Scaling Automated Collections

As more merchants adopted metafields, they hit a bottleneck: they couldn't use metafields to build automated collections. That meant:

• Duplicate data entry. Merchants copied the same information into variant options, tags, and metafields just so products would be sorted correctly.
• No single source of truth. With data split across three places, it was hard to know which one was accurate.
• A ceiling on adoption. The manual work kept merchants from relying on metafields as a way to organize their stores at scale.

Automated collection conditions could use built-in product fields like tags, type, and vendor, but not metafields. Supporting metafields as a condition would:

• Stop duplicating data across variant options, tags, and metafields
• Cut the manual work of re-creating that data so conditions could read it
• Encourage metafield adoption among merchants who had not started yet

I worked closely with engineering, product, the Polaris design system team, and the Merchandising team, which owned collections, so the feature would be easy to use, scale well, and stay consistent with existing patterns in the admin.

• Grounding the work in merchant feedback. I reviewed support tickets and merchant feedback to understand how merchants were working around the gap and where they got stuck.
• Rebuilding the condition dropdown. I replaced the default browser dropdowns for both the field and the relation with a custom component, reviewed with Polaris, Shopify's design system team. It added search and a labelled Metafields section, so merchants could find the right field quickly, even with a long list.
• Designing around technical limits and performance. Only some metafield types could be supported as conditions, so I designed the experience around that. Merchants turned on eligible definitions one at a time, up to 128 per store, and a clear warning explained what to do when they reached the limit.
• Handling deletion dependencies. Merchants couldn't delete a metafield definition while a collection still used it. A warning listed each affected collection by name, so they knew exactly where to remove the condition first.
• Documenting for handoff. I created annotated handoff specs for web and mobile, with callouts on each change and interaction, so engineering could build each update without guessing.

The patterns we set here became the foundation for bringing metafields to other parts of the admin, starting with filtering in the product index table.

The feature was officially released as part of Shopify Winter Editions 2023. The launch eliminated redundant data workflows and allowed merchants to automate collections using metafields in conditions for the first time.

• 3.17M active smart collections using metafield rules
• 77.8K shops actively using the feature as of March 2026
• 16× growth in collection creation over 15 months