Supply Chain Chapter 12: Demand Planning: Forecasting and Demand Management

The longer the lead time..

... the greater the forecast error

Managers try to manage demand by using variants of three basic tactics:

1. Influence the timing or quantity of demand through pricing changes, promotions, or sales incentives ex: automobile dealerships offering promotional packages 2. Manage the timing of order fulfillment Examples: disney places signs saying how long wait is 3. Substitute by encouraging customers to shift their orders from one product to another, or from one provider to another example: Dell is famous for selling what it has

Fluctuations causing operational inefficiencies all across the supply chain:

1. requiring extra resources to expand and contract capacity to meet varying demand 2. Backlogging (delivering later than originally promised) certain orders to smooth out demand fluctuations 3. Customer dissatisfaction with the system's inability to meet all demands 4. Buffering the system with the use of safety stocks, safety lead time, or safety capacity

Rules for indicating how situational characteristics tend to affect forecast accuracy:

1: Short term forecasts are usually more accurate than long-term forecasts 2: Forecasts of aggregated demand are usually more accurate than forecasts of demand at detailed levels 3: Forecasts developed using multiple information sources are usually more accurate than forecasts developed from a single source -difficult for a single source of information to comprehend all of the forces -unlikely all sources would be "wrong" in the same direction

Forecasting Performance

= primary measure is forecast error positive forecast = overly pessimistic negative forecast = overly optimistic

Statistic Model Based Forecasting

= transforms numerical data into forecasts using one of three methods 1. Time series analyses, which extrapolate forecasts from past demand data 2. Casual Studies, which look for causal relationships between leading variables and forecasted variables 3. Simulation Models, which try to represent past phenomena in mathematical relationships and then evaluate data to project future outcomes 4. Artificial Intelligence, in which a "smart" computer program "learns" from a combination of causal and simulation analyses using a wide array of data

Focused Forecasting

A combination of common sense inputs from frontline personnel and a computer simulation process

Stable Pattern

A consistent horizontal stream of demand examples: mature customer products like shampoo and milk

Weighted Average Model

A forecasting model that assigns a different weight to each period's demand according to importance typically more weight is given to more recent demand weight given to the demand value should all end up equalling 1

Moving Average Models

A forecasting model that computes a forecast as the average of demands over a number of past periods

Marketing Research

A forecasting technique that bases forecasts on the purchasing patterns and attitudes of current or potential customers

Historial Analogy

A forecasting technique that uses data and experience from similar products to forecast the demand for new products

Regression Analysis

A mathematical approach for fitting an equation to a set of data most commonly used method for estimating relationships between leading indicators and demand

Collaborative Planning, forecasting, and replenishment (CPFR)

A method by which supply chain partners periodically share forecasts, demand plans, and resource plans in order to reduce uncertainty and risk in meeting customer demand -supply chain partners must first come to an understanding regarding their relationship and the roles they will play -market planning: The partners collaboratively discuss such issues as the introduction of new products, store openings/closings , changing inventories prices and product demands -Demand and Resource planning: customer demand and shipping requirements are forecasted -Execution: ordres are placed, delivered,received and paid for. -Analysis: Execution is monitored and key performance metrics are collected with the goal of identifying opportunities for future improvement

Mean Squared Error MSE

A more sensitive measure of forecast errors that approximates the error variance.

Exponential Smoothing

A moving average approach that applies exponentially decreasing weights to each demand that occurred farther back in time By using the exponential smoothing equation again and again and again from one period to the next, each new forecast is implicitly built upon many past actual demands, and each will receive less and less weight as one goes back in time. Really just a sophisticated form of the weighted average model Still only reactive models, do not anticipate the effects of a trend The forecast can reduce the lag effect by increasing the value of the smoothing coefficient, but this also increases the risk of adding unwanted variability to forecasts as they overreact to random variations in demand.

Shift or Step Change

A one-time change in demand, usually due to some external influence on demand, such as major product promotional campaign, or sudden economic shock.

