Chapter 12

demand planning

both forecasting and managing customer demand to reach operational and financial goals

helps operations managers know what customers they should serve and at what levels of service

combined process of forecasting and managing customer demands to create a planned pattern of demand that meets the firm’s operational and financial goals

demand forecasting

predicting future customer demand

decision process in which managers predict demand patterns

a basic fact of forecasting is that the longer the time period over which you have to forecast, the greater the forecast error

adaptive forecasting

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

in adaptive forecasting, the smoothing coefficients in exponential smoothing models are automatically adjusted as a function of the tracking signal

demand management

influencing either the pattern (quantity or timing) of demand

proactive approach in which managers attempt to influence patterns of demand

factors that are influenced by business executives using demand management

• certainty of demand

• pattern

• timing

role of demand planning in operations management

managers need to make good predictions of product quantities that will be demanded at a given time or place

costs of inaccurate forecasts

too high: money lost in holding inventory, lost capacity, lost wages to workers that arent required
too low: lost sales, overused capacity, overworked employees, lower product availability

members of the supply chain who are affected by a high forecast

the firms that own the supply chain incur the expenses of high forecasting

the customers pay for it in the form of higher prices

elements of demand planning

forecasting, demand planning, and capacity planning

autocorrelation

describes the relationship of current demand with past demand

forecasting basics

understanding your data, time series analysis, forecasting metrics, forecast models

purpose of forecasting

explain as much variability as possible and describe or quantify the variability that cannot be explained

forecasting is a process

importance of forecasting

used in production planning, scheduling, inventory control, finance investment decisions, and workforce planning

users' characteristics and needs that have to be kept in mind when designing a forecasting process

level of detail

accuracy versus cost

fit with existing business processes

time horizon

forecasting basics principles

the forecast is always wrong
the forecast will always change
the further out in time the forecast, the less accurate it is
the greater the level of detail, the less accurate it is

why forecasts are always wrong

demand is uncertain, forecasts use historical data, forecasts are highly disaggregated, conditions change

how to handle forecast uncertainty

do not fixate on a single point, include a range of possibilities, use buffer capacity

granularity in forecasting

product aggregation (family vs. SKU), temporal aggregation (daily, weekly, monthly, yearly), geographical aggregation (country, region, market)

forecasting quality

measuring historical performance of forecasts to determine accuracy in predicting future demand

forecasting accuracy vs. bias

accuracy: closeness to actual observations
bias: persistent tendency to over or under predict

• forecast bias is the average forecast error over a number of period

• a positive forecast bias indicates that over time forecasts tend to be too low

true statements about situational drivers of forecast accuracy

the random forces that affect demand for individual products tend to be inconsistent across all products

as the time horizon for forecasting increases, more and more potentially unknown factors can affect demand

categories of forecasting methods

judgmental methods (Delphi method, surveys, scenario building)
causal methods (estimating relationships between variables like advertising and sales)
time-series methods (using past data to predict future values)
other methods (simulation, prediction markets)

forecasting rules

rule 1 - short-term forecasts are usually more accurate than long-term forecasts

rule 2 - forecasts of aggregated demand are usually more accurate than forecasts of demand at detailed levels

rule 3 - forecasts developed using multiple information sources are usually more accurate than forecasts developed from a single source

lead time

the time between the initiation and completion of a production process

ways to reduce lead time

• by speeding up poorly executed processes

• by eliminating unnecessary processes

• by speeding up processes which are redundant

true statements about simulation models

a relatively simple simulation-based approach is known as focused forecasting

a forecaster makes new forecasts using the combination of rules that would have provided the best forecasts for the past demands

focused forecasting combines common sense inputs from frontline personnel with a computer simulation process

time series models

a historical record of values listed in time order (such as sales history)

components of time series models

trend, seasonality, cyclicality, randomness

linear trend

results when demand rises or falls at a constant rate

forecasting metrics

mean deviation, mean absolute deviation, mean squared error, root mean squared error, mean percent error, mean absolute percent error

forecast error

unexplained component of demand that seems to be random in nature

forecast error formula

error = demand - forecast (et = Dt - Ft)

naïve forecasting

assumes demand in the next period is the same as the most recent period

moving average forecasting

averages demand over a set number of past periods to predict future demand

seasonality in forecasting

recurring fluctuations in demand at regular intervals (e.g., holiday sales spikes)

