Sales forecasting: scatter graphs, lines of best fit and extrapolation

Scatter graphs and correlation

A scatter graph plots pairs of values, such as advertising spending (x) against sales (y), or time (x) against sales (y).

• Positive correlation: as one rises, the other tends to rise.
• Negative correlation: as one rises, the other tends to fall (price and quantity sold).
• No correlation: no clear pattern.

Correlation does not prove causation: a third factor (such as rising incomes) may drive both variables.

Line of best fit

A line of best fit is a straight line drawn through the middle of the points so that the points are spread as evenly as possible on both sides. In the exam it is drawn by eye; it should follow the trend, not join the dots. If the points sit close to the line, the relationship is strong and forecasts are more reliable.

Extrapolation

Extrapolation extends the line of best fit beyond the existing data to forecast future values. A quick numerical version uses the average change per period: (last value - first value) / number of intervals.

Example: sales ($m) for years 1 to 6 were 20, 23, 25, 29, 31, 34. Average change = (34 - 20) / 5 = 2.8 per year. Forecast for year 8 = 34 + 2 x 2.8 = 39.6.

A least-squares line through the same points (y = 17.20 + 2.80x) gives 39.6 for year 8, the same as the average-change forecast here; with less regular data the two methods differ.

Limitations

• Assumes the past trend continues; markets change (new rivals, recession, changing tastes, technology).
• The further ahead the forecast, the less reliable it is.
• A line drawn by eye is subjective.
• Correlation may be weak or caused by another factor.
• Ignores qualitative information such as market research and managers' experience.

In the January 2025 WBS13 paper, Question 1(c) asked candidates to assess quantitative sales forecasting; the mark scheme noted that actual premium subscribers came in above the forecast and that forecasts are less suitable for dynamic technology markets.

Worked example

Revenue ($m) for years 1 to 5 was 48, 52, 57, 60, 66. Use the average change per year to forecast year 7, and state one limitation.

1. Average change = (66 - 48) / 4 = 4.5 per year.
2. Year 7 is two years after year 5: 66 + 2 x 4.5 = 75.
3. Limitation: it assumes the trend continues; a recession or a new rival would make the forecast wrong.

Answer: About $75m, assuming the past trend continues.

Check yourself

1. Sales ($m) for years 1 to 6 were 20, 23, 25, 29, 31, 34. Calculate the average change per year.
   Show answer

   $2.8 m. (34 - 20) / 5 = 2.8.

2. Using an average increase of $2.8m a year from year 6 sales of $34m, forecast sales for year 8.
   Show answer

   $39.6 m. 34 + 2 x 2.8 = 39.6.

3. Revenue ($m) for years 1 to 5 was 48, 52, 57, 60, 66. Use the average change per year to forecast year 7.
   Show answer

   $75.0 m. Average change = (66 - 48) / 4 = 4.5; year 7 = 66 + 2 x 4.5 = 75.

4. A scatter graph of price (x) against quantity sold (y) shows points sloping down from left to right. This shows
   • positive correlation
   • negative correlation
   • no correlation
   • that price causes all changes in sales
   Show answer

   negative correlation. As price rises, quantity sold tends to fall.

5. Ice cream sales and sunglasses sales are strongly positively correlated. What is the best conclusion?
   • Buying ice cream causes people to buy sunglasses
   • Sunglasses sales cause ice cream sales
   • The data must be wrong
   • A third factor, such as hot weather, probably drives both
   Show answer

   A third factor, such as hot weather, probably drives both. Correlation does not prove causation.

6. Why is a forecast for five years ahead usually less reliable than one for next year?
   • Lines of best fit cannot be extended that far
   • There is more time for conditions in the market to change
   • Sales always fall in the long run
   • Moving averages cannot be calculated
   Show answer

   There is more time for conditions in the market to change. The assumption that the trend continues becomes weaker.

7. Explain one limitation of extrapolation as a way of forecasting sales.
   Show answer

   It assumes the past trend will continue, but the market may change, for example a new competitor or a recession, so the forecast may be inaccurate.. State the assumption, what could change, and the consequence.

Practise with instant feedback Print this lesson

Sources

• Pearson Edexcel International Advanced Subsidiary/Advanced Level in Business specification (XBS11/YBS11, 2018), Unit 3 Business decisions and strategy and Appendices 6, 7 and 9
• Pearson Edexcel IAL Business WBS13/01 mark scheme, January 2025

Last updated 8 November 2026. Spotted a mistake? Tell us.