Smoothing Coefficient

A parameter indicating the weight given to the most recent demand between 0 and 1

Demand Management

A proactive approach in which managers attempt to influence the pattern of demand involves use of pricing and promotional activities

Postponable Products

A product designed so that it can be configured to its final form quickly and inexpensively once actual customer demand is known

Naive Model

A simple forecast approach that assumes that recent history is a good predictor of the near future assumes tomorrow's demand will be the same as todays ignores the trend, seasonal, competents of historical series and creates highly erratic forecasts

Adaptive Forecasting

A technique that automatically adjusts forecast model parameters in accordance with changes in the tracking signal

Simple Linear Regression: Time Series

A technique that finds optimal values for the parameters a and b, that is parameters that will most closely equate the independent variable t, and the dependent variable d over a set of values

Grassroots Forecasting

A technique that seeks inputs from people who are in close contact with customers and products

seasonal Index

An adjustment factor applied to forecasts to account for seasonal changes or cycles in demand = Each period's actual demand / estimate of the average demand across all periods in a complete seasonal cycle

Mean Percent Error (MPS)

Average Error represented as a percentage of demand

Demand Forecasting

Decision process in which managers predict demand patterns

Step 4

Document and apply the proposed technique to the data required for the appropriate business process

Time series Analysis Models

Forecasting models that compute forecasts using historical data arranged in the order of occurrence Forecasts are generated by summing weighted values of past demands, and the weighting schemes range from very simple to very complex -the type of weighting used depends upon the demand pattern

Executive Judgement

Forecasting techniques that use input fro high-level experienced managers

Delphi Method

Forecasts developed by asking a panel of experts to individually and repeatedly respond to a series of questions -prevents any single individual from dominating the process

Step 1: Designing a Forecasting Process

Identify the users and decision-making processes that the forecast will support -forecasting process needs to be designed with the following users' characteristics in mind: Time Horizon (lead time) Level of Detail Accuracy vs Cost: important to weigh the costs created by forecast errors against the costs of achieving greater accuracy Fit with existing processes In order to be useful, forecasting process must be integrated into existing business model

Big Data

Large amounts of data made available through sensors and interconnected systems

Step 5

Monitor the performance of the forecasting process for continuous improvement

Artificial Intelligence

Refers to learning and decision making capabilities that stems from software algorithms next generation approach that combines time series analysis, casual modeling, simulation and focused forecasting techniques

Seasonality and Cycles

Regular demand patterns of repeating highs and lows restaurants experience this with lunch rush, dinner rush etc. Economic, political and demographic and technological factors influence these patterns

Step :3

Select forecasting techniques that will most effectively transform data into timely accurate forecasts over the appropriate planning horizon

Simulation Models

Sophisticated mathematical programs that offer forecasters the ability to evaluate different business scenarios that might yield different demand outcomes helps forecasters to better understand how different variables and drivers of demand relate to one another

Root mean square error (RMSE)

Square root of MSE gives good approximation of standard deviation

MAPE Mean absolute percentage error

The MAD represented as a percentage comparability across product

Mean Absolute Deviation (MAD) also called mean absolute error

The average size of forecast errors, irrespective of their directions

Demand Planning

The combined process of forecasting and managing customer demand to create a planned pattern of demand that meets the firm's operational and financial goals Basically Demand forecasting and demand management

Autocorrelation

The correlation of current demand values with past demand values

Forecast error

The difference between forecast and the actual demand -simply the unexplained component of demand that seems to be random in nature Difference between forest and actual demand

Trend

The general sloping tendency of demand, either upward or downward in a linear or nonlinear fashion

Forecast Accuracy

The measurement of how closely the forecast aligns with the observations over time

Tracking Signal

The ratio of a running total of forecast error to MAD that indicates when the pattern of forecast error is changing significantly

Forecast Bias

The tendency of a forecasting technique to continually over predict or under predict demand. Bias = mean forecast error

Determining Trend Factors

The trend component of a time series normally results from some market force that causes a general rise or decline in values over time.

Step 2: Identify likely sources of the best data inputs

Today's information rich environs

Casual Models

Where time series models use only past demand values as indicators of future demand, casual models use other independent, observed data to predict demand -these models concentrate on external factors that are thought to cause demand Examples: The amount of household disposable income in an economy might be a good leading indicator of the sales of luxury items like sailboats

Demand Management activities

adjust product characteristics including price, promotion, and availability the purpose is the influence product demand to achieve sales objectives and to acccomdate the supply chain resources and capacities that the firm has in place

Judgement- Based Forecasting

built upon the estimates and opinions of people, most often experts who have related sales or operational experience

Primary goal in designing forecasting process:

generate forecasts that are usable, timely and accurate

Adjusting Forecasts for Seasonality

seasonal variations in demand