Rejoinder for
The proper interpretation of sales promotion effects:
supplement elasticities with absolute sales effects I would like to express my appreciation to the editors for initiating this special issue, and to the two practitioners that have reacted to my article. The practitioners are very knowledgeable in this area, since they both have published in international top journals and they have worked for many years on sales promotion models. I concur with Tom Wilms that we should not express sales effects in terms of multipliers. Although these are already easier to interpret than sales elasticities, we need absolute sales effects to make proper comparisons between smaller and larger brands. Eijte Foekens' addition that managers want outcomes in terms of sales effects fully confirms my view.
I also have a number of specific recommendations 'to harvest the low hanging fruit in the sales promotion tree' (using Foekens' expression). First, I would focus on just short-term effects, since these are far stronger than long-term effects, at least for mature categories [1]. Consumers have been 'trained' to react to promotions for these categories, causing strong temporary category expansion effects yet no long-term effects (reacting to the second note by Tom Wilms). For new products that are promoted with free samples, promotions do tend to have long-term effects [2].
Second, within the set of short-term effects there is quite a clear ordering in the reliability with which effects can be estimated [3]. Own-item effects are estimated with lowest standard errors, followed by cross-item effects. These latter effects equal about 1 3 of the own effect [4]. Temporary category expansion effects and pre-and post-promotion dips are hard to capture although they seem to equal about 1 3 of their own effect [3]. It can be even harder to find significant storeswitching effects (which are part of the category expansion effect) [3]. In addition, cross-category effects are small and estimated with large standard errors [5]. So from a practical perspective, I would recommend focusing on own effects, cross-brand effects, category expansion effects, and pre-and post-promotion dips until marketing scientists have developed better models for more 'distant' promotion cross-effects such as those between categories or stores. I strongly agree with Eijte Foekens that it is quite interesting to study how the decomposition can be moderated by category role (traffic, impulse and routine categories) and category type (fresh versus packaged products).