Heineken + Visualfabriq

Flexible trade promotion AI prediction models

Heineken’s Lars Nijland and Victor de Graaff share insights on how Heineken leverages Visualfabriq’s AI prediction models to enhance their trade promotion strategies. One of the standout features of Visualfabriq’s software is its ability to allow customers to build, train, and adjust their own prediction models, providing a level of flexibility and granularity that is rare in SaaS solutions.

How adaptability improves accuracy

Heineken has benefited greatly from this adaptability. It enables them to tailor the prediction models to their specific needs and market conditions. For example, by filtering data per retailer and SKU, and creating submodels that take into account factors such as promotion price elasticity and seasonality, they have seen significant improvements in the accuracy of their trade promotion forecasts. This has led to better decision-making and increased profitability.

Victor de Graaff, a freelance Data Scientist who has created prediction models for Heineken and Unilever, and Lars Nijland, Demand Planner and Project Lead for the Visualfabriq prediction model at Heineken Netherlands, discuss the three predictive models they use— in-market, ex-factory, and phasing —and how these models have outperformed manual predictions. They also highlight improvement topics such as factoring in upward or downward trends, and the ongoing benefits of using AI-driven predictions.

“What we see now is that (AI) predictions are better than manual predictions” − Victor de Graaff

Watch the presentation

Interested in learning more about Heineken’s experience with Visualfabriq? Watch the on-demand recording by filling out the form below.