NOT condition

Rent the Runaway: Data

-chief analytics officer was one of their first hires -data drives the value chain (imbound= demand patterns, operations=how long do materials last? how does this impact cost, etc?) -analytics help inform length of rental period , turnaround, time expectations, inventory needs, on time returns, locations of warehouses and stores, etc -web and mobile app deliver engagement -email campaigning, FB, twitter, etc, marketing campaigns -leveraged demand pricing -agile software development, continuous deployment

Netflix: intro

-co-founder Reed Hastings: had gone back and forth as far as people's viewpoint of him -netflix went from DVD to internet service -they poorly communicated this transition and attempted to split into two firms, Quikster and Netflix, which caused an exodus of a million customers, their share price collapse, and hastings had to call to redesign -however, they came back from this, and have since expanded their customer base abroad and been able to sponsor original content -DVD streaming allowed netflix to dominate the sector, and hasting knew that the internet was the firm's future; focused on costs, content, revenue, and rivals were all different

Disruptive Technologies: Understanding Giant Killers and Tactics to Avoid Extinction: Concerns

-consumer benefit needs to be stronger -difficult to understand and use -reputation problem -security concerns -unsure about regulation -volatility -increased transaction volume will help stabilize, but early bitcoin tech cannot handle (scalability problem) -lack of standards

Streaming business: everything is different (netflix)

-content availability -content acquisition costs -legal and regulatory environments -potential opportunities for revenue and expansion -potential partners -competitors and their motivation

content adjacency problem (google)

-contextual advertising based on keywordsL lucrative, but limited -can avoid adjacency issues by using keywords to not show -networks refine based on previous feedback

Triangulate: Pivoting

-continued to pitch start-ups to VCs and angles two to three times a week, hoping to raised 3mil to pursue B2B model -went to advisor Misiek Piskorski for advice... encouraged him to keep building dating site -reasoned facebook might provide window of opportunity for market entry -triangulate could create a dating application for facebook, using an API to integrate with social graph -users would be able to import info from profiles and other networking sites -"wingman" services -"our consumer dating site will leverage data" -revenue generated through subscription service (lifetime value of $X per user, acquisition cost of $X per user, X% conversion rate of paying users) -network effects are key -lower acquisition costs per user

creativity and the role of the leader

-creativity is at the heart of business, but not the top of the management agenda -creativity is essential for an entrepreneur -considered unmanageable -ignored because didn't produce immediate payoff -most work in creativity is available to any businessperson who steps back from the day to day and asks questions -now, creativity is an urgent concern for many -competition has turns into a game of who can generate the best and greatest number of ideas and creative scholars are being asked questions about their research -harvard business school convened for two days with business leaders from all around to make the business practices theory, and an agenda for new business leadership took place -there IS a role for management in the creative process, though it is different than traditional thoughts of mangement -managers don't manage creativity; they mange for creativity

google's ad network type of market

-cross sided market: the advertisers fo through google to ad network partners and consumers

Business Model Canvas: Cost Structure

-describes all costs incurred to operate a business model -creating and delivering value, maintaining customer relationships, and generating revenue incur costs -cost driven: focus on minimizing cost wherever possible (southwest) -value drive: focus on value creation; personalized service (luxury hotel) characteristics: -fixed costs -variable costs -economics of scale -economics of scope

Business Model Canvas: Key Partnerships

-describes network of supplies and partners than make the business model work -optimize business models, reduce risk, and acquire resources -4 types: strategic alliances between non-competitors, coopetition; strategic partnerships between competition, joint ventures, buyer-supplier relationships to assure reliable supplies 3 motivations: 1. optimization and economy of scale 2. reduction of risk and uncertainty 3. acquisition of particular resources and uncertainty

Business Model Canvas: Key Resources

-describes the most important assets required to make a business model work -can be physical, financial, intellectual, or human; they allow a business to create and offer a value proposition, reach markets, maintain relationships with customer segments, and earn revenues

Business Model Canvas: Key Activities

-describes the most important things a company must do to make its business model work -required to create value relationships and earn revenues categories: -production -problem solving -platform/network

Business Model Canvas Customer Relationships

-describes the types of relationships a company establishes with specific customer segments -clarify the type of relationship it wants with each segment -driven by customer acquisition, customer retention, boosting sales -examples: 1. personal assistance 2. dedicated personal assistance 3. self-service 4. automated services 5. communities 6. co-creation

Triangulate's Inception

-developed idea that began Triangulate while at Harvard, thought used of algorithm could fix -took independent student research project and explored if mining people's computer usage could be effective -reasoned patterns could be 100% effective and could "triangulate" -chose 1.2 billion dating industry because triangulate was about accurate information of people, and identify patterns of compatibility -initially experimented with downloadable application called rescuetime to capture computer's usage data (needed dataset, though didn't really test hypothesize or target consumers) -by spring of 2009, had recruited 100 volunteers, found harvard student to write code and analyze data -value of being on cloud -had blind faith

Big Skinny: Display Ads

-display advertising seemed similar to a booth at a street fair (intrigue users) -definitely favored in-person sales for comparison -display ads came with complications: users are conditioned to ignore display ads and required Big Skinny to determine where the ads would appear; approached ad networks on smaller ad networks, especially ad exchanges

google the search king

-google is the dominant search provider, for now -microsoft's bing is on the rise -mobile advertising is still young -can google continue its dominance?

strategic issues

-google leads in search/ads and offers unmatched network reach -strong brand: switching costs for search low -defeating google with some sort of technical advantage will be difficult, since web based innovation can often be quickly be imitated

other google projects

-google pay: near field communications; will be able to use to leverage advertising -YT and blogger successful, okurt successful in Brazil, Buzz and Wave failures, google + lacks loyalty; hangouts, photos are good tools (AI used); hard to dominate in this market -google apps are collaborative; google has cloud, pay service, google + lets you share with friends, google drive (people are reluctant to give up office, however, thousands of people use though schools -competition .w/ microsoft and amazon (cloud) -google leads in AI -google has massive potential across many scopes; studying infrastructure and changes in industry is essential

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: Warehouses

-early warehouses: inefficient; couldn't copy walmart and their individual items... focused on data, tech, and costs -fixed slow processing for pre-packaged -warehoused powered with as much homegrown code as their website; holidays can be 65 mil a day -some warehouses are 1/4 mile in length; people and robots... 1/2 sprint, 1/2 symphony -shipments are scanned, workers examine for defect, staffer flips light from green to red if issue, "problem solver" swoops in to fix -robots are being developed; brings shelves to workers... kivas carry 720+ lbs -shelvers scan products added to inventory, shelf is also scanned, software knows; soon whisked away by robot... different robots can process same items at same time -"random stow" no two similar products sit next to each other, which reduces mistakes -pickers wait for robots, monitor indicates where product goes, humans do a second quality check and electronically scan item -> placed on conveyer belt to be packaged -> software shows how to do so -SLAM: scan, label, apply, manifest; only takes about a second; weighted, labels printed, loaded separately on conveyor belt, sent to trucks (sometimes trucks leave as little as 3 minutes later) -using over 100,000 robots; can travel 3-4 mph... never collide bc of motion sensors; rechargable, repair staff on site -amazon is now 50% more efficient than time of acquisition (kiva: closer shelves, more product, Imbound inventory unloaded 30 minutes, fulfillment 15 minutes -spent 26 million on equipment, 46 million getting warehouses on board -key to customer experience

Disruptive Technologies: Understanding Giant Killers and Tactics to Avoid Extinction: Benefits

-easy and fast transfer w/o bank -cuts transfer free; micropayments -international commerce -no data miners in credit card company data; blockchain straddles transparency and privacy -decentralized -Deloitte University says more than money replacement; a standard for securely exchanging value and recording ownership over a network without an intermediary -network effects are necessary

blogs

-emerged a decade ago -tumblr hosting 415 million -long-tail phenomenon (remains discoverable) -trackbacks (citation links to o.p.) and blog rolls (shoot outs) -2 way dialogue: electronic bulletin board; comments keep blogger honest -can be great for businesses; immediate dist. of ideas, no limits; directly to the public; bill marriott, zapps tony hsien, virgin group's richard branson -can be supplemental content -downsides: comments can be spammy, poor business posts, hard to control employee blogging, posts live forever

Disruptive Technologies: Understanding Giant Killers and Tactics to Avoid Extinction: Don't Fly Blind, Improve Your Radar

-firms must pay attention to the trajectory of emerging fast/cheap tech; increase conversations, utilize researchers, rotate staff for idea sharing, pay attention -many disruptive tech startups come from big firm ex-employees

Content acquisition: escalating costs, limited availability, and the "long enough tail"

-first sale doctrine: good for DVD service, not streaming -6 firms control 90% of US media consumed; increased bidders and have to buy directly from suppliers -cost of streaming rose 43x in 5 years -different licenses and costs -windowing: limited amount of time -some premium channels have exclusive access to movies -some simply refuse to provide content -long tail to long-enough tail

selling strategy

-fixed price (price as labeled, one price for one product/service) -auction: specify the maximum cpc they are willing to pay, the rank of ads based on both maximum cpc and the quality of advertisers' web pages -actually pay just one cent more than the second highest bid

google=miracle

-founded september 4th, 2018 -IPO august 9th, 2004 -market cap 04/2015: $375 billion -sergey brin and larry page are both billionaires -their mission is to organize the world's information and make it universally accessible and useful -unofficial motto is don't be evil

Rent the Runway: Conclusion

-founders represent a new generation of entrepreneurs leveraging tech to recast consumer relationships -mobile + social + big data + tech operations + collaborative consumption came together -raised $210 million in capital -valued $800 million -women entrepreneurs growing, but underrepresented in tech industry, founders of rent the runway want to change this (host talks, stories of women entrepreneurs posted on their website, and they partnered with UBS for project entrepreneur)

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: Amazon prime

-free 2 day shipping -videos and songs -whole foods disocount -free same day delivery -100 million customers; 1/2 US homes -shipping costs: 600 million (25%), upped prime membership

IP address and geotargeting (google)

-geotargeting occurs when a computer system identifies a user's physical location to deliver tailored content -users geotagged based on their ip address, since they're public -ip addresses change depending on where ad how you're connected -proxy hide addresses -use wifi or GPS

