qualitative sales forecasting

8.1 Marketing Analysis – Sales Forecasting

8.1.1 Purpose of a Sales Forecast

• Provides the quantitative basis for setting realistic marketing objectives (sales, market‑share, profit, budget).
• Helps allocate resources – advertising spend, distribution channels, staffing levels.
• Informs product‑development decisions – size of launch, R&D investment, production capacity.
• Acts as a benchmark for performance monitoring and control.

8.1.2 Demand‑Side and Supply‑Side Drivers

Demand‑side drivers	Supply‑side constraints
Price of the product and of substitutes. Consumer income and purchasing power. Tastes, preferences and lifestyle trends. Population size, demographics and geographic concentration. Seasonality and weather patterns. Advertising, promotion and brand perception.	Production capacity and plant utilisation. Availability and cost of raw materials. Technology and automation levels. Supplier reliability and lead‑times. Distribution network efficiency. Regulatory or legal restrictions.

8.1.3 Forecasting Methods

Quantitative Forecasting (requires reliable historical data)

Statistical techniques that generate a numerical forecast.

• Time‑Series Methods
  • Moving Average (MA) – smooths random fluctuations. $$\text{MA}_{t}= \frac{1}{n}\sum_{i=0}^{n-1} S_{t-i}$$ Limitation: Ignores trend and seasonality; sensitive to the choice of n.
  • Exponential Smoothing (ES) – gives more weight to recent periods. $$F_{t+1}= \alpha S_{t}+ (1-\alpha)F_{t}$$ Limitation: Requires a suitable smoothing constant α; not ideal when data contain strong trend or seasonality unless modified (e.g., Holt’s method).
• Regression Analysis – relates sales to one or more independent variables (price, advertising spend, income, etc.).

  Simple linear regression:

  $$\hat{Y}= a + bX$$ where \(\hat{Y}\) = forecast sales, \(X\) = explanatory variable, \(b\) = slope (often the price‑elasticity coefficient), \(a\) = intercept. Limitation: Assumes a linear relationship and that the chosen variables remain stable; outliers can distort the slope.
• Trend Projection – fits a straight line or curve to historical sales and extrapolates. Limitation: Assumes the identified trend will continue; can over‑estimate when the market is approaching saturation.

Checking Forecast Accuracy

After a forecast is produced, its reliability is assessed using simple error‑measurement tools required by the Cambridge syllabus.

Measure	Formula	Interpretation
Mean Absolute Deviation (MAD)	$$\text{MAD}= \frac{1}{n}\sum_{t=1}^{n}\big|A_{t}-F_{t}\big|$$	Average absolute forecast error – lower MAD = more accurate.
Mean Absolute Percentage Error (MAPE)	$$\text{MAPE}= \frac{100\%}{n}\sum_{t=1}^{n}\frac{\big|A_{t}-F_{t}\big|}{A_{t}}$$	Expresses error as a percentage of actual sales; useful for comparing forecasts of different sizes.
Tracking Signal	$$\text{TS}= \frac{\sum_{t=1}^{n}(A_{t}-F_{t})}{\text{MAD}}$$	Indicates whether a forecast is consistently biased (TS > ±4 suggests a problem).

Qualitative Forecasting (judgement‑based)

Used when data are scarce, the product is new, or the market is rapidly changing.

1. Delphi Method – iterative questionnaire rounds with an expert panel; anonymity reduces group‑think.
2. Sales‑Force Opinion – front‑line staff estimate demand from customer contacts.
3. Market‑Research Surveys – consumer questionnaires or focus groups that capture purchase intentions.
4. Executive Opinion – senior managers apply strategic insight.
5. Historical Analogy – compare the new product with a previously launched, similar product.

Steps in Conducting a Qualitative Forecast

1. Define the forecasting horizon (short‑term ≤ 12 months, medium‑term 1‑3 years, long‑term > 3 years).
2. Select the most appropriate technique(s) for the product and market.
3. Identify and brief participants (experts, sales staff, consumers).
4. Collect opinions using questionnaires, interviews or workshops.
5. Summarise responses, look for common themes and calculate a consensus figure (average, median or weighted average).
6. Present the forecast as a single figure, a range, or a set of scenarios (optimistic, realistic, pessimistic).
7. Review and update regularly as new data become available.

Advantages & Disadvantages of Qualitative Methods

Advantages	Disadvantages
Quick to produce when historical data are unavailable. Can incorporate non‑quantifiable factors (brand perception, upcoming legislation). Useful for new‑product launches and volatile markets. Relatively low cost compared with extensive data collection.	Highly subjective – risk of personal bias. Limited statistical reliability; difficult to test accuracy. Depends on the expertise and honesty of participants. Hard to validate against actual outcomes.

8.1.4 Elasticity and Its Use in Forecasting

Elasticities measure how a change in one variable (price, income, promotion) influences the quantity demanded.

• Price Elasticity of Demand (PED) $$\text{PED}= \frac{\%\Delta Q}{\%\Delta P}$$ If |PED| > 1 demand is elastic; if |PED| < 1 demand is inelastic.
• Income Elasticity of Demand (YED) $$\text{YED}= \frac{\%\Delta Q}{\%\Delta I}$$ Positive YED > 0 indicates a normal good; YED < 0 indicates an inferior good.
• Advertising (Promotional) Elasticity (AED) $$\text{AED}= \frac{\%\Delta Q}{\%\Delta A}$$ Shows the percentage change in sales for a 1 % change in advertising spend.

