Quantifying promotion effectiveness for a retailer

Abstract

Retailers spend a substantial part of their marketing budgets on promotions. While promotions already account for a large share of retailers’ revenues, promotion intensity continues to increase. Despite substantial investments in promotions, companies often have little understanding of the true performance and thus struggle to determine which promotions are working and which are not.

The objective of this thesis is to quantify the impact of sales promotions, which can help retailers to identify the most and the least effective promotions. This thesis addresses the research problem by seeking answers to three key questions. First, it studies whether promotions generate a lift in sales during the promotional period. Second, the extent to which the sales lift may cannibalize the sales of other non-promoted products in that category is determined. Lastly, this thesis investigates whether promotions affect the quantities sold in the weeks after the promotion.

The promotion effects were quantified using a 2-year dataset of weekly point-of-sales (POS) data obtained from a Finnish retailer. During the study period, over 140,000 promotions were implemented in nearly 500 different product categories.

The main findings of this thesis are threefold. First, about three quarters of promotions generate additional sales. The magnitude of this sales lift is about 130% in comparison to normal level of sales. The interquartile range of the sales lift measure falls between 17% and 381% indicating high variation among products. Second, approximately 40% of the sales lift accounts to switching from other non-promoted products in the same category. This indicates that on average only about 60% of the additional sales are truly incremental to retailers. Lastly, this thesis did not find evidence for stockpiling effects which refers to decreased quantities sold in the weeks after the promotion period.

The results imply that not all promotions generate the desired results and that the effectiveness of promotions is significantly diluted due to switching. This means that retailers should focus promotional activities to products and categories that are more effective and generate a positive impact. Systematic analysis based on quantitative data applied in this thesis helps retailers to gain insights from a large set of data and understand the true performance of promotions while guiding effective promotional decisions in the future.