Experiment Mixtape: Advancing Your Solution Via Prototyping: Intro

-getting the team to stop just discussing and start building, as well as help you engage in a different way 1. prep and schedule 2. launch: 15 min, warm up and get building 3. prototype: 90 minutes, get building 4. debrief: 45 min, recap the feedback, and plan next steps 6. going forward

the march to global dominance (netflix)

-global expansion is challenging, but netflix is still biggest player, having pull in 190 countries -entering new markets is expensive upfront -different laws -now, international profit and make up over half use base

Triangulate: Matching Method

-had deprioritized matching . method by august 2010 -more practical matching engine, Nagaraj still believes in his original matching engine

exclusives and original content (netflix)

-have exclusive rights to some shows -don't rely on partners; original offerings have won 2 oscars, 4 golden globes, and emmys -costs of orignal content are 7.5 billion -however, continue to out-do other streaming services with 125 million sub -seeking complete ownership with many of these, because as of now, many original titles are not owned by netflix; value netflix has showed will help -netflix "wants to be disney"

most valuable global brands

-in 2006, apple ranked 49th and google ranked 24th -in 2014, google was first, and apple was 2nd

the sharing economy: boom times and looming challenges in the sharing economy

-influential factors: world-wide recession, technology, environmental benefits, convenience -2 sided markets; network effects... scale, brand, resources to expand competitive advantage -need to understand value to participate -need suppliers -uber ahead of lyft -save on capital investment, utilities, and maintenance -even though peer to peer, inventory and organization necessary -highly-fragmented markets: lots of suppliers = high search costs... get between suppliers and customers/extend value chain (angie's list); can be more efficient (drizly alcohol delivery, classpass: fitness, 150 mil revenue, 400 mil value (47% of because of sp) -happier and more trusting customers -crowdsourced ratings (can be biased) -neighborhood rentals: hard to manage; maintenance supply of goods at fair price) needs a lot of producers and consumers... needs strong value proposition to overcome setbacks -neigh goods and snapgoods: producer > suppliers -trust in sharing economy, but still issues; how does insurance work? is more regulation needed? -local firms lobby for taxes and fee for sharing economy firms (airbnb concerns, ride-sharing concerns) -will their workers continue to be considered independent contractors? don't receive as many benefits.... low tax revenues (instacart did, homejoy shot down)... could raise wages 20%; lawsuits and opposition -boston supports; works w/ uber

The Big Customer Based: Delivering True Economies of Scale

-infrastructure: 300 million; anyone can buy, but it won't be easy to maintain -size of firm's customer base -thanks to economies of scale, bigger firms have a greater chance to maintain operations (netflix 23 million vs blockbuster's 1.3 million subscribers)

Design For Action: Designing a New Peru

-intercorp group is one of Peru's biggest corporations; CEO Carlos Rodriguez Pastor JRr. (owns 30+ companies) -with newly renamed interbank, he wanted to rebuild Peru's middle class 1. seeking culture of innovation: make bank competitive; tried to learn from US banks, but learning about himself would not be enough... needed managers; diversified company, 55,000 employed and $5billion revenues -started educating team -launched urban innovation team with IDEO 2. From Wallets to Hearts and Minds -Peru lacking in educating middle class; started mapping design processes; preparing for hesitation to a large business heading up education 1. reward fro "teacher who leaves footprint" 2. purchased a small school business, San Felipe Neri; would soon learn to re-engineer model... needed highly skilled teachers. IDEO created a new model, Innnova Schools 3. launched a 6 month human design process; engaging needs of students, teachers, parents, and stakeholders; resulted in tech integrated model, but risk disdain from parents and teachers, so they tested it out... people loved it, word spread, and it was integrated. teachers wanted to work at an innovative school

ways to get SMART

-interior and exterior -embassy: establishing a presence on several platform with same name, some even establish communities -be prepared for users responding poorly -only do so if you have excellent product and service -effective social media: megaphone, magnet, monitoring, mediation -engaging with customers makes them happy, but know who you're engaging with -respond to feedback (sometimes) -engage w/ influencers (sometimes) -engage w/ influencers -don't respond to just anything -have a game plan in case... silence can be deadly -respond to negative reviews on yelp and tripadvisor

profiling and privacy (google)

-internet based ads user 3rd party based cookies -forrester: 1/3 sales had mobile involved at some point -google partnering w/ Axciom (tracks customer email addresses); cookies verify EA, user goes to store and buys something, vendor can match them to that purchase (anon) -google's android firms can connect 15-20% of customers to mobile accounts, fewer IPhones; for some, enough ROI on spending (homedepot, target) -negative backlash: us federal trade commission says more transparency and user control is needed; the user needs to opt in -Google ad preferences manager: remove categories, interests -tech avoids targeting sensitive issues -some firms use a plug-in to ensure ability to use cookies even if user opts out

busting the bad guys (google)

-ip makes bad guys easy to spot -click farms spread fraud across dozens of ip addresses -botnets: hackers exploit security holes, spread viruses, or use phishing to trick users into installing a software that is dormant w/o central location commands, which use bots to generate traffic -the activity an ad network can monitor, the easier can detect fraud -social influence fraud: pepsi, coke, mercedes, paris hilton, 50 cent, us state department -software networks can stop all types of fraud, even more sophisticated ones -google's scale = more info... power drill can process 1/2 trillion cells of data in less than 5 seconds; detects fraud, finding patterns... turing tests emerges

Triangulate: Facebook Application

-january 2010 launched wings, targeting california users -used data from people's pages to match -used Bayesian modeling -badges, like social butterfly -coins rewarded to users to communicate -FB took 30% wings revenue -used data to assess accuracy of predicting based on profile -didn't invest time to refine matching engine -with wings launched, nagaraj approached VCs again for 750,000; said matching engine didn't matter -signed a term sheet february 2010for 750,000 from trinity ventures -hired Jason Radovan in april 2010 as web designer, Nadarak felt site improved

Inside Rent the Runway Warehouse

-library for designer clothes -100,000s of clothes, shoes, accessories -located in New Jersey, called "dream fulfillment center" -450,00 + items to check out 3 categories: 1. cleaning: sanitation; brought by truckload. separated and steamed, bagged and shipped 2. repairing: expert seamstresses 3. renting: online, app, in store 3 ways to rent: 1. check out an item for up to 8 days .2. choose 4 items for $90 a month 3. gain access to unlimited for $160 a month

Google and android

-manufacturers don't have to write their own software, cost of devices will go downn -ads in operating system, and it is by far the most popular OS -80% smartphone share, largest tablet share, exploring wearable, auto, chromecast, android TV, android apps; 60% education sales, purchased next, google cardboard = stronger network effects -challenges= fragmented platform, amazon alters android so much most android apps don't run

A Conversation With Jenny Hyman

-massive logistics company (traditional outbound, r+r inbound) -they can't be automated: vertically integrated, especially dry-cleaning warehouse 1. receive Inventory 2. repair [both have 0 day turn-around] -thought tech and logistics: 90% employees, core of business and barrier to entry would be outsource; a good leader is flexible -distribution partnership w/ we-work: less-around (120 days a year -> 150) -think of clothing as investment for impulse buys -rtr is more sustainable -rtr has scaled: "responsibility about more than making money" -treat employees equally (salary v hourly) -best ideas come from people who have experienced problems -diversify: races, gender, economic, ethnicity -if I don't speak up, who will? -more should speak up about what matters to them -sexual harassment: all raise and others raising awareness about issues... plurality of voices... diversify! -Kristen Green: prove herself vs guys don't have to -40% drop out after child -advice: go fot -confidence: worst possible scenario is not that bad -be aggressive

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: from personal storage to AWS

-media -> bits= consumer based offerings -alexa and firetv: thin computing... bulk computing happening elsewhere -AWS: amazon web services allow firms to rent industrial strength computing capacity at a need basis -EC2: elastic computing cloud -S3: simple storage serice -O.S., AI, security, programming, database -amazon workspace: fully functioning virtual windows pcs thru cloud -amazon app stream for gamers -beat out IBM and HP -helped insta; rivals netflix rely -NASA, NYT, ESPN use -reliant, but not fool proof -introuduced 2000... beat google and microsoft -17.5 billion profits 2017 -more than 18x capacity of next 14 combined -goal was monetize expertise in scalability and reliability -2005-2013: 12 billion on infrastructure

social networks

-most dominant and FB and linkedin -FB users > twitter and snap -insta ad revenue: $7 billion, 98/100 FB advertisers choose insta -pinterest: traffic nearly passed Fb, referrals 3x more likely -social media= customer engagement platforms -feeds are viral; low cost promotion... encourage user interaction; though, controversial -FB targeted college students, but linkedin went to business right away; great for networking, learning, recruiting, advice -corporations employ social media for their own needs -deloitte, goldman sachs, dow chemical use social medai for alumni; recruiting former employeers, leads -IBM: 42% work elsewhere, so SM important; sharing, expertise; use select minds and liveworld -slack emerging: the app that may kill email

Why is customer development is done by founders?

-need visions -if people say no.... employee was wrong -you can change strategy (product change, make pivots, hear feedback firsthand)

so, what's it take to run this thing (netflix)

-netflix's infrastructure runs on amazon's servers -using amazon's web services allows the firm to grow millions of customers w/p adding data capacity -biggest AWS customer: 125 million hours of TV shows and movies, collaborative filtering, diigtal rights manngement, video encoding, etc -12 PB of data per day -when a firm uploads a file to netflix, the firm uses amazon's cloud to configure 100+ copies -netlfix uses a complexity-based encoding to shrink files based on content type -runs nightly tests to determine what titles will be most popular the next day -could use content delivery networks, but relying on special purpose servers around the globe allows firm to operate at fast pace annd low cost (colocation facilities) -built most software from scratch; localization specialists make sure content designed well for all interfaces -netflix shares tools they think useful for cloud

Disruptive Technologies: Understanding Giant Killers and Tactics to Avoid Extinction: Potential Disruptor Spotted, Now What?