Seasonality can occur over a variety of time periods. Which of the following time periods are most likely to be used to forecast demand in operations? - daily, week, month

adjusting for seasonality

deseasonalize the data, generate a forecast, reseasonalize the forecast

moving average with seasonality

moving average model adjusted for seasonal fluctuations

postponable product approach

the keys to this approach are redesign of the product and redistribution of production resources so that the products can be easily configured close to the source of demand

when products are redesigned as postponable products, components rather than finished goods inventory is stocked closer to customer demand

CPFR

collaborative planning, forecasting, and replenishment

coordinated demand forecasting, inventory replenishment, and production and purchasing

systemic process that is being used as a means to integrate members of a supply chain

simple exponential smoothing

uses weighted averages of past data with a smoothing factor to predict demand

exponential smoothing with trend and seasonality

adjusts exponential smoothing to account for both trends and seasonal patterns

choosing the right forecasting technique

judgmental (Delphi techniques), experimental (focus groups), time series (historical data), causal (econometric models)

indicators of future demand in causal models

external factors, independent data, observed data

judgement-based (Delphi) techniques

grassroots forecasting - it is a technique that seeks inputs from people who are in close contact with customers and products

executive judgment - it is a technique used by high-level managers for long-term sales and business patterns

historical analogy - this approach to forecasting uses data and experience from similar products to forecast the demand for a new product

marketing research - this approach bases forecasts on the purchasing patterns and attitudes of current or potential customers

Delphi method - this approach develops forecasts by asking a panel of experts to individually respond to a series of questions

Which of the following scenarios are examples of the application of causal forecasting techniques?

A ski resort has increased occupancy when the snow base increases. Occupancy increases by 15% after a fresh snow fall.

When a firm spends on advertising, sales increase. The more ads placed, the more sales increase.

A firm has a new service idea, which has no current customers. The firm has no prior experience in this area. It wants to test the idea before committing the resources. The firm has resources to invest in two judgment forecasting techniques. Which of the following techniques will provide a rational forecast in this scenario? 

marketing research, the Delphi method

focused forecasting

relatively simple simulation-based approach that combines common sense inputs from frontline personnel with a computer simulation process

statistical model based forecasting techniques that transform numerical data into forecasts

time series analysis - this method extrapolates forecasts from past demand data

simulation models - this method tries to represent past phenomena in mathematical relationships and then evaluate data to project future outcomes

causal studies - this method looks for causal relationships between leading variables and forecasted variables

artificial intelligence - this method uses a smart computer program to learn from a combination of causal and simulation analyses using a wide array of data

• sources of data include internet, external data feeds, company sales and transaction systems

artificial intelligence combines..

focused forecasting techniques

causal modeling

time series analysis

simulation

forecasting process

forecasting model, demand planners, marketing and sales, management review

integrates information gathered from the market, from internal operations, and from the larger business environment to make predictions about future demand

seasonal index

ratio of each period’s actual demand to an estimate of the average (or base) demand across all periods in a complete seasonal cycle

computed by dividing each period’s actual demand by an estimate of the average (or base) demand across all periods in a complete seasonal cycle

possible sources of data when designing a forecast process

expert information

government economic information

past sales

efficiency vs. efficacy in forecasting

balance between forecasting accuracy and cost, minimizing total cost of forecasting and forecasting errors

handling inconsistent demand patterns

find alternative data sets, tap local knowledge, use ensemble modeling, test repeatedly

big data

consists of sensor data, transaction data, and descriptive data

demand versus sales

forecasting demand predicts what customers want, forecasting sales is based on past sales data and may not reflect true demand

demand planning summary

forecasts are judgment or statistical model-based
both accuracy and bias should be considered
demand management involves influencing customer demand
supply chains should be responsive to changes in demand
big data, social media, IoT, and artificial intelligence improve demand planning

• big data refers to the voluminous amount of information that are easily accessible through interconnected systems