Our 2 previous posts (intro and part II) include an introduction to the 3-part incrementality series and the second example use case. Those two use cases we mentioned in our previous posts are the most obvious ones, and can be combined to solve more complex use cases, for example product replacement. The treasure trove of use cases is further increased by detaching the substitution matrix – how likely a product is substituted by another product – from the incrementality model, and combining that with other elements, such as demand estimation, shopping basket profiles, or product attribute matrices.
With incrementality, you no longer need to ignore those the interactions between products, which is the real crown jewel of this treasure chest.
- Evaluating customer satisfaction, or service level, by calculating the likelihood that the customer demand is satisfied, given available assortment. This can be calculated per store, per product category, or even for each individual customer.
- Finding optimal size of the assortment, for a given service level.
- Ranking, and protecting, key products in the assortment. Dropping these hard-to-substitute items from the assortment may have drastic effects because the potential customer may take their entire shopping basket to another store.
- Forecasting sales for new products, based on common product attributes.
- Evaluating the effects of a supplier change, for example, when a supplier product is being replaced by a private label product. How are the overall sales affected, and what does the change mean for each individual supplier?
- Creating new product groupings, or entire product trees, based on how products are related to each other in terms of incrementality.
Many of these use cases can be answered with traditional sales forecasting models, but in that case the interactions between products are ignored. With incrementality, you no longer need to ignore those effects, which is the real crown jewel of this treasure chest.