CHAPTER 4

DEMAND FORECASTING

Demand forecasting is essentially a judgment of future probabilities of the future demand of a product or
a service. Since the future is uncertain, no forecast can be 100 cent percent correct. Several methods of
demand forecasting are available to business decision-makers. However, there is no unique method that
will always guarantee the best result. Therefore, it is important to get acquainted with the whole range of
techniques of forecasting.

Classification of demand forecasts

1. Active vs passive forecasting

Active forecasting is a method of forecasting demand based on the consideration that a firm is likely to
initiate some action like changes in product quality, size, price etc. Passive forecasting on the other hand
is based one based on the assumption that the same product is being offered without any changes.

2. Short-run vs long-run forecasting

Short-run forecasting are those demand forecasts that will extend to the period of up to one year. These
forecasts are useful for product scheduling, inventory planning, and mobilization of working capital etc.
Long-run forecasts extend for a period beyond one year. They are helpful in capital budgeting, product
diversification, personnel recruitment etc.

3. Company forecasting vs Industry forecasting

This relates to forecasting of the demand of products of a particular firm. Company forecasting are firm
specific and are designed to serve the individual firms in planning their policies. Industry forecasting
forecasts the demand of the products of the industry as a whole. The association of manufacturers or trade
associations undertake them and serve the needs of all the firms in the industry.

4. Micro level forecasting vs Macro level forecasting

Micro level forecasting may take the form of company forecasting or industry forecasting. Macro level
forecasting is concerned with forecasting of the general economic environment and business conditions in
the country as a whole. National income growth rates, production indices, price indices among others
provide useful insights into the future demand for most of the commodities.

Factors Influencing Demand Forecasting

There are a number of factors that can significantly impact on demand which need to be taken into
account prior to forecasting. Here are the five most common influencers impacting forecasting and
demand management.

Seasonality: As seasons change, so can demand. A highly seasonal brand, or a cyclical business, may
have a peak season when sales are booming followed by off-seasons when sales are steady or even very
slow. Some demand forecasting examples based on seasonality include products used during specific
seasons (boating gear during the summer), holidays (costumes and candy on Halloween) or events
(wedding season, for example).

Competition: When competition enters or exits the scene, demand can drop or skyrocket. For example, if
a new player enters the market and starts vying for its share of the pie, established businesses may suffer;

1
on the other hand, if an existing competitor folds, or begins losing ground because of bad product,
service, or PR, other businesses will be in greater demand as consumers make a switch.

Geography: Where your customers reside and where you manufacture, store, and fulfill orders from can
have a huge impact on inventory forecasting (not to mention shipping costs). So, it pays to be strategic
when choosing geographic locations of your supply chain. For example, if you sell swimwear, you’ll
probably want to store the majority of your product in a county like Mombasa where they are ordered
most; that way, you don’t have to transport to faraway locations.

Economy: Economic conditions can have a big impact on forecasting product demand. For example, if
an economy enters into depression or recession, and fewer people are working, the demand for high-
priced, luxury products is likely to fall, while demand for low-priced, generic products is likely to
increase.

Types of Goods: Different products and services have very different demand forecasting. For example,
forecasting demand for perishable goods with a short shelf-life must be very precise or a lot of inventory
could be lost. On the other hand, demand can more or less be predicted for a subscription box service
that’s shipped to the same customers at the same time each month (assuming customer retention and
attrition are relatively steady).

Objectives of Demand Forecasting:

The objectives of demand forecasting can be divided in two categories namely:

A. Short term objectives,
B. Long term objectives.

A. Short Term Objectives:

(a) Formulation of Production Policy: Demand forecasts help in formulating suitable production policy
so that there may not be any gap between demand and supply of product. This can further ensure:

(i) Regular Supply of Material: By the determination of desired volume of production on the basis of
demand forecasts, one can evaluate the necessary raw material requirements in future so as to ensure
regular and continuous supply of the material as well as controlling the size of inventory at economic
level.

