goPuff model favors convenience over groceries
Former head of strategy at a global on-demand giant on the economics of grocery delivery
goPuff’s early surge showed that instant delivery worked best first as a digital convenience store, not as a full supermarket. Its model was strongest on small, urgent baskets like snacks, alcohol, nicotine, medicine, and household basics, where speed mattered more than broad selection. That let goPuff scale in the US before the later wave of dark store startups tried to stretch the model into larger weekly grocery baskets with tougher spoilage and picking economics.
-
The core dark store formula was simple, small urban warehouses, roughly a few thousand SKUs, very short delivery radii, and fast handoff to couriers. That setup favored repeat convenience purchases and lower spoilage, but it was built around limited assortment, not a full weekly shop.
-
The main operating lever was basket mix. Research on the category found mature ultrafast stores could work around a $25 order with positive contribution margin, but broader online grocery often needed closer to a $50 basket to absorb picking, delivery, and inventory costs. That is why moving from impulse items into fresh grocery was a much harder jump than it looked.
-
That early success also attracted a rush of copycats in Europe and the UK, where players like Gorillas, Getir, and others chased the same dark store playbook. Interviews across the category describe a discount war, falling AOV as customers ordered for convenience, and eventual consolidation pressure because too many nearly identical players were fighting for the same habit.
The next phase of instant delivery keeps pushing the same boundary, how far a convenience model can expand before it starts to inherit grocery’s harder economics. The winning companies are likely to be the ones that keep the speed and frequency of goPuff’s original use case, while layering in fresh food, ads, and subscriptions only where density and basket behavior support it.