Applying Elasticities to a Forecast

1. Start with a base quantitative forecast (e.g., 130 000 units).
2. Adjust for planned price change: $$\text{Adjusted Forecast}=130\,000 \times \bigl(1+ \text{PED}\times\%\Delta P\bigr)$$
3. Further adjust for expected income growth or advertising budget using YED and AED in the same way.

Example: Base forecast = 130 000 units. Planned price reduction = 5 % and PED = ‑1.4. Adjusted forecast = 130 000 × [1 + (‑1.4 × ‑0.05)] = 130 000 × 1.07 ≈ 139 200 units.

8.1.5 Linking Forecasts to Product Development

• Scale of launch – the forecast determines the initial production volume and the size of the first‑year budget for raw materials and labour.
• R&D investment – a higher forecast justifies greater spending on product refinement, packaging design, or new features.
• Timing of introduction – seasonal forecasts help decide whether to launch in a high‑demand period (e.g., summer for soft drinks) or to stagger releases.
• Capacity planning – forecasts are compared with existing plant utilisation; if demand exceeds capacity, the firm may plan plant expansion or outsource production.

8.1.6 Market Research – Primary & Secondary Data

Both types of data feed into quantitative and qualitative forecasts.

• Primary Research – data collected specifically for the forecast.
  • Surveys, questionnaires, face‑to‑face interviews.
  • Focus groups and observation.
  • Test‑market sales figures.
• Secondary Research – existing data not originally collected for the forecast.
  • Industry reports, government statistics, trade publications.
  • Competitor annual accounts and market‑share data.
  • Historical sales data from the company’s own records.

From Research to a Numerical Forecast

1. Convert purchase‑intention percentages into units: $$\text{Forecast Units}= \frac{\text{% intending to buy}\times \text{Target Population}}{100}$$
2. Use secondary market‑size figures to estimate total market demand, then apply a desired market‑share target to obtain a sales figure.
3. Adjust the raw figure for known supply‑side constraints (capacity, lead‑time) and for elasticity effects (price, income, promotion).

Sampling, Reliability & Bias

The quality of primary data depends on the sampling method.

Sampling Issue	Effect on Forecast
Non‑representative sample (e.g., only high‑income respondents)	Skewed demand estimate; over‑ or under‑forecasting.
Small sample size	Low reliability; high margin of error.
Response bias (social desirability, leading questions)	Inflated purchase‑intention figures.
Sampling error (random variation)	Uncertainty range that should be reflected in the forecast.

To improve reliability:

• Use stratified random sampling to reflect key market segments.
• Pilot test questionnaires to remove ambiguous wording.
• Calculate confidence intervals (e.g., 95 % confidence level) and report them with the forecast.

8.1.7 Illustrative Examples

Quantitative Example – 4‑Period Centre Moving Average

Monthly sales (000 units) for the past six months: 112, 118, 124, 130, 136, 142.

4‑period centred moving average for month 4:

$$\text{MA}_{4}= \frac{112+118+124+130}{4}=121\text{ (000 units)}$$

The marketing team sets a short‑term sales objective of 122 000 units for month 5 and uses the figure as a baseline for budgeting.

Quantitative Example – Applying Price Elasticity

Base forecast = 130 000 units. Planned price increase = 4 % and PED = ‑1.2.

$$\text{Adjusted Forecast}=130\,000 \times \bigl[1+(-1.2\times0.04)\bigr]=130\,000 \times 0.952 = 123\,760\text{ units}$$

Qualitative Example – Delphi Method

A beverage company runs two Delphi rounds with five industry experts to forecast a new eco‑friendly sports drink.

Round 1 forecasts (thousands of units): 120, 135, 150, 140, 130.

After feedback, Round 2 forecasts: 138, 145, 152, 148, 140.

Consensus forecast (average of Round 2):

$$\frac{138+145+152+148+140}{5}=144.6\text{ thousand units}$$

The firm plans production of 145 000 units for the first year and targets a 4 % market share of a 3.6 million‑unit market.

8.1.8 Choosing the Right Approach

• Quantitative – established products, abundant reliable historical data, need for statistical accuracy.
• Qualitative – new‑product launches, entry into unfamiliar markets, rapid decision‑making, volatile external environment.
• Hybrid (Quantitative + Qualitative) – combine a time‑series forecast with expert adjustments to reflect upcoming market changes (e.g., competitor entry, regulatory shift).

8.1.9 Key Take‑aways

• Sales forecasts translate market information into concrete marketing objectives, budgets, and product‑development plans.
• Both demand‑side drivers (price, income, tastes, etc.) and supply‑side constraints (capacity, raw‑materials) must be considered.
• Quantitative methods rely on historical data; each technique has specific limitations and should be checked with accuracy measures (MAD, MAPE, tracking signal).
• Qualitative methods rely on expert judgement; they are essential for new products or rapidly changing markets but must be documented and periodically reviewed.
• Elasticity coefficients allow forecasters to model the impact of price, income and promotional changes on sales.
• Primary and secondary market research provides the raw data; proper sampling, reliability checks and conversion of percentages to units are crucial for a credible forecast.
• Regular monitoring, updating and, where appropriate, blending of quantitative and qualitative techniques give the most accurate and flexible forecasts for A‑Level business planning.