(ii) Maximum Utilization of Machines: The operations can be so planned that the machines are utilized
to its maximum capacity.
(iii) Regular Availability of Labour: Skilled and unskilled workers can be properly arranged to meet the
production schedule requirement.
(b) Price Policy Formulation: Sales forecasts enable the management to formulate some appropriate
pricing mechanism, so that the level of price does not fluctuate too much in the periods of depression or
inflation.
(c) Proper Control of Sales: Sales forecasts are calculated region wise and then the sales targets for
various territories are fixed accordingly. This later on becomes the basis to evaluate sales performance.
(d) Arrangement of Finance: On the basis of demand forecast, one can determine the financial
requirements of the enterprise for the production of desired output. This can lead to minimise the cost of
procuring finance.
2
B. Long Term Objectives:

If period of a forecast is more than a year then it is termed as long term forecasting. The following are the
main objectives of such forecasts:

(i) To Decide about the Production Capacity: The size of the plant should be such that the output
conforms to sales requirement. Too small or too large size of the plant may not be in the economic
interest of the enterprise. By studying the demand pattern for the product and the forecasts for future the
enterprise can plan for a plant/output of desired capacity.
(ii) Labour Requirements: Expenditure on labour is one of the most important components in cost of
production. Reliable and accurate demand forecasts can ensure best labour facility and no hindrances in
the production process.
(iii) Long term production planning can help the management to arrange for long-term finances.
The analysis of long term sales is more significant than short term sales. Long term sales forecast help the
management to take some policy decisions of great significance and any error committed in this may be
very difficult expensive to be rectified.

.
Steps in Demand Forecasting
Demand forecasting is an intricate exercise and conducted in a systematic and meticulous manner.
Demand forecasting involves the following steps.

I. identification of the objective

The first step in Demand forecasting is identification of the objective or purpose for which Demand
forecasting is undertaken. Based on the objective the type, level and the method of forecasting will be
selected.

2. Nature of the product

Before proceeding towards Demand forecasting, the forecaster should be clear about the nature of the
product whether a consumer good or capital good, Durable good or an intermediate product. This is
essential since the factors affecting the demand and the decision variables widely differ from one product
to the other.

3. Determination of demand
Demand for different products depends on different variables. Therefore, a forecaster should identify the
relevant determinants of the demand for a commodity before he proceeds with demand forecasting.

4. Choice of appropriate method

The forecaster should select an appropriate method of forecasting based on the nature of the product.
There is nothing like a universal method of Demand forecasting. The techniques used for forecasting
demand for an existing product may not be suitable for a new product. Hence, choice of the correct
method of demand forecasting is an essential prerequisite for forecasting accuracy.

5. Analysis

3
Once the demand of the product is forecasted, the forecaster has to apply his analytical powers to properly
understand and interpret the results. If necessary he should subject the results to available statistical tests
like coefficient of variation, percentage absolute tests etc.
Importance of Demand Forecasting

1. Production Planning and product scheduling

A business firm cannot function in wilderness. It has to take crucial decisions about what to produce and
how much to produce. This in turn depends upon its estimates of future demand for the product. If the
forecasted demand is likely to rise, the firm can plan expansion of its production capabilities to meet the
growing demand at the right point in time. In the eventuality of declining demand the firm should resort
to product improvement, diversification, design changes or even pursue an aggressive sales promotion
strategy

2. Inventory planning
Demand forecasting is useful for the firm in enabling it acquire the right quantum of inventory at the right
point in time so as to meet the needs of the production department without unnecessarily locking up the
finances of the firm in inventory accumulation.

3. Capital planning
Increased production requires increased capital resources fixed as well as working capital. Availability of
demand forecasts helps the firm to mobilize the capital resources in time.

4. Marketing strategy
Demand forecasting is useful in devising appropriate sales promotion or marketing strategies. If the
demand forecasts indicate a declining trend in sales, the firm should resort to intensive sales promotion
campaign so as to sustain its sales. Demand forecasting will also help the firm in setting sales targets for
the sales personnel.

5. Manpower planning
A firm has to recruit and train the appropriate level of work force it engages in the production process.
This therefore calls for forecasting the demand of the products well in advance so that the required
contingent of the labour resources could be obtained.

6. Pricing strategies
Devising and setting the optimum price depends upon the forecasted demand. If the forecasts indicate a
declining share in the market demand for the firm’s products; then it has to slash prices so as to sustain
demand. Conversely, if the forecasts indicate increased demand for the product over a longer period then
the firm can charge higher prices subject to the other considerations.

