Cache

Cache attempted to solve the last-mile delivery economics problem for convenience items by building tiny, fully automated dark stores that could fulfill orders placed through DoorDash and UberEats. Urban consumers experienced acute friction—they wanted snacks, drinks, and household essentials delivered quickly but existing convenience stores couldn't operate profitably at delivery scale. The problem was measurable: delivery fees often exceeded product costs, making 15-minute delivery economically unsustainable for low-margin items. However, Cache missed critical warning signs. Competitors like Gopuff and Instacart already dominated convenience delivery with less capital-intensive models. The automation technology, while innovative, required significant upfront investment per location with unproven unit economics. Cache's reliance on DoorDash and UberEats meant they had no direct customer relationship and faced commission rates that squeezed margins further. The fundamental issue wasn't solvability—it was that the problem's economics didn't support the proposed solution. By betting heavily on automation rather than validating demand and unit economics first, Cache built an elegant answer to a question the market hadn't asked.

Cache launched on DoorDash and UberEats in 2021, targeting customers seeking ultra-fast grocery delivery through automated micro-fulfillment centers. Early behavioral signals looked promising—users repeatedly opened the app and browsed inventory, with strong click-through rates on staple items like beverages and snacks. The team measured genuine interest through conversion rates and repeat order frequency, which initially exceeded industry benchmarks for new delivery services. However, early traction masked fundamental problems. While users *browsed* frequently, they rarely completed purchases at profitable margins. The company discovered that stated interest ("I want 15-minute grocery delivery") diverged sharply from actual purchasing behavior—customers preferred established players like Instacart for serious shopping and used Cache only opportunistically. The automated dark stores, positioned as a cost advantage, became a liability when inventory mismatches frustrated customers seeking specific items. The critical warning sign emerged too late: high engagement metrics didn't translate to unit economics that worked. Cache confused traffic with demand, mistaking curiosity for genuine product-market fit before the model collapsed.