Our client already evaluated its promotional activity, based on simple estimations of volume sales uplifts on each SKU. It knew that there were several factors influencing these sales uplifts that they weren’t considering:
Because these factors were not being controlled for, it appeared that the effectiveness of promotions, even with the same discount and store position, varied randomly over time. This made planning promotions and predicting the expected volume, and revenue gains subject to large errors. Our client wished to reduce these errors, and potentially gain more insights allowing a more optimal promotional plan across a range of SKUs and accounts.