Methods of Demand Forecasting

The methods of demand forecasting are:

A. Opinion survey methods and

B. Statistical methods.

A. Opinion Survey Methods

4
1. Market Research/Surveying: Market research is another form of demand forecasting, with customer
surveys being important demand forecasting tools. Today, online surveys make it easy to target your
audience and survey software makes analysis much less time-consuming than in the past.

Using surveys, forecasters can gain a lot of valuable insights that simply can’t be mined from a sales
figure. They can help paint a better picture of your customer and their needs, inform marketing efforts,
and identify opportunities.

Some of the most popular surveys with sales and marketing teams include:

Sample surveys, in which a select sample of potential buyers are interviewed to determine their buying
habits.

Complete enumeration surveys, in which the largest possible sample of potential buyers are interviewed
to gather a broader data set.

End-use surveys, in which other companies are surveyed to determine their view on end-use demand.

2. Sales Force Composite Method: Also known as the "collective opinion," the sales force composite is
a demand forecasting method in which sales agents forecast demand in their territories. This data is
consolidated at the branch, region, or area level, and then the aggregate of all factors is considered to
develop an overall company demand forecast. This “bottom-up” approach is valuable because salespeople
are very close to the market and can often provide more accurate predictions based on their direct
experience with customers.

When using this method, remember that factors like product price, marketing campaigns, customer
affluence, and competitors can differ based on region, so it’s important to take this into account when
forecasting. Some inventory management platforms have built-in features allowing sales executives to
gather and send this data electronically, while others will use market research surveys to gather data.

3. Expert Opinion: The expert opinion method of demand forecasting involves seeking insights and
predictions from industry experts, professionals, or individuals with specialized knowledge in a particular
field. This method relies on the expertise and judgment of knowledgeable individuals to provide informed
opinions about future demand trends. Here's how the expert opinion method typically works:

Expert Selection: Identify and select experts who have relevant experience and knowledge in the
industry or market under consideration.

Experts may include industry analysts, economists, market researchers, or individuals with a track record
of accurate predictions in the specific domain.

Structured Interviews or Surveys: Conduct structured interviews or surveys with the selected experts to
gather their opinions on factors influencing future demand.

Develop a set of specific questions related to market trends, consumer behavior, economic conditions, and
other relevant variables.

5
Collation of Responses: Compile and analyze the responses from the experts, focusing on common
themes, patterns, and consensus. Some methods involve aggregating individual forecasts, while others
may give more weight to the opinions of experts with a proven track record.

Qualitative Analysis: Analyze the qualitative aspects of the expert opinions, considering factors such as
emerging trends, technological advancements, regulatory changes, or any other qualitative insights
provided.

Quantitative Integration (if applicable): In some cases, quantitative methods may be used to integrate
expert opinions into a more formal forecasting model. This could involve assigning numerical weights to
different expert predictions based on their perceived credibility or relevance.

Documentation of Assumptions: Clearly document the assumptions and reasoning behind each expert's
opinion to provide transparency and context for the forecast.

Regular Updates: Periodically update the expert opinions to reflect changes in market conditions,
emerging trends, or any new information that may impact demand forecasts.

4. Delphi Method:

The Delphi method is a qualitative forecasting technique that involves a structured, iterative approach to
gather and distill the opinions of a panel of experts. It is commonly used in demand forecasting to make
predictions about future trends, market conditions, or technological advancements. Here's an overview of
how the Delphi method works in the context of demand forecasting:

Selection of Experts: Identify a diverse group of experts with relevant knowledge and experience in the
industry or market under consideration. Experts can include professionals, academics, or individuals with
a deep understanding of the subject matter.

Round 1 - Generation of Forecasts: The experts are provided with a set of open-ended questions related
to the demand forecasting topic. Each expert independently formulates their forecasts and provides
reasoning for their predictions.

Compilation and Anonymity: The facilitator collects and compiles the responses while keeping the
identities of the experts anonymous.

The anonymity helps prevent biases, groupthink, or dominance of opinions by influential participants.

