Instacart is preparing to launch a new in-app substitutions engine that recommends replacement items when a shopper cannot find a requested product. The feature has shown promising early results in a limited internal pilot, and leadership wants a recommendation on whether to launch nationally before the holiday peak in 8 weeks. You are the program manager coordinating product, engineering, data science, operations, and legal across a 16-person cross-functional team.
The project matters because substitutions affect order completion, customer satisfaction, and refund costs. However, the available data is incomplete: the pilot only covered 3 mid-sized cities, one model input feed is known to be noisy, and operations leaders are concerned the feature could slow shoppers during peak demand.
The VP of Product wants a national launch this quarter to support revenue goals. The Head of Operations prefers delaying until shopper workflow impact is better understood. The Data Science lead believes the model is directionally strong but not fully calibrated. Legal requires updated customer messaging before any broad rollout.
You have 8 weeks until the holiday freeze, a remaining project budget of $180,000, and no approval for additional headcount. Engineering capacity is limited to 6 engineers because 2 are tied up on checkout reliability work. Legal review takes 10 business days, and the experimentation platform team can support only one major launch in this period.