Roblox Recommended for You: A Verified 28-Day Creator Audit

Roblox began rolling out changes to its Recommended for You algorithm on June 15, 2026. The update expanded familiar recommendation signals from seven days to 28 days and divided their evaluation into three windows: Day.

Roblox Recommended for You: A Verified 28-Day Creator Audit

Roblox began rolling out changes to its Recommended for You algorithm on June 15, 2026. The update expanded familiar recommendation signals from seven days to 28 days and divided their evaluation into three windows: Day 1, Days 2–7, and Days 8–28.

Verified recommendation signals

Roblox named six signal families evaluated across those windows:

• Playtime
• Play days
• Qualified play sessions
• Intentional co-play days
• Spend days
• Robux spent

Roblox states that each signal is based on users who joined organically through the Recommended for You sort on Home.

The update also removed qualified play-through rate, or QPTR, and introduced separate play-through rate and first-play bounce-rate signals. Roblox says this separation allows play-through, initial session quality, and retention to be measured independently.

Creators can inspect the updated signals in Creator Analytics → Acquisition → Home Recommendations.

A conservative audit workflow

Use the three evaluation windows as an organizational framework, while avoiding assumptions that any single metric proves why player behavior changed:

1. Review the Home Recommendations signals for the relevant acquisition cohort.
2. For Day 1 concerns, examine the core loop, first-time user experience, and technical performance.
3. For Days 2–7 concerns, inspect progression goals, rewards, content variety, and difficulty pacing.
4. For longer-term concerns, inspect the experience's ending systems, recurring content, and social mechanics.
5. Compare cohorts associated with major updates, events, or marketing campaigns.
6. Treat benchmark scorecards as comparison tools, not recommendation inputs.

This workflow follows Roblox's retention guidance. The Creator Hub recommends focusing Day 1 improvements on the core loop, first-time user experience, and performance; Day 7 improvements on progression systems; and Day 30 improvements on ending systems. Its retention charts organize cohorts by each user's first-play date, and Roblox says cohorts can be used to examine the effects of updates, events, and marketing campaigns.

Interpret distribution carefully

Roblox warned that distribution could fluctuate during the rollout. Its announcement identifies four influences on Home recommendation impressions: creator changes, Roblox or discovery-system changes, seasonal variation in Home traffic, and the relative performance of other experiences.

An impressions change should therefore be investigated alongside release timing, acquisition source, retention cohorts, and technical data rather than attributed automatically to one design change.

Keep benchmarks separate

Creator Analytics provides benchmark scorecards for comparison with similar experiences. Roblox explicitly states that these benchmarks are not used as direct signals in the discovery algorithm. They can help identify metrics worth investigating, but they should not be presented as Recommended for You inputs.