Feedback and Consensus: The compiled responses are shared with the experts in the form of a summary
or statistical analysis.

Experts are encouraged to revise their forecasts in light of the collective insights without knowing the
identities of other participants.

Round 2 (and subsequent rounds):

Steps 2-4 are repeated through multiple rounds until a convergence of opinions or a clear consensus is
reached.
6
The facilitator may provide additional information or insights from the previous rounds to guide experts
towards convergence.

Consolidation of Results: The final set of forecasts and the rationale behind them are consolidated into a
cohesive report.

The report may include a range of opinions, highlighting areas of consensus and divergence among the
experts.

The Delphi method offers several advantages in demand forecasting:

 Diversity of Perspectives: Involving a range of experts ensures a diverse set of opinions, reducing
the risk of tunnel vision or bias.

 Anonymity: Experts can freely express their views without the fear of being influenced or judged,
fostering independent thinking.

 Iterative Process: The iterative nature of the Delphi method allows for refinement and adjustment
of forecasts based on evolving information.

 Reduced Group Dynamics: By avoiding face-to-face interactions, the method minimizes the
impact of social dynamics and power structures within the group.

Because the Delphi method allows the experts to build on each other’s knowledge and opinions, the end
result is considered a more informed consensus.

5. Barometrics: The barometric method of demand forecasting is a technique that relies on leading
indicators, often referred to as "barometers," to predict changes in economic activity and subsequently
forecast demand. This method assumes that certain economic indicators can serve as precursors or signals
for future changes in consumer behavior. Here's an overview of how the barometric method works:

Selection of Leading Indicators: Identify and select specific economic indicators that are believed to
precede changes in demand for a particular product or in a specific market.

Leading indicators could include factors like changes in consumer confidence, interest rates, industrial
production, stock market indices, or other variables that have demonstrated a historical correlation with
changes in demand.

Data Collection and Monitoring: Regularly collect and monitor data on the chosen leading indicators
over time. Establish a historical relationship between these indicators and changes in demand for the
product or service being forecasted.

7
Analysis and Correlation: Analyze the historical data to identify patterns or trends in the leading
indicators and correlate them with subsequent changes in demand. Establish statistical relationships that
can be used to predict future demand based on the observed patterns.

Forecasting: Use the current values of the leading indicators to make predictions about future changes in
demand. The assumption is that if the leading indicators exhibit certain patterns or movements, it can be
inferred that a corresponding change in demand will follow.

Regular Updates: Continuously update the data on leading indicators and adjust the forecasts as new
information becomes available. Economic conditions and the relationships between indicators may
change, requiring ongoing monitoring and adjustment of the forecasting model.

Limitations and Considerations:

Recognize the limitations of relying solely on leading indicators; external factors and unforeseen events
can impact demand.

Validate the reliability of chosen indicators and their historical correlation with demand before relying on
them for forecasting. This forecasting method uses three indicators to predict trends.

Leading indicators attempt to predict future events. For example, an increase in customer complaints
due to shipping delays or backorders could lead to a decrease in sales.

Lagging indicators analyze the impact of past events. For example, a spike in sales the month prior
could indicate a growing trend that needs to be watched closely for inventory purposes.

Coincidental indicators measure events happening right now. For example, real-time inventory turnover
demonstrates current sales activity.

Each indicator can be used to conduct better inventory planning and improve supply chain management.

6. Econometric Method: The econometric demand forecasting method accounts for relationships
between economic factors. For example, when the COVID-19 pandemic became widespread in 2020,
there was an increased demand for online shopping as customers locked down and avoided the in-store
experience. Another economic example could be an increase in disposable income coinciding with an
increase in travel, as more people book vacations with their extra money.

While it may sound simple in theory, the econometric demand forecasting methodology can be extremely
challenging, as forecasters are rarely able to conduct controlled experiments in which only one variable is
changed and the response of the subject to that change is measured. Instead, econometrics are determined
using a complex system of related equations, in which all variables may change at the same time. There’s
a reason those that employ this method aren’t just forecasters. They have their own title: Econometricians.