-no guarantee a disruptor will be dominant, however, could shrink corporate revenues, lower profits, and incur losses -some tactics include: 1. build a portfolio of options of emerging technologies... invest, consider acquiring; keep separate from partner firm (autonomy) -easy to mention, hard to do (FB w/ mobile-$$$ on insta and WhatsApp) -can't acquire yourself out of missing the market (AT&T) -sometimes, there is no answer (newspaper and internet) -think of Porter's 5 Forces Model

SQL structure

-not case sensitive; semicolon used to separate SQL statements -commands, accompanied by SELECT extracting data from the database allows you to perform the query

Google: Introduction

-often considered a one-trick pony, as 86% of its revenue comes from alphabet -they spend more on advertising than US -a lot of their services, such as google docs, are free -their $ allows for innovation, especially as the open source operating software android tops the market and extends google's platform -google is more focused on AI more than mobile -"moonshot" projects are risky projects -focus on acquisitions -alphabet non-ad revenue is 187th on fortune 500 -Larry Page announced a re-organization of various businesses recently -though not all efforts are successful, their revenue keeps them moving forward -they control 1/3 of all time advertising $'s -their market cap (the value of the firm, found by multiplying the share price times the number of shares) made them most valuable media company on the planet -5 years in: M> disney + news corp + time warner + viacom + cbs + nyt

twitter and the rise of microblogging

-originally a side project to podcast startup odeo -330 million MAUs -500 million tweets a day, 80% smartphones -more like FB, but asymmetrical following -the pulse of the planet -hashtags created by users -corporations and orgs use twitter for time-sensitive info, scheduling, communication, etc (starbucks, some use for sales from amazon, dominos, chipotle, dell getting customer feedback, and fundraising) -posts from algorithms; has allowed more virality -corporations know to tweet during live TV -limited advertising w/ emoji hashtags (1 mil+, but 420% up benefits) -bulk of business is advertising: promoted tweets (use CRM for targeting; charge w/ action) -mopub connects w/ iAd, google's admob, FB audience -developed and sold SDK to google for software management -video is biggest $$.. makes half of their revenue; promoted video and amplify (automated)... 30% to rest of company -tried APIs, but people weren't visiting twitter (free ride problem); purchased app, bought tweetdeck, launch bit.ly service, blocked users from other apps from seeing -"fake news" 1.4 users interacted w/ Russian government, 50,000 bots during 2016 election -struggle to balance free speech and abuse -AI used to manage abuse and bots -November 2013: IPO, 73% above price, stock fell after (growth is slow compared to others and features are hard for new users)

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud

-people predicted "Amazon.toast", "Amazon.bomb" -went 7 years without turning profit, losing $3 billion -stock price: 100 - 6 -Bezos: expanded warehouses, built up ecommerce (grew one of the largest cloud computing platforms), lead pack in e-readers, launched video streaming/gaming devices, defined new category w/ echo -shipping and logistics: primeair, hundreds of tractor trailers, trans-oceanic shipping, expanding local, drones -amazonGo, amazon books, acquired wholefoods (13.7 billion) -firephone= disaster -first profit: after kmart went bankrupt 3 years after kindle introduced: 4.76 million -> 1.15 billion (meanwhile, B + N, 150 to 37 to red zone) -future: unstoppable monster of the tech industry vs charitable organization run by investment community -April 2018: 1.6 billion profits, 51 billion sales -AWS: growing faster; 54 billion -amazon is the world's most valuable retailer and largest provider for cloud computing (sales are a fraction of walmart's, but worth 2x as much)

traditional display ads

-pop up ads -banner ads -mass advertising problems: unmatched audience and a way of pushing -context based targeting marketing can eliminate

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: the lord of logistics

-primeair, trailers, trans-atlantic shipping -cost and control drive logistics -spent 21.7 billion shipping costs 2017 -enhances customer experience -offers delivery for other businesses -uses data to track costs, fuel prices, weather, traffic, etc -long and mid-haul trucking: thousands of their own trucks -40 boeing 767-3000; low traffic areas, light loads... use software to determine if efficient -the last mile: 70 sortation facilities... 44% population within 20 miles -amazonkey: inside delivery w/ ring -onstar for car delivery -sea and global reach: china shipping; bulk up fulfillment centers and control more of value chain -stake in 2 european logistics companies (6700 tuckers, 170 million shipments... UK and France) -drones: goal is delivery in 30 minutes or less (regulation is challenge)

Leveraging the Data Asset: Collaborative filtering and beyond (netflix)

-proprietary system: cinematch= software tech... collaborative filtering; use data to personalize customer experience -recommended titles make up 60% of content -data is a switching cost -churn rate fell with rival entrance -cinematch operational advantages: inventory availability... recommends in stock movies -revenue sharing systems w/ studios

Triangulate: Marketing

-purchased wings ads on facebook and ad networks -33,000 users: 30% had been acquired through word of mouth -cpc of .50 -effective cost per acquistion (eCPA): $3 to $6 -quality of users from ads was lower (used less) -Nagaraj disappointed that users mostly in california

What do customers get from you?

-rank customer segments; crucial v trivial, frequency of occurrence, constraints or limitations context -day in the life (know where your product fits) -going out is how you know

Business Model Canvas: Revenue Streams~

-represents the cash a company generates from each customer segment (revenue - costs) -arteries -successfully answering what value each customer is willing to pay -2 types: transaction revenues and rewiring revenues -ways to generate: 1. asset sale 2. usage fee 3. subscription fee 4.lending/renting/leasing 5. licensinng 6. brokerage fee 7. advertising pricing mechanisms:: ~fixed "menu" pricing: price is a pre-defined (list, price, product feature dependent, customer segment dependent, volume dependent) -dynamic: prices change based on market conditions (negotiation, yield management, real time-market, auctions)

Hasting's letter to himself (netflix)

-responsible people who are self-motivated and disciplined -don't shell out stock options -honesty and maturity culture enables freedom with vacation -employees among highest paid (add 10% to peer's payment, severance) -no cutthroat behavior -like a team

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: scale and competition

-scale = negotiating power... low prices, longer payment terms -smaller than walmart, though more efficient (11,500 stores,485 billion vs 400 stores and 136 billion) -"a smaller firm picking a fight with amazon will not survive" -amazon is cheaper than e-commerce and walmart, and amazon has inventory through dynamic pricing -amazon can compete with whole foods ($ back to firm = upward 3 pillars) -scale has allowed creation of their own brands... cuts costs, reinforces power from suppliers -amazon business: 7.2 trillion (features targeted at business customers; multipe accounts, negotiate bulk accounts) -1st choice in tons of categories -relentless customer focus: every employee spends two days on the phone with a customer, pioneered 1 click ordering and recommendations, user customer reviews

The San Diego City Schools: Enterprise Resource Planning Return on Investment: A quiet revolution

-schools had not been capitalizing on ERP benefits, but SAP, PeopleSoft, and Oracle had been tailoring products to education market, so products were geared towards budgeting, reporting, and student information systems -most importantly, prices dropped, and products were discounted for schools -more interest in ERP systems -new focus on education with top line consulting firms like Deloitte, Price Waterhouse Coopers, and Accenture, maturing their practices to understand unique challenges confronting schools

The San Diego City Schools: Enterprise Resource Planning Return on Investment: Risk

-scope creep -wide-scale deployment -issue escalation -organizational change -negotiations -IT experience -software modifications -also possibility of potential benefits not achieved

search advertising (google)

-search engine marketing: the process of running and advertising search engine ad campaigns -2/3 google's ad revenues come from its own site -at first, there were no ads, but two years after INC, they changed their minds -keyword advertising is based on the user's search -ads show up when users have more purchasing intent -google has developed a formula ranking 1. maximum cpc an advertiser will pay 2. the ad's quality score 3. impact of extensions 4. ad format -ads are intentionally vague -the quality score: ctr, overall history of performance, relevance of ad to user's search, ad assessment of user experience of landing page -google provides to tools to help with relevancy, and offers automatically generated dynamic search ads based on your website -google offers discounts for minimum necessary to maintain ads position on page how much: usually .30 to 1.00

Why netflix?

-see how firms use tech to reinforce competitive advantage -successfully shifted their business model -see how tech helped: brand, scale, and switching costs and able to compare with rivals -How netlfix works: mail rental service, aware they would transition, and subscription for available titles

Design For Action: Spreading the Wealth

-since expanding to middle-class, job creation was necessary... expanded supermarket chain by purchasing Royal Ahold, renaming supermercados peruanos (102 stores nationwide today) -could hurt local businesses and farmers; knew needed to stimulate local production with engagement with local businesses: Peru Pasion program helped upgrade local farmers to supply S.P.

SQL in

-specify multiple values in a WHERE clause -shorthand of multiple OR conditions SELECT column_name FROM table_name WHERE column_name IN (value1, value2)

No business plan survives first contact with customers

-start with operating plan and financial model (the goal was to simply do right plan) -start-ups are unpredictable -5-YR plan -need to plan before you write the real business plan -plan comes from business model canvas (organizes thinking, turn hypothesis to focus on update canvas)

streaming and scale advantages (netflix)

-studios might charge more for deals involving more users, and the firm with more users is able to pay more -size based advantages: come from scale of a firm's streaming library and scale of virtuous base; virtuous cycle: more titles, more customers -price advantages from size -growth = vital competitive asset -aware of scale advantages

Triangulate: Optimizing ad Marketing

-summer and fall of 2010, wings continued to refine marketing campaigns; put less emphasis on wingmen -july and august 2010: wings user base grew 3500 - 17500 -hired public relations manager -FB added wings to application dashboard, more users -started to listen to consumers, used A/B testing and saw people wanted big photos -july 2010: surveys from top 200 users, paid testing service -added feedback button to wings -august 2010: wound down wingmen feature

