FEATURECross-Content Recommendation BlocksRead more
← Back to Changelog

Smarter Ranking for Interaction Recommendations

Two recommendation panels side by side: raw confidence ranking a same-category jacket first, versus lift score ranking complementary products first and labeling them

Better recommendations, no setup required.

Frequently Bought Together, Viewed Then Bought, and other behavior-based blocks now rank results by lift, not raw popularity. Lift measures how much more likely shoppers are to buy or view two products together than chance alone predicts. Recent activity counts more than old activity.

New products used to sit in empty blocks until they built up their own history. Now Layers backfills with strong performers from the same category, then a broader category, so every anchor shows useful recommendations right away.

  • Complements over duplicates - Frequently Bought Together, Viewed Then Bought, and Customers Who Bought Also Bought now favor complements over near-duplicates from the anchor's own category.
  • Relationship labels - Each interaction strategy now shows whether it surfaces a Complement, a Substitute, or a Sequence, right in the block editor.
  • Old ordering still available - Set order_by to confidence_score or interaction_count in strategy_config if you'd rather sort by raw co-occurrence.

This is on by default for existing interaction blocks. No configuration changes needed.

Learn more

Subscribe to the changelog

Releases land wherever your team already reads.

Slack

Subscribe to changelog

1

Choose a Slack channel

Open the Slack channel where releases should land. Most teams use #product-updates.
2

Subscribe to releases

Paste this command into the Slack channel.

/feed subscribe https://www.uselayers.com/changelog/rss.xml
3

Approve the confirmation post

Approve the confirmation Slack post. From then on each release arrives as a message with its title, date, and link.