Econometric methods of demand forecasting involve the use of statistical techniques, economic theories,
and mathematical models to analyze historical data and make predictions about future demand. These
methods aim to capture the relationships between various economic variables and the demand for a
product or service. Here's a general outline of how econometric methods work in demand forecasting:
8
Model Specification: Identify the relevant economic variables that influence demand. These variables
can include price, income, advertising expenditure, and other factors specific to the product or market.

Data Collection: Gather historical data on the chosen economic variables and the corresponding demand
for the product or service. Ensure that the data cover a sufficiently long period to capture different
economic conditions and trends.

Model Estimation: Choose an appropriate econometric model (such as linear regression, multiple
regression, or time-series analysis) based on the nature of the data and the relationships being studied.
Use statistical techniques to estimate the parameters of the model, determining the strength and
significance of the relationships.

Variable Transformation: Transform variables if necessary to meet the assumptions of the chosen
econometric model. Common transformations include taking logarithms, differencing time-series data, or
standardizing variables.

Model Evaluation: Assess the goodness of fit of the econometric model by examining statistical
measures such as R-squared, F-statistic, and the significance of individual coefficients. Validate the
model using techniques like cross-validation to ensure its predictive accuracy.

Forecasting: Once the model is validated, use it to forecast future demand by inputting values for the
relevant economic variables. The forecasted values provide an estimate of how changes in the
independent variables are likely to impact the dependent variable (demand).

Sensitivity Analysis: Conduct sensitivity analyses to understand how changes in key variables or
assumptions affect the forecast. Assess the robustness of the model under different scenarios.

Regular Updates: Update the model regularly with new data to account for changing economic
conditions and ensure the accuracy of ongoing forecasts.

7. A/B Experimentation: A/B experimentation, also known as split testing or randomized controlled
trials, is a method commonly used in demand forecasting to assess the impact of changes in marketing
strategies, pricing, or other variables on consumer behavior. While it may not be a traditional demand
forecasting method, it can provide valuable insights into how changes affect demand. Here's how A/B
experimentation is typically applied to demand forecasting:

Hypothesis Formulation: Formulate a hypothesis about how a specific change, such as a new marketing
campaign, pricing strategy, or product feature, is expected to influence demand.

Variable Isolation: Identify the variable (or variables) to be tested and isolate it from other factors that
could influence demand. This involves creating two or more versions (A and B) that differ only in the
variable of interest.

Random Assignment: Randomly assign a subset of the target market to each version (A and B). This
helps ensure that any differences in demand can be attributed to the changes being tested rather than other
external factors.

9
Implementation: Implement the changes in the respective versions and monitor the impact on demand
over a specified period. This could involve changes in pricing, advertising channels, website design, or
any other factor relevant to the demand being forecasted.

Data Collection: Collect data on key performance indicators (KPIs) such as sales, conversion rates, or
customer satisfaction for both versions during the experiment.

Statistical Analysis: Conduct statistical analysis to compare the performance of version A with version
B. This helps determine whether the observed differences are statistically significant.

Insights and Decision-Making: Draw insights from the results of the A/B experiment. If there is a
statistically significant difference in demand between the two versions, it provides evidence to inform
decision-making regarding future marketing strategies, pricing structures, or other relevant factors.

Iterative Testing: Iterate and conduct additional A/B experiments to refine and optimize strategies based
on ongoing market dynamics and consumer behavior.

B. Statistical Methods

Statistical methods include:

1. Trend projection: which is probably the easiest method of demand forecasting. Simply put, you
look at the past to predict the future. Of course, be sure to remove any anomalies. For example, if
you had a brief sales spike the previous year because a story about your product went viral for a
month, or your eCommerce site was hacked and sales temporarily dropped as customers heard the
news. Both of these events are unlikely to repeat, so they should not be factored into the trend
projection.

2. Regression technique or method of least squares

Regression Technique: which enables companies to identify and analyze the relationships between
different variables such as sales, conversions, and email signups. Taking a holistic view of how each
impacts the other can help a company allocate resources to the right area in order to boost sales.

The regression technique is popular as the method of best fit. In this case sales forecasts are made
assuming:
(i) That there is single determining variables that is sales or function of time

(ii) It is also assumed that there is a linear relationship between the variables.