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: alexa

-taps in to cloud; tons of features... little rocky, but improvement with more usage -category lead against google and apple -echodot, kids, echolook, and echoshow -extending to toyota, mercedes, BMW, ford, hyundai -alexa for business -machine learning -amazon wants to make ordering easier through alexa, the kindle, dash buttons, etc

customer profiling and behavorial targeting

-targeting: using ip address, type of browser, computer type, operating systems, and cookies (a line of identifying text, assigned and revealed by a given web browser and stored by your browser) to better identify customers based on certain attributes -IP addresses: geolocation, employer, university; ibm uses to recruit... 5-30% click rates, average is down 1% -Double click: personal ads to modem media -tech ads: computer, browser, or operating system -apple -> windows users -cookies: highest degree of personalization and targeting (tags your browser and stores it, builds a profile from your activities; matches to advertise) -remarketing (retargeting): site operators tag users according to which page they visit on their site -cookies: remember me feature... net person can log in as you -cookies can't be read if not given to website -3rd party cookies: websites you haven't visited -don't 100% work (users are assigned different IP addresses often, cookies can get mixed up, and users have different browsers and/or computers)

creativity and the role of the leader: drawing on the right minds

-the first priority of leadership is the engage the right people, at the right times, to the right degree is creative work 1. tap ideas from all ranks: google has found that ideas that have been executed without being directly backed by the founders have done beteer. secondlife claims that greatest successes come from workers own initiatives. depending too much on a single founder will be detrimental to the firm. 2. encourage and enable collaboration: innovation today draw many contributions. the most creative organizations are not centralized and top down. contributing to an independent network is its own reward and motivates innovation. organizations fail to fake full advantage of internet technologies to tap into the creativity of many smart people working on the same problem. the use of coordination totems to promote innovation. development teams use not only protoypes, but also metaphors, analogies, and stories to encourage their thinking. if the field was more level, more people would be willing to speak up. missions should be getting people to shut up at the right time, 3. open the organization to diverse perspectives: innovation is more likely when people of different disciplines, backgrounds, and areas of expertise share their thinking; diversity enhances creativity. social identities often have distinct knowledge associated with them, and to the extent an individual is comfortable integrating multiple identities, his or her knowledge sets can combine productively. if managers cause people to suppress their identity, they will suppress creativity. managers can enhance diversity by looking outside the organization for sources of creativity. innovative thinkers are not motivates by the desire to get rich but by the technical challenges and quest of the human fight. open source innovation, with the ability to tap into the passion and integrity of tinkerers, offers enormous potential for creative output, but they only work in certain endeavors and limited windows of time

creativity and the role of the leader: marrying research to practive

-the group parted with a sense that . theory and practice would increasingly come together to advance the understanding of creativity in business -three conditions seemed to be necessary for novelty: slack, hubris, and optimism; these . suggest mechanisms that organizations could employ -slack in an organization setting means sufficient time and resources for exploration -increasing hubris means inspiring managers -optimism takes hold when a vision of something truly different is made to seem more promising than the status quo

The San Diego City Schools: Enterprise Resource Planning Return on Investment: The Ideal ERP system

-the ideal school district ERP system was one where the full power of the ERP system was used to face the challenges schools systems faced -finance, HR, procurement, delivery, food services, and student tracking functions to be seamlessly integrated

ACT II: netflix and the shift from mailing atoms to streaming bits

-the shift from atoms to bits is happening in almost every industry (newspaper, apple/spotify, ebooks, videogames) -DVD service: 2010; 19.5 subs, 2018: 16.5 -they closed 20 distribution services -DVDs were still $56 million on $99 million revenue in first 3 months of 2018

Disruptive Technologies: Understanding Giant Killers and Tactics to Avoid Extinction: The Characteristics of Disruptive Technologies

-the term stems from the way that so many technologies create market shocks and catalyze growth; their main characteristics are: 1) come to market with a set of performance attributes that existing customers don't value 2) over time, performance attributes improve to the point they invade established markets -these attributes bring along many innovations, such as the shift from analog to digital

Business Model Canvas: Customer Segments

-these segments define the different groups is an organization that they are buying to solve -customers are the heart of the business model -grouping them by common needs etc will be satisfying them (different needs, distribution channels, relationships, willingness to pay) -decide which to serve and which to ignore -types of customer segment examples: 1. mass market: focus on one group of customers with similar problems (don't distinguish); example is consumer electronics 2. niche market: cater to specialized customer segments; supplier relationships 3. segmented: focus on segments with different needs and problems (medical, watch, and automation) 4. diversified: serves two unrelated customer segments with various needs and problems (amazon) 5. multi-sided and . platform: two or more interdependent customer segments (credit card company users and merchants)

Disruptive Technologies: Understanding Giant Killers and Tactics to Avoid Extinction: Why Big Firms Fail

-they focus on the bottom line/listen to their customers -disrupting tech initially has poor margins/poor performance... not attractive to the firm or their customers -early customers of disruptive tech are part of a different tech are part of a different "value network"... care about different features (free > quality) -over time, disruptive tech improves, and incumbent customers join in

Rent The Runway: Customer Engagement (Physical)

-they needed a physical aspect, so they introduced physical stores -high value customer engagement; appointment typically -can use data for the future -can replace fast fashion with last minute -rolled out cautiously; partnered with Henri Bendil -several locations, pop-up shops, and stores inside larger ones -patterns differ by location, but can still use data from engagement -helps with inventory

Rent The Runway: Customers like it- but what about suppliers? Growing customer base and creating a win-win

-they needed to buy from other designers in bulk, at a discount, which was risky with their inexperience and the possibility of brand dilution -however, helps brands by giving each their own page and educating users about the brand and its quality -highly trained stylists -tech as vital to these brands; 98% customers try brands they have tried before -550 + brands

Design For Action: The New Challenge

-typically, a new model from a company is not seen as threat, however, some risks include: -failure in a marketplace -phase out other models -regardless, designer does not think of those things -the more complex, the less feasible to ignore design artifacts; can even require business model change ex: massmutual "society of Grownups" was a multichannel effort that was disruptive to the businesses typical process -the larger the artifact, the more complex to integrate a new design; approach as parallel challenges of the design of the artifact and the intervention it brings

Learn your customer's pains

-undesired costs, situations, risks, and negative emotions -costs: same time, eliminate costs -underperformance? there is a lack of features, performance, and malfunctioning -main differences and challenges? know the top 10: big issues, concerns, and worries -barriers to adopt: upfront investment costs, resist change, and learning costs -feel bad?: frustrations and annoyances -risks/fear: financial, technical, social

social media development

-unexpected and quick -2018: 9/10 most downloaded apps are social media; users spend 135 minutes a day on it -rise of musical.ly and tik tok -microsoft bough linkedin for 26.2 billion -wikipedia became 5th most visited site -google purchased youtube for 5 billion -FB is largest "nation" in the world; users spend 52 minutes a day on FB- 40 billion in revenue and 16 billion in profits 2017 insta: 800 million MAUs, 95 million photos a day -whatsapp: 1.5 billion MAUs, 1.6 billion photos, 250 million videos, 65 million messages, 2 billion minutes in calls... largest messaging platform -china's tencent: market cap=n1/2 trillion; 2017 profits = 11 billion -snap: <800 million users, though 65% of 18-24 year olds have one -slack: 1/2 million DAUs, 300 mil messages; fastest growing business app, ever. 2019 value of 1 billion and after 120 million of in funding -2/3 of US adult get new from social media. though rise of fake news; countries like china and iran blocked social media bc protesting efforts -twitter: 336 monthly users, 500 million tweets a day; smaller, but big platform for politics. purchased by yahoo for 1 billion -stock overflow and github: software development tools; 5.7 million participate in Q and A, 73% questions in 11 minutes -pinterest: fastest growing website of all time... 2017 12.3 billion value -zynga and fornite; 1/4 billion march 2018 -twitter, yelp, tripadvisor= reviews -google purchased waze 1 billion -tinder: 22 million users; 1.4 swipes, 26 million matches a day; valued 3.3 billion 2015, grew 960% in 6 months

why the old way of building startups was wrong

-use to build by managing processes.... concept -> development -> testing -> launch -marketing: create materials/positioning; early buzz; launch event/branding -sales: hire during testing; simply executed plan given -business development: hired to do partnerships to attract mainstream -engineering: marketing requirements doc. during concept, development was waterfall, testing was QA, launch was hired technology publishing department

At times more is less and less is more

-value proposition: solving a customer problem or satisfying a need 1. products and services 2. gains for customers 3. solving pairs -not judge about product features -understand smallest possible feature set to solve issues (impossible); MVP

Get SMART: the social media awareness and response team

-viral complaints have cause UAL to go down 600 million -domino's employees gross video -uber ceo kalanick seen being rude to employee on video; gone within months -social media impacts whether you intend or not; get SMART -team members should be leaders; need company wide support; communication and tech skills; experts take up certain aspects

The Hierarchy of Engagement: Does the engagement of users creates virtuous loops in the product

-virtual loops are hard to create, but they propel a company forward -can also be inn form of growth and re-engagement -pinterest: more content, better discovery experience, more users pin

Big Skinny: Social Media

-wanted to use social media to mimic success at street fairs and bring new viewers to webpage -FB, twitter -social networking allowed allowed engagement

social media: introduction

-web 2.0: the new collaborative culture of the internet now; different than the static of the internet's past -social media= peer production

Amazon: An Empire Stretching From Cardboard Box to Kindle to Cloud: Amazon vs publishers

-when kindle rose, 6 publishers dominated the industry with 60% of trade publishing -amazon has started publishing; royalties 35-70% -faster to market -2016: 16/100 exclusive to amazon -independent authors on the rise -2017: 1,000 authors = 100,000 royalties

prediction markets and the wisdom of crowds (social media)

