↵
Home/Devs/Creator Analytics changed Discovery measurement: do not compare October 12 like an ordinary day
Devs

Creator Analytics changed Discovery measurement: do not compare October