Shipt launches 'Kits' feature in three weeks

Composable commerce typically depends on a limited set of primary entities that point to one another instead of copying data. There is a continual temptation to add a fresh entity for each emerging use case, which generally indicates that the current model should adapt rather than expand. If a capability must meet multiple separate constraints simultaneously—like determining the stores that can fulfill a shopper or identifying categories with additional eligibility criteria—modeling the issue as the intersection of independently evaluated sets proves more scalable than embedding conditional branches within one workflow.

Store-Aware Resolution and Kit Types

Behind every kit view is a calculation of which retailers can fulfill it. A customer’s eligible retailers form one set. Each kit has its own allow-list of retailers. Alcohol kits add a third layer of eligibility based on store licensing.

Once the effective retailers are known, the system runs concurrent searches against each catalog. Items are scored against actual inventory. Anchor items are checked. If a retailer call times out, that result is simply omitted rather than blocking the request. This keeps customer-perceived latency flat regardless of the number of retailers.

Kit types handle complex compliance rules. An alcohol kit’s lifecycle includes editorial creation plus store-eligibility checks and product-level filtering.

Three Weeks to Build a Store-Aware System

In June 2026, Shipt introduced Shipt Kits, a feature allowing customers to build their shopping carts with a single click, coinciding with a major international soccer tournament. Users could select a ‘Match of the Week’ kit, instantly adding a full evening’s groceries to their cart. Each item in the kit was tailored to the customer’s local store inventory, ensuring availability.

While it appeared seamless to the public, the development team had a tight three-week deadline. They completed the API in ten days and spent the remaining eleven days on the admin interface, user experience, and system stability.

Before Kits, building a themed shopping experience involved a manual assembly workflow. A merchandiser would query the data warehouse for products fitting a theme, pick relevant UPCs by hand, translate each UPC to a retailer-specific product ID, build a category, and then build a product list. For a national campaign requiring dozens of fresh kits across multiple retailers and weeks, this process was impossible.

Items are the atomic catalog units, generic, typed, and tagged. An item knows what kind of thing it is and carries arbitrary metadata in a JSON tag bag. Items are not retailer-specific; the resolution to a particular store’s product ID happens at view time. Shopping lists are curated sequences of items, which can be merchandiser-curated or customer-built. Kits are the composable bundles customers see, carrying editorial metadata and references to items.

A key feature is that items are shared across lists and kits, ensuring consistency. For example, the same ketchup item can appear in multiple kits and lists. Updating an item’s details automatically reflects across all references. Adding a new retailer for an item makes it available in all relevant kits and lists without manual intervention.