-wisdom of crowds: a group of people has more insight than professionals; key to many online resources -diverse crows can be very helpful -prediction market: diverse crowd polled to predict outcome (bestbuy tag trade for employees make 99.5% accurate predictions) -crowd needs to be: diverse, decentralized, collective verdict, and independent -google found geographic location important -can have legal and ethical drawbacks.. gambling, disclosure -would be more of a market w/ more people, but many are unsure if illegal -decentralization is key... blockchain

startup partner strategies

-you're not a peer -asymmetrical; don't confuse partners for earlyvangelists (know what they want now) v mainstream (guide to the future) -find partners that give unfair advantage; be bigger than them, could be you acquirer

creativity and the role of the leader: fanning the flames of motivation

1. Provide intellectual challenge -motivating people to perform at their peak is especially vital in creative -to discover that drivers of creative productivity, they looked at data on more than 11,000 R&D employees in manufacturing and service companies who had been routinely surveyed by the national science foundation -workers who are more intrinsically motivated are more creative than extrinsically motivated -early stage researchers who were more motivated by intellectual challenge tended to be more productive -a stronger desire for independence was also associated with somewhat higher productivity 2. allow people to produce their passions -if the keys to creative output are indeed intellectual challenge and independence, management must find ways to provide them -some people are simply more revolutionary in their thinking than others and therefore more suited to radical projects -when people are well matched to a project, granting them independence holds less risk 3. be an appreciative audience: -the fact that creative works are intrinsically motivated does not mean that managers' behavior makes no difference -people are highly attuned to a manager's engagement with and attitude toward a project -the wrong managerial behaviors or simply careless neglect can be demotivating -employees doing creative work are more motivated by managerial behavior even seemingly little things like a sincere word of public recognition -creative workers are reactions to failure -managers must decrease fear of failure and that the goal should be to experiment constantly fail early and often and learn as much as possible in the process -managing creativity are the ones that have been most successful because they develop an aversion to failure -ironically, companies' success at establishing the economic viability of an activity can lead to increased scrutiny and therefore to the companies' increases sensitivity to failure and desire to avoid it -fear of failure also seems to rise with the scale of a business -managers are more likely to deny that failure has happened and are more likely to erase all memory of it -management must create an environment of psychological safety, convincing people that they will not be humiliated, much less punished, if they speak up with ideas, questions, or concerns, or make mistakes -failures fall into unsuccessful trials, system breakdowns, and process deviations 4. provide the setting for good work -the potential for passionate engagement in one's work is hghest when the work iteself seems noble -work that is excellent technically, meaningful, and engaging to the workers, and carried out in an ethical way -rules tend to be bent in situations where market forces are dominant

There are four types of startups

1. existing 2. resegmented 3. new 4. clone market: size, cost of entry, launch type, competitive barriers, positioning sales: model, margins, sales cycle chasm width finance: original capital, time to profitability customers: needs, length to adoptions

3 drivers behind increased online ad spending

1. increasing user time online 2. improved measurement and accountability 3. targeting

4 main differences between mobile and desktop

1. smartphone apps can access the user's address book... easier to build social graph 2. smartphones can can access media library and cloud service; easier to share photos and videos 3. apps can have notifications (not great on desktop) 4. icon on home screen, although desktop allows for menu or icon, so mobile users favor single-purpose, specialized apps 5. authenication; sign on with mobile instead of password

The San Diego City Schools: Enterprise Resource Planning Return on Investment: Payroll

1. time and labor collecting: used time cards and time sheets 2. paycheck processing was done twice a month, processed weekly to correct errors. printed, not electronic.

Deep learning

A type of machine learning that can process a wide range of data resources. It requires less data preprocessing from humans, yet produces more accurate results than traditional-machine learning approaches. -interconnected layers of software-based calculators know as "neurons", which form neural networks -the network can injest large inputs of data, processing them through multiple layers that learn increasingly complex features of the data at each of those layers. -once this occurs, the network can make a determination about the data, learn if this determination is correct, and use what it has learned to make new determinations about the data. -for example, once the machine had learned what an object looks like, it can recognize the object in a new image

Design Thinking Crash Course Playbook

Gain Empathy: 1. interview, 2. dig deeper (ask why) 3. capture findings (synthesize) Ideate; generate alternatives to test: 4. take a stand with a point-of-view 5. sketch to ideate (go for volume) 6. share solutions and listen for feedback Iterate Based on Feedback: 7. reflect and generate a new solution 8. build a prototype 9. share your solution and get feedback Reflection and takeaways 10. group gather and debrief

Reporting systems

Reporting systems: integrate and process data by sorting, grouping, summing, and formatting; produce, administer, and deliver reports. They improve decisions by providing relevant, accurate, and timely information to the right person. They aggregate data across the organization and provide a one time snapshot of the data (a report)

SELECT TOP

SELECT TOP is used to specify # of records to return (MySQL=LIMIT9, oracle=ROWNUM) 1. SELECT TOP (# or %) column_name FROM table_name WHERE condtition

SQL

Structured Query Language -using for storing, manipulating, and retrieving data from databases -SQL can: execute queries, retrieve data, insert records, update records, delete records, create new databases, create new tables, create stored procedures -create views -set permissions on tables, views, etc ~there are different versions, but all follow basic commnands ~it is run by a relational database management system; data is related and stored in rows and columns -a record is an individual entry is a row -a field is an individual entry in column -standardization of database language

Understanding Mixtape: Discovering new insights via human engagement, Track 3: extreme users

WHY engage with extreme users? their needs are amplified, so engagement with them provides even more meaningful insights, especially needs most users might not reveal -so who is extreme? depends on what you're exploring; engage; look at the extreme in all of us for inspiration

Design For Action: Track 2: Facilitate a Brainstorm

WHY: can come with wide-ranging ideas and a good facilitator sets the stage for the team to be successful doing this HOW: 1. energy: keep it flowing, the questions you are brainstorming is important... create variation to other direction 2. constraints: add some to spark new ideas... "what if I had to" 3. space: be mindful; plenty of writing area, balance so that everyone can easily write

Design For Action: Track 3: How-Might-We Questions

WHY: generating these questions are seeds for your ideation that fall from P.O.V. statement, design principles, or insights... create H-W-M statement that is not too broad or narrow HOW: begin w/ POV or problem statement that retain your perspective -amp up the good -rename the bad -explore the opposite -question the assumption -go after adjectives -ID unexpected resources -create analogy from need or context -play against the challenge -change a status quo -Break POV into pieces

Design For Action: Track 6: Brainstorming

WHY: great way to come up with ideas you would not be able to generate by sitting w/ pen and papers; leverages collective thinking. turn off generative, turn on evaluative HOW: be intentional; invest energy; use HMW question, write down you're brainstorming; everyone shares Rules: -1 convo at a time -be visual -encourage wild ideas -go quantity -defer judgement -headline -build on ideas of others -stay on topic

Design For Action: Track 6: Selection

WHY: harvest brainstorms so ideas don't go to waste HOW: don't narrow too fast, don't immediately worry about feasibility; focus on what gets everyone excited 1. post it voting 2. the four categories method: rational, most likely to delight, the darling, and the long-shot) 3. bingo selection method: choose ideas based on different factors: a physical prototype, a digital prototype, and an experience prototype

Design For Action: Track 2: Prototype To Test

WHY: opportunities to examine solution as perceived by users HOW: start low-resolution; create experience and focus on what you are trying to test; think of context of feedback -start building -don't spend too long on one prototype -build with the user in mind -ID a variable

Understanding Mixtape: Discovering new insights via human engagement, Track 5: interview for empathy

WHY? you want to understand a person in order to innovate for them 1. ask why 2. never say "usually" when asking a question 3. encourage stories 4. look for inconsistencies 5. pay attention to non-verbal cues 6. don't be afraid of silence 7. don't suggest answers to questions 8. ask questions neutrally 9. ask one word answer questions 10. make sure you're prepared to capture the answer

machine learning systems and data mining

artificial intelligence systems learn from data and can be used to find patterns or even predictions from that data

The Hierarchy of engagement: the more important thing to measure: cohort performance

cohorts: # of weekly users completing action, and % of weekly users completing the action -visualize: triangle heats maps, charts, line graphs -shows: growth through size of cohort, engagement through ratio of performing core action, and retentionn through cohort performance over time

The San Diego City Schools: Enterprise Resource Planning Return on Investment: Choosing an HR System

criteria: product functionality, price, support, and organizational strength -would SCDS select a vendor who provide all modules of the ERP system or choose a best of breed system that would have all of the software systems merged together, which would guarantee all the functionality they wanted -although, this would increase complexity (gartner group advised against) -SAP, PeopleSoft, Lawson, SCT corporation claimed to provide a full package of functionality (unlikely) -wanted a company that had experience with education software, where success was uncommon

Design For Action: Intro

design has evolved from a process for physical goods to business processes -each design process is an improvement from the one before it because they learn from previous one to improve -the design challenge is the acceptance of "the designed artifact"

Online Reputation Management

firms track online presence to see if positive or negative and provides intelligence -google alets -bit.ly -salesforce -FB audience insights

The Big question

how do you know if a start-up had the potential to become a 1B+ company -in the early days: art (understanding how products) -mature days: science (knowing how to measure it)

Understanding Search (Google)

how it works: 1. perform search or query 2. see results that are organic/natural (google uses pagerank algorithim, which ranks web pages. while it does not accept money, it is based on the site with the most clicks) -results are improving through search engine optimization; working to improve link fraud and are further analyzing user behavior -some links from some websites carry more weight than others (influential, mobile friendly) -40 million searches per second, 3.5 billion searches a day, 1.2 trillion a year -spiders, web crawlers, software robots ask the public network for a list of public websites, and they then go through a list to follow every link until pages are uncovered -cached: tempory storage which speeds up computing -HTML code stops spiders from indexing a page -content hides in the deep web often -estimated 1.4 million servers on server farms (google invested 7.3 billion, which was mostly to server farms... strategic, as this investment presents another barrier to entry, profitability, and economies of scale) -these are custom built, use AMD processors, low end hard drives, and makes google intel's 5th largest customer -it also allows google to be more fault tolerant -they design their own hardware security chips -their tensor processing outperforms others -batteries backup each server -1.0= perfect power usage; google=1.1 -use AI to monitor efficiency -equipment operates using the world's largest grid computer -servers are all across the globe

Ideate Mixtape: Generating Unexpected Ideas via Reframing Your Challenge: Intro

ideas come from good questions... provoking prompts from specific and meaningful insights -What To Do? 1. prep: schedule the day and prepare yourself 2. reframe: what's your unique perspective? what do you and your team know about this challenge that no one else has considered? 3. brainstorming prompts: how-might-we questions 4. launch: 20 min, get excited 5. brainstorm: 60 min, go for quantity 6. select: 40 min, choose and discuss questions 7. wrap: 30 min, recap the work, plan next steps 8. going forward

search engines, ad networks, and fraud

types: -enriching click fraud: site operators generate ridiculous ads to boost CPC income -enriching impression fraud: when site operators generate false page views to boost cpm earnings -depleting click fraud: clicking a rival's ad to exhaust their cpc advertising budget -depleting impression fraud: when ad rank is base don ad performance, fraudsters repeatedly search keywords linked to rival ads, generating impressions w/o clicks.... drops rank -disbarring fraud: trying to frame a rival by generating clicks or impressions to get the rival banner from the network -list fraud: creates random websites, linking back to a page in hopes it will increase organic search -keyword stuffing: packing a site w/ unrelated keywords to attract different customers -social Influence: generating fake followers

