
The most common mistakes in GenAI adoption
Why organizations that start by asking "how do we automate this?" often end up frustrated and what to ask instead.
What transaction-level data reveals about profitable omnichannel design

Online grocery profitability hinges on a small number of high-impact design choices, most notably delivery fee pricing and online assortment size. Using millions of transaction-level observations from a large grocery retailer operating both online and physical stores, this analysis examines how customers trade off convenience, variety, and cost when deciding where to shop. These behavioral insights are then embedded into a prescriptive AI model to identify which online service configurations are truly profitable.
The results challenge several common practices in omnichannel retail:
These findings come from a study on delivery fee pricing and online assortment design in omnichannel grocery retail, which I coauthored with Daniela Fernandes, Fábio Neves-Moreira, and Jan Fransoo, and which is published in the European Journal of Operational Research.
Overall, the study shows that online grocery main levers should not be optimized in isolation. For omnichannel retailers, delivery fees and assortment size are strategic decisions that determine how demand is allocated across channels and how value is created across the entire retail system.