Measuring the impact of a sales promotion campaign and analysing the ensuing results over a period of time represents a critical task for a product manager. This paper applies and compares three statistical approaches which assess the impact of promotions or marketing events on the sales of a product. The first approach is based on Winter’s method with an iterative technique to estimate a baseline. The second approach uses an extension of Winter’s exponential smoothing model which takes marketing events into account. Each sort of marketing event, a promotion or price reduction, for example, is linked to an index. Winter’s model includes a promotion index adjustment equation for each sort of promotion. Each equation is updated whenever the same sort of promotion occurs. This approach is suggested by Goodrich (1994). The third modelisation consists in modelling time series of sales contaminated with outliers due to the influence of marketing events from ARIMA models and intervention analysis. The procedure used to identify and describe AO (additive outlier), IO (innovation outlier), LS (level shift) and TC (temporary change) is the one suggested by Chang et al (1988) and Chen and Liu (1993). The data analysed concerns monthly consumption of orange juice and savoury cocktail biscuits which have been the object of promotion campaigns.
INDJEHAGOPIAN, J.P. et MACE, S. (1994). Etude comparative de mesures d’impact de promotion des ventes. ESSEC Business School.