Data mining systems

use sophisticated statistical techniques (regression analysis or decions tree analysis) to find patttenrs and relationships. They improve decisions by discovering patterns and relationships in data in order to predict future outcomes (market basket analysis or predict donations)

step 1 of data mining

business understanding: establish the goals you want to achieve, then think of how data mining will help you achieve those goals. from there, draw up a plan with timelines, actors, and roles

Big Bets: Bringing Potential Rivals and Platform Powerhouse into the Facebook Family

-"it's facebook that's going to be the next facebook" -instagram: at the time of Fb's IPO, insta had 13 employees, was 15 months old, had 50 million users, with 5 million more every week -whatsapp: mobile messaging was a bigger threat than photosharing. whatapp was the global leader, sending more messaged than sms text -their numbers were 12-64x facebook's messaging -72% of users used it every day -FB acquired for 19 billion; 16 billion in cash and 3 billion in stocks -lost 138 million, but 1.5 million monthly active users -oculus VR: it was purchased for 2 billion; it had no revenue, no customers, no complete product, -originally, it had 24 in commitments to support the firm -the goal was to build a computing platform for strategic advantage in the future -FB provides for development

facebook's success stats

-1 in 5 of the world's population have facebook; its daily traffic is 3x larger than the superbowl viewing -30% of usage in the united states is mobile -it is the 5th most valuable company in the world -85% of its users are from abroad

"facebook in 60 seconds"

-4 thousand users in two weeks -1 million users in 10 months -zuckerberg passed on offers from MTV and yahoo -he started by aiming toward the college market, but then expanded to the high school market, then the world -microsoft pays $100 million a year to broker ads -microsoft pays $240 million for 1.6% stake =15B -almost a billion users -private company: projected revenue 4.6 billion -facebook IPO

Artificial intelligence

-AI is a branch of computer science, seeking to use computer models as a tool to perform tasks that are normally associated with human cognitive functioning. -AI had adavanced recently due to algorithmic advancements, explosion of data, and exponential increases in computing and storage -advancements have been achieved most recently by applying machine learning to large data sets, where patterns can be detected and predictions can be made. These machines adapt and improve over time -in 2017, AI was a 2 billion dollar industry, while in 2021, it is expected to be a 58 billion dollar industry -85% of firms think they use AI towards their competitive advantage, but onl 20% do

Beacon Busted-Management Lessons on Tech Planning and Deployment

-Beacon= platform; could energy and nature of social networks be harnassed to offer truly useful consumer information to users? -the firm worked with vendors to see consumer activity at the point of purchase, putting it into newsfeed -opt out instead of opt in -moveon.org petition -shut down, 9.5 million donated to privacy groups -example of not fully considering implications

APIs, Playing Well With Others, and Value Added (FB)

-FB open graph allows developers to link web pages and app usage into the social graph -with just a few lines of HTML code, any developer can add a FB like button to their site -some websites can enter log in credentials not just on site -firms can leverage FB data (see likes of friends on yelp)

Move Fast and .Breaks Things: Fumbles, Fake News, and Global Growth Challenges

-Facebook's campus is "edgy" -has been modified to better tolerate mistakes; bold risks of new iniatives

Can FB's Ad Network Best Google:

-Fb has limited reach from fan pages; now offers promoted pages -Google: 20% ad revenue from 3rd party sites... split revenue -FB launched similar site: FB Audience Network; matches data to servers and delivers appropriately targeted app -FB can use dark web

Facebook Takes on Search

-Fb is private; content is not accessible through google -FB is introducing graph search, allowing users to draw info from the site's social graph -critics over privacy concerns

Attention Challenges: The Hunt vs the Hike

-Hunt: Google and other search sits; a task-oriented expedition to collect information that will drive a specific action (only charge advertisers with clicks, most profitable activity among any internet firm, 12.6 billion in profits and $10 billion in tax, FB has 16B) -Hike: FB; sort of know what they will encounter (goolge has 2% click rates, FB 0.2% click rates)

Understanding Mixtape: Discovering new insights via human engagement, Track 2: Assume a beginner's mindset

-WHY? going forward with assumptions can limit your ability to empathize 1. don't judge 2. question everything 3. be truly curious 4. find patterns 5. listen

Walmart and predictive technology

-Walmart used data mining to realize that during hurrricane frances, people purchased water, beer, and pop tarts the mostn

The Rise of Facebook and growing concerns:

-Zuckerberg: he was successful by age 23, was Time's person of the year, his net worth is 19 billion -the 100 billion dollar public offering and 16 billion that was raised made it the biggest tech IPO in history -in 2017, its revenue was 50 billion and its profits were 16 billion -there are privacy concerns -instead of depending on their IPO, they acqui-hire, aka acquiring start-ups to find new hires -so, where is FB going? will stock be valuable to workers and their 401k in the future? should investors invest? is FB a partner worthy firm? ... answer should be based of knowledge of the tech industry

Understanding Mixtape: Discovering new insights via human engagement (intro and what to do)

-a core principal of design thinking is human engagement (understanding users =new solutions) -what to do? 1. prep: schedule the day and prepare yourself 2. frame: structure the work in a human centered way 3. launch: 30 min, get the team prepared and excited 4. empathize: 120 min, engage with real people 5. unpack: 50 min, get the data visually on the wall (work in small groups, using the tensions, contradictions, surprises, to be aware) 6. needs and insights: 50 min, synthesize to find meaning (same groups, use insights) 7. wrap (30 min): recap the work, and plan next steps 8. going forward: now, consider idea mixtape to move on

Ads in Feeds: Better Performance Even As Fewer Ads Are Shown

-ads are not easy to ignore; they have 26x return on investment than side bar ads -more 18-24 on FB during primetime than watching TV -video ads on FB cost per thousand impressions: $4 -cutback ads= higher proce

Value of business intelligence tools

-analyze data -look for patterns -use patterns to make business decions (share info with business partners, manage inventory seasonally or drop items, and design marketing and advertising strategies)

RFM analysis

-analyzing and ranking customers according to purchasing patterns -R: recently customer ordered; to produce an R score, the program sorts customer purchase by data of most recent purchase. Essentially, the program would divide the customers into five groups, scoring them from 1-5. The top 20% of recent orders would have an R score of 1 -F: frequently customers orders; to produce an F score, the customers would be re-sorted based on their order frequency- again, the top 20% of the customers would have a score of 1 M: money spent on orders; to produce an M score, customers would be sorted by the money spent. The top 20% of the biggest spenders would receive a score of 1 -in the example, you would give a coupon to bloominghams, since they spend the most, but have not purchased from your firm in a while

Genetic Algorithms (GAs)

-are search algorithms that mimic the process of natural evolution. They are used to generate solutions to optimization and search problems using such techniques as mutation, selection, crossover, and chromosome. Model building techniques where computers examine many potential solutions to a problem, modifying mathematical models + comparing to choose the best option -refine data

Lessons from Platform FB: Big Growth, Bad Partners, APIS

-at the first F8 conference, FB published APIs that specified how programs could be written to run within and interact with FB -increase vendors, takes 50% revenue from goods sold through the site -more 3rd party= more network effects -security concerns, intellectual property violations -one year in, had marshalled efforts of 400,000 developers, 24,000 applications, and 140 apps each day -95% users had installed at least one of these -some succeeded, although some too FB reluctant and couldn't benefit from FB's changes

cons of data mining

-big data has issues with volume, variety, veracity (not all data is accurate), and velocity -some models are over-fitted (too complex) -there is cost of scale; not only do you have to purchase tools, but maintain them -concerns with privacy and security, especially as there is a turn towards the cloud

The Explainer: Design Thinking Video

-businesses struggle to make strategy by collecting backward data or making risky bets based on instinct instead of real data -design thinking is a strategy making process that avoids these mistakes by applying tools from the world of design and shifting focus to human behavior -popularized by David M. Kelley, Tim Brown, and Roger Martin

Content Adjacency Challenges: Do You Really Want Your Brand Next to That?

-content adjacency refers to concern over where a firm's advertisements will run/ think about all the questionable content on facebook that could be seen along with your ad -FB has tried to expel concerns by updating their guidelines on bad content, improve compliant reviewers training, restrict ads, and work with groups concerned about graphic content

pros of data mining

-decision making is automated, which is time-efficient and enhances day to day functions of a firm -allow for prediction and forecasting -costs are reduced -insights about customers and discovered

Problems with operational data

-dirty data (inaccurate, inconsistent, or incomplete/missing values data) -raw data might not be compatible to sophisticated data mining -data not integrated -too much data (curse of dimensionality, when data has too many features, and too many rows of data)

Why Mobile is Different (Sometimes Better) than desktop

-dominance of mobile not as definite as on desktop; lost to leaders instagram and whatsapp -the four key differences between mobile and desktop -smartphones can do so much; location, address book, apple pay, traffic, APIs, etc

Beyond "Right Margin" Display Ads: Big New Winners Emerge (FB)

-enhance ads with social context= social proof, which allows for 50% greater likelihood of higher recall and 35% sales lift

Facebook and network effects

-facebook is a classic example of network effects, where the user base is the key to the value (metcalfe's law) -friend's=value... the more fb friends you have the more value -fb tools bring users together (games, apps, chat, email, media)

Faked out by Fake News

-fake news spreads like wilfire, places pay for clicks -Russian Government: "internet research agency" -top 20 fake news get more traffic than top 20 real news -original estimate: 18 million, real was half a bilion -people accused facebook of turning a blind eye -hard to limit fake news; quarter of the planet uses Fb and it's free speech -option to report, notice patterns of fake news, see if image has been doctored, they're trying to use AI

Facebook Feed: Viral Sharing Accelerated (FB)

-feeds allow people to share with all their friends; uses algorithm to identify most interesting content -timeline offers a digital scrapbook -100,000 individual weights determining if a post is seen -accept $ for advertising, allows "subscribe" feature and hashtags -winner in traffic to newspaper sites and ahead in social traffic referrals

challenges in monetizing the social graph

-google's advertising is based on pay per clicks and a good click through rate of 2% -facebook's CTR is 0.04%, they aren't directly related to the search/visit, and content adjacency problems

FB Ads: Despite Post-IPO Concerns, A Massive Potential Upside Is Realized

-growth slowed 2 quarters prior to the firm's IPO; although, overcame skeptics.... 2017, revenue grew 40% -supply and demand, improve pricing, leverage data, new ad formats -precise targeting: large firms have tried to develop user profiles; a photography company benefited largely from FB targeting their ads toward newly engaged women -FB partnered with AcXiom, Datalogix, and Epsilon allow data purchasers to have access to loyalty programs -FB allows brands to upload info from their proprietary CRM ti show ads to existing customers

Content Providers, Step Deeper Inside My Walled Garden?