Y = a + bx

Where:
10
Y = sales
a
= Y axis intercept i.e, constant
b = Coefficient of the determining variable x
x = time.

The normal equations for obtaining the values of a and b are:

∑ Y =Na+b ∑ x
∑XY= a∑x +b∑ x 2

Given sales data over a period of time the sales forecast can be made by applying the above equation and
solving it to obtain the values of a and b respectively.

Illustration

Using the sales data in Table 1 forecast the demand by use of linear regression.

Year Sales in ‘000 units

1995 25

1996 23

1997 30

1998 35

1999 43
11
2000 54

2001 49

2002 58
Solution:

Year Sales in Shs. (000) X X2 XxY

1995 25 1 1 25

1996 23 2 4 46

1997 30 3 9 90

1998 35 4 16 140

1999 43 5 25 215

2000 54 6 36 324

2001 49 7 49 343

2002 58 8 64 464

N=8
∑ y=317 ∑ x =36 ∑ x 2=204 ∑ xy=1647

Substituting these values in the normal equations

317 = 8a + 36b…………… (1)

1647 = 36a + 204b ………. (2)

Multiplying the equation (I) by 4.5 and subtracting it from equation (2)
1426.5 = 36a + 162b
1647 = 36a + 204b
220.5 = 0 + 42b

¿ ¿
b=220.5=
42 = 5.26

Substituting the values of b in equation (I)

1426.5 = 36a + 162 x 5.26
1426.5 = 36a + 850.5
576 = 36a
a = 16
12
Thus the trend equation is
Y =a+bx

Y=16 + 5.25x
Applying these values the sales for the years 2003 and 2004 forecasted

Sales in 2003 Y=16 + 5.25 x 9 = 16 + 47.25 = 63.25

Sales in 2004 Y=16 + 5.25 x 10 =16 + 52.5 = 68.5

Note that 9 and 10 are the 9th and 10th years, respectively.

Criteria/Qualities of a Good Demand Forecasting Method

The criteria/qualities of a good demand forecasting method in economics:

1. Accuracy: Accuracy denotes near to actual demand. A firm should forecast its demand very close to
the actual market demand so that required quantities could be made available for the market. Inaccurate
forecast may cost huge to the firm. It may create over or under production. Forecast should be explicit.
For example, there would be an increase in sales in the next year than the current is not a good forecast
but there would be an increase in sales by 20% in the next year is an accurate forecast.

2. Longevity or Durability: Demand forecast generally takes huge time, money and planning. Since a
forecast takes a lot of time and money, it should be usable for longer span of time or multiple years. A
forecast for short span of time may not be effective for the organization.

3. Flexibility or Scale-ability: A demand forecast should be flexible and adaptable to any kind of
changes. Now a days there is a rapid change in the tastes and preferences of consumers. This affects the
demand for different products up to a great extent. Therefore, the demand forecasts made by a firm should
be able to reflect those changes accordingly. Apart from this, a business firm, while making forecasts,
should consider various business risks that may take place in the future.

4. Acceptability and Simplicity: Acceptability is one of the most important criterion of a good demand
forecasting method. That means a forecast should be acceptable to all. It should also be as simple as
possible. A business firm should forecast its market demand by using simple and easy methods so that
the organizations do not face any complexities. However, some companies generally prefer advanced
statistical methods, which may prove difficult and complex.

5. Availability: A good a good demand forecasting method should have adequate and up-to-date data
available. The forecasts should be done in timely manner so that necessary arrangements could be made
related to the market demand. Data should be available to the decision makers at all time.

6. Plausibility and Possibility: It denotes that the demand forecasts should be reasonable, so that they are
easily understood by individuals who will use it. Again, it should have the quality of application in the
changing business conditions.

13
7. Economy: A good demand forecasting method should have a relationship with costs and benefits. It
should be economically effective. The forecasting should be made in such a way that the costs do not
exceed the benefits that will be derived from it. Costs should be less and benefits should be high.

8. Yielding quick results: good demand forecasting method should yield quick result rather than taking
longer period to respond. It should match with the changing business environment.

9) Maintenance of timeliness: It should take care of timelines. Data should be available to users as and
when requires so that decision making does not hamper.

14