Chapter 04

Estimating incremental sales

The first step towards calculating financial payback is to estimate the incremental sales generated by your activity.

To do this, you need to compare actual sales with 'base sales'; how much you would have sold if you hadn’t run the marketing activity in question. Incremental sales can be calculated by subtracting base sales from actual sales (the shaded area in Fig1).

Estimating incremental sales is by no means easy, but various common methods include:

  • Econometric modelling

    Econometric modelling is a mathematical technique that allows you to identify the various different factors that drive your sales, and to separate out their effects. For example 30% of IPA Effectiveness Awards used econometrics to measure the contribution of advertising to overall sales:

    For further information on econometric modelling, see Econometrics Explained, available from the IPA.

  • Test and control

    Another way of measuring the incremental sales effect is to look for some kind of control; for example a group of people or products that has not been exposed to the activity in question. The classic approach is the regional test. However, regional testing is not the only approach. For example, one might compare sales of advertised with non-advertised products. Or one might compare sales amongst individuals who were exposed to communications with sales amongst those who weren’t exposed.

  • Extrapolating from a trend

    In some cases, you may be able to plausibly argue that, without the marketing activity in question, sales would have remained static or continued along a certain trend. In that case, it may be valid to estimate the incremental effect by extrapolating from the trend. For example, the 2004 IPA Effectiveness Awards paper for The Guardian argued that, without the ’Fresh‘ campaign, the newspaper would have continued to lose market share.

    However, be sure to consider the other factors that were affecting your brand at the time. In order to use this method, you must be sure that other factors were not responsible for the deviation from the trend.