-instant articles service helps overcome slow page load problem; articles on FB 10x faster, formatted for mobile; will use FAN to sell unsold inventory for 30% cute; some think FB has too much power over journalism

Walmart focused on "cart"

-instead of focusing on the indidual customer, walmart focuses on what the typical customer would put in their cart -for example: cold medicine on sale is a loss leader (sale one something people come in for to generate traffic and get people to buy more of others things), but walmart knows they can also sale tissues, aspirin, cough drops, chicken soup, etc -this is called cross selling, where consumers come in for one category of good, and buy another (businesses use market basket analysis to determine what this basket consists of) -up-selling: consumers end up paying more, or buying a higher level of service

The Admirable Goals and Unintended Consequences of Internet.Org

-internet is a human right -solar-powered drone aircraft -85% of the world's population lives within at least 2G -plus, data is expensive, services aren't designed for emerging markets that need ultra-low bandwidth services, and content isn't compelling enough to . draw in users -in India, this could create jobs, cut poverty, and reduce infant morality -internet.org: non profit organization crafting partnerships with local carriers to offers free services on a low powered app -rocket carrying a satellite to Africa blew up -some say this simply a two tiered internet -"free basics" blame FB for country's political issues (hate speech and fake news)

New Ad Models (FB)

-marketers need to rethink how they advertise and take advantage of the social graph -advertisers hope to reach consumers with similar interests via engagement ads -privacy concerns come along with effective marketing

Messenger: A Pillar Business Building FB'a Future

-only whatsapp is bigger, 900 million users -Zuckerberg got paypal CEO to help develop messenger -features: group messaging, stickers, location sharing, notifications, audio messaging, VOIP -FB send customer updates, messenger= support chat -even can use AI chatbots -key is balance between critical vs spam -possible venom like features -can use app-stone like feature in future possibly

Examples of reinforcement learning in business

-optimizing the trading streatgy for an options trading-portfolio -balancing the load of varying electricity grids in varying demand cycles -stock and pick inventory using robots -optimize the driving behavior using robots -optimize pricing in real time for an auction with a limited supply

Lengthening Leaders, Quick Catch-up, and the Challenging Rise of Mobile (FB)

-social graph: concept from Zuckerberg, referring to Facebook's ability to collect, express, and leverage connections between the site's users -everything on FB is a node, connected to other stuff -trust is a huge part of FB's success -network effects are key, as are FB's powerful switching costs

Storage capacity and data mining

-storage capacity is increasing, and storage cost is decreasing; essentially, it is becoming somewhat unlimited -although this is beneficial, buried in this data is relationships, and that is where data mining comes in to play -finding these relationships can allow your business to make better, more informal decisions

Facebook's Dominance on the Desktop

-successfully implemented across several different services, although challenge of new desktop vs the old success of mobile -FB didn't acquire photosharing company; simply turned on the feature -chat is a key feature, video sharing, video streaming, marketplace, dating feature -network effects and switching costs allow firm to access new markets

data mining and its importance

-the process of using computers to find and analyze hidden, yet meaningful patterns in data, especially in large data sets -goal of data mining is the discovery of non-trivial data relationships and patterns (historical and future) -used to build machine learning models and power artificial intelligence

Understanding Mixtape: Discovering new insights via human engagement, Track 1: empathizing... empathy is the foundation of the human process

-to empathize you observe (view users and their behaviors), engage (interact with and interview users), and immerse (experience what your user experiences) -WHY empathize? human centered design projects need to understand people; to solve their problems, you must understand them -uncover insights= solutions -understand values, most through behavior 1. uncover needs 2. guide innovative efforts 3. identify the right users to design for 4. discover the emotions that guide behaviors

Walmart responding to disaster

-walmart gave Mississippi and Louisiana managers during the hurricane three rules 1. You have the authority to make decisions above your level 2. Make the best decision you can with the information available at the time 3. Above all, do the right right -sent 2, 498 emergency truckloads of supplies for relief

2 critical conditions of data mining

1) clean, organized data 2) data should reflect current and future trends

3 major stages of design thinking

1) form a few theories about what customers want, but don't have by immersing yourself in their lives 2) test products; experiments and prototypes; see how customers respond and adjust the product the, the price, and the positioning accordingly 3) bring the product or service to life; identify the activities , capabilities, and resources they need ex: proctor and gamble wanted to turn around olay, so they first noticed through observing the market that they were only targeting women over 50, missing huge markets... they tested, developed, and brought new products to market quickly

Facebook as a platform

1. apps: users derive additional value through the apps, which are often free or cheap (highly MAU base) 2. APIs allow programmers to post the application directly on their site

applications of data mining

1) marketing and targeting in a database: retailers can use data mining to better understand their customers, as well as better segment groups in order to target their promotions 2) credit risk management and scoring: data mining can be used to predict an individual's ability to pay off debt, which is then used to select their interest rate. 3) fraud detection: data mining models can be used to automatically detect and stop fraudulent spending habits 4) healthcare bioformatics: statistical models can be used to predict a person's likelihood of particular health conditions, based on risk factors they posses 5) spam filtering: email spam can be combatted by systems analyzing common characteristics of malicious messages; security software is informed, and sometime the email is never even seen. 6) recommendation systems: online retailers use data mining to predict customer behavior 7) sentiment analysis: data mining from social media, which uses text mining, aka using an input from social media to gain insights as a result of a pattern in the text. additionally, natural language processing can be used to find the contextual meaning of a language. 8) qualitative data mining: text mining can be used to make sense of a large set of unstructured data 9) financial modeling: building and trading systems to capitalize based on historical trends 10) hiring and promotion: use data to see characteristics typically associated with employee loyalty and success 11) customer churn: determining which customers are likely to leave and finding ways to avoid that

Data mining trends

1. Language standardization: like the way SQL developed to become the preeminent language for databases, users are begging to demand a standardization among data mining to allow users to conveniently interact with many different mining platforms while learning only one standard language 2. Scientific mining: it is popular in business, and now becoming more popular in science, like psychologists using association analysis to track and identify broader patterns in human behavior, or economists using forecasting algorithms to predict future market changes based on present day variables 3. Complex data objects: new methods are being implemented to analyze varied and complex data, like google using a visual searach tool for users to conduct a search using a picture 4. Increased computing speed: as data size, complexity, and variety increases, mining needs for complex tools and cycles of mining. 5. Web mining: as the internet expands, uncovering patterns and trends over it is valuable to organizations. Uses same techniques as data mining, but applies them to the internet

Primary business intelligence systems

1. Reporting systems 2. Data mining systems 3. Knowledge management systems: share knowledge of products, product uses, best practices etc, among employees managers, customers, and others. They improve decisions by publishing employee and others' knowledge, create value from intellectual capital, foster innovation, improve customer service, increase organizational responsiveness, and reduce costs 4. Expert systems: they encode human knowledge in the form of if/then rules and process those rules to make a diagnoses or recommendation. They improve decions making through non experts encoding, saving, and processing expert knowledge -these e

Patterns in data

A pattern within data is simply a relationship within the data -sometimes, these patterns are not very helpful/valuable, for example: there is data that people who buy diapers on Thursday also tend to buy beer

Machine learning: living in the age of AI

AI is everywhere; cars, emails, photographs • AI's rate of change has increased significantly, thanks to software - companies are spending billions on AI, yet it is becoming more accessible to children at the same time • Some people are thrilled; others are scared-regardless, dystopian talk / ideas are not fact based * A I will be one of the most important advancements... it will change everything * self-driving cars are one of the most exciting advancements; they will be more efficient and more safe ↳ however: still need more data (shared data to eliminate mistakes)... there are some great demos, but they aren't ready yet * racing toy cars are a fun example; they use neural networks to improve • Neural networks are AI based on the function of the brain; synapses form strength In learning, which defines layers, hooking them together through similarities... these can be trained for image classification. ↳ after enough examples, the computer learns how to form its own rules • one AI innovator was a young man who used A- I to create a system that eventually may be able to scan the texture of a pancreatic tumor (hard to access) to determine whether it is malignant or benign based on its texture • AI is very computationally heavy as far as training systems and running speed • hopefully some of these technology advancements will be helpful to farmers, although some are reluctant about technology replacing people ↳ genes can be modeled for seeds, pick the crops • Essentially, the goal of AI is to free humans from repetitive work ↳ "being more productive moans loss people" ↳ although people are concerned with job displacement, A-I will "bring new jobs and innovation" as well • so, how are we gonna prepare people? ↳ set rulers now! • China introduced an "AI" news anchor, that isn't really AI (no A- I can communicate that perfectly---yet • live digital animated humans ↳ ready in 24 hours through neural network training ↳actors can play younger or older versions of themselves ↳however, one concern stems from how it is hard to distinguish between real or fake • Still, no AI system is capable of one stop learning • some offer real improvements, like the girl who used the app to help with her visual impairment • AI helps connect people, and has the potential to help us in a multitude of facets of life ❤ what makes A- Is what it is today is the way that everyone Is invited

Challenges of AI

Challenges with AI begin with "Naked algorithms", where the public domain can access As I from Cloud providers or APIs. - quality of data - not enough data - tech staff needing training - prohibited data - "change management" - the negative consequences of misuse - hardware is monitored - competition in data, which typically results from moving early and scale

supervised data mining

In supervised learning, the algorithm learns the relationship between an input and an output. A human will label every point of data so that the algorithm can find this connection. The goal of supervised learning in data mining is prediction or classification; a process is considered supervised if the goal is to look for a single output variable. For example, spam filters use supervised learning to classify incoming emails as unwanted content and automatically remove them from the inbox once the algorithm has been trained, it can be applied to new data. Overall, there are two main types of supervised learning: classication, which separates the data, and regression, which fits the data. -common examples of models: 1. Linear regressions: predict the value of a continuous variable using one or more independent inputs, such as realtor using linear regressions to predict the value of a house based on square footage 2. Logistic regresssions: predict the probability of a categorical variable using one or more independent inputs, such as banks using using them to predict the probability that a loan applicant will default based on factors like their credit score, housebhold income, and age 3. Time series: forecasting tools that use time as the primary independent variable. Realities like macy's Use time series models to predict the demand for products as a function of time, using it to forecast and plan to stock stores with the right amount of inventory 4. Classification or regression trees: predictive modeling techniques that can be used to predict the value of both categorical and continuous target variables. Based on the data, the model will create sets of binary rules to split and group the highest proportion of similar target variables together, and through following these rules, the group that a new observation falls into will become its predicted value. Includes decision trees, which splits data features into branches at decision nodes (color) until a final decision output is made, a random Forrest, which improves the accuracy of a simple decision tree by taking a majority vote to predict the output (continuous like age for regression and discrete like color for classification), or gradient-boosting trees which is a technique that generates trees sequentially, where each tree focuses on correcting errors from past trees. 5. Neural networks: a neural network is an analytical model that is inspired by the brains and the connections of its neurons. They use inputs, and based on their size, will either be fired or not fired based on if they meet a certain requirement. The signal, or the lack of the signal, is combined with other signals within the layers of the network, where the process is repeated until an output is generated 6. K-Nearest Neighbor: this method is used to categorize a new observation based on past observations. Instead of being model-driven, it is data-driven. Does not make underlying assumptions about data and doesn't use complex processes to interpret inputs. Essentially, it classifies new observations by identifying the closest neighbors and assigning it the majority's value 7. Linear/quadratic discriminant analysis: this upgrades logistic regression to deal with nonlinear problems, aka those which changes to the input do not mean the output value will change 8. Support vector machine 9. Naive Hayes: applies Bayes theorem, which allows the proabability of an event to be calculated based on knowledge of related factors that might affect that event (emails containing the word money being spam) 10. Adaboost: uses a multitude of models to come up with a decision, weighting them based on their accuracy to come up with a prediction

Operational dashboard

Include metrics or KPIs to assess performance of key business processes day-to-day to quickly identify need for management intervention to maintain or improve performance

Snapping Snapchat's Lead with Stories

Snapchat lead October 2013; FB did march 2017, insta prior of august -by mid-2018, insta and whatsapp features were larger than snap's daily active user base -FB opened up to advertising and 3rd party (spotify, soundcloud, musical.ly) which was called platform play -Fb used network effects and competitive advantage to surpass snap

unsupervised data mining

These tasks focus on understanding and describing data to reveal underlying patterns. Recommendation systems, for example, employ unsupervised learning to track user patterns and provide them with personalized recommendations to enhance their customer experience some common models include: 1. Clustering: group similar data together. They are best used for complex data sets describing a single entity. Consists of k-means clustering, which pushes data in to number of groups based on similar chractersitsics based on the model, not in advance by the human. -Gaussian mixture modeling is a more general and flexibility model of k means clustering with the size and variation of the groups, -Hierarchcial clustering: splits or classifies the data along a hierarchical tree to form a classification system -Recommender system: often uses cluster behavior to identify the important data necessary for making a recommendation 2. Association analysis: also known as market basket analysis, it is used to identify items that frequently occur together 3. Principal component analysis: used to illustrate hidden correlations between input variables and create new variables, called principal components, which capture the same information contained in the original data, but with less variables. Reducing the number of variables allows analysts to convey underlying relationships in data

Understanding Mixtape: Discovering new insights via human engagement, Track 4: what? how? why?

WHY? a tool that helps you dive deeper into observation (concrete to abstract); especially good for photos WHAT: concrete. what is the person doing (objective) HOW: understanding how is the person doing what they're doing WHY: interpretation

step 3 of data mining

data preparation: clean existing data, and add any missing data. processing data can be intensive, distributed data systems are helpful and often used in DBMSes. -using distributed systems can also be more secure, since all of your data is not being stored in one data warehouse. this way, you can be sure that all of your data is not lost in the future.

Knowledge management systems

share knowledge of products, product uses, best practices etc, among employees managers, customers, and others. They improve decisions by publishing employee and others' knowledge, create value from intellectual capital, foster innovation, improve customer service, increase organizational responsiveness, and reduce costs -piazza is a great example

Expert systems

they encode human knowledge in the form of if/then rules and process those rules to make a diagnoses or recommendation. They improve decions making through non experts encoding, saving, and processing expert knowledge

Recurrent neural network

↳ a multilayered neural network that can store information in context nodes ↳ this allows the network to learn data sequencers and output a #, or another sequence ↳ USE WHEN you are working with time-series data or sequences (audio recordings, text) ↳An RNN can be used to predict the next word in sentence. In this case, it is "are you free ________?" - First, it will receive a command that indicates the start of a sentence. - Next, the neuron receives the word "are" and then output a vector of numbers that feeds back into the neuron to help it "remember" that it received "are", and that it receives this word first. -this process is repeated for the rest of the words in the sentence, and the state of each neuron is updated upon receiving each word -after receiving "free", the neuron will assign a probability to every word in the English vocabulary that could complete the sentence. In this case, a well trained RNN would predict that sentence to say, "are you free tomorrow?" ↳In business, RNNs can be used to: -generate analysis reports for securities traders -provide language translation -track visual changes to an area after a disaster to assess potential damages -assess the likelihood that a credit card transaction is fraudulent -generate captions for images -power chatbots that can address customer service issues

Convolution neural network

↳ multilayered neural network with special architecture inoroasingly designed to extract complex features of the data at each layer to determine the output ↳ USE WHEN: you have an unconstructed dataset and you need to infer from it ↳Can be used for image processing; for example the CNN receives an image and processes it as a collection of pixels. In the model's inner hidden layers, it will identify unique features, such as the individual lines that make up the letter A. After doing this, the CNN is able to classify the letter A when it finds its unique features that it has identified to make up the letter. ↳In business, it is used to diagnose health diseases from medial scans, can detect a company logo in social media to better understand marketing opportunities, understand customer's brand perception and usage through images, and detect defective products on a production line through images

data visualization and dashboard projects: getting started

↳consider things like project objectives, core KPIs, user needs, data readiness etc ↳behind every dashboard project is a visualization need ↳get the right concrete information (such as KPIs) by coordinating with strategic project sponsors that are driving the business initiative with those who will use the dashboard day to day. continuously circle back to them for feedback and schedule regular meetings and updates ↳things to consider: -what is the organizational need being addressed -have objectives been agreed upon with sponsors -what are the main data sources with sponsors -what are the core KPIs for users to understand -is the data integrated, accurate, and clean? ↳consider the audience your data is intended for a model it to them; think of your goal as telling a story using analytics... the who, the what, the how, the when, the where ↳since the goal is to boost usage of analytics, consider multiple dimensions ↳have data in order before identifying the key metrics to showcase, ensure that it can be integrated and showcased

data visualization and dashboard projects: Dashboard Layout and Design

↳shows the best practices for placing different metrics and visual analytics on a dashboard for effective presentation to your audience ↳if data visuals are not organized, it will miss the mark ↳amount of constant: no more than 3 or 4 charts of graphs for your audience; if you need more than that, make accessible with toggle replacing button ↳what goes where: have higher level metrics at the top with an overview if data with aggregate KPIs over time below; more granular as you go down -place most important info where people start to read ↳themes and formats: -good contrast and balance in color pallete -balance saturation, hues, and lightness to include more color data -titles, labels, and user instructions present, but not overhwhelmingly

data visualization and dashboard projects: data visual solution

↳the way to identify visualizations that both visualize different types of metrics and access useful tipis for grouping data by category ↳quantities: the bread and butter for KPIs: wide variety, but bar or column charts are a good place to start because they are best for quantities tracked over time; scatter plots are good for direct relationships between two quantities... will form a line if there is one and identify outliers ↳showing trends and changed over time: time series data is a deep analysis topic, but not crucial for most visuals; instead, it adds another dimension to the analysis with time on the horizontal axis -too much can be confusing, so less than 4 is ideal. useful for changes over time (growth and decline rates) are best indicated by clear text with a directional arrow ↳relative shares and proposals: display a relationship between the parts as a whole -pie chart to display aggregate quantity at one point in tine -stacked bar or area chart to show proportions over time -radar/spider web chart to capture relative importance of different aspects within the company ↳ranked lists -not really a data visualization, though it provides need to know info ↳geography and location -gives a richer experience by allowing exploration of spatial and physical relationships -don't include overlays, like street traffic, if they aren't relevant -limit colors, lines, and other details so that KPIs stand out -code regions differently -have ability to zoom in or out ↳dealing with multiple dimensions of once to become richer without confusing users -bubble scatter: size of bubble indicates numeric value of that point as well as color representing another factor -heat grid leverages 'temperatures' of horizontal and vertical access... good for multiple Y axis values for every X axis size of chart is is constant and colors represent value scale

data visualization and dashboard projects: user interaction

↳things to consider: dynamics filters, drill-down capabilities, etc ↳statics have their place, but dynamic is more interactive ↳filters and selectors: when providing a time range selector, provide a balanced mix of pre-formatted data ranges, as well as a full calendar for defining a custom range -time selector should drive all relevant metrics -dynamic period selectors that can be be expanded or contracted by circling or dragging an area can allow adjustment to time frame a user sees fit ↳Hover and Drill Down Interaction: when a user hovers over an item on a dashboard, it can be used to trigger some sort of highlighting or sharing so the point sticks outs; label a piece of data when a user hovers it to make it stand out -allow users to drill down to make grammar underlying information; data and context presented upon click must be distinct from original view ↳going further: project doesn't end with roll-out; have to be able to change data over time.... anticipate long-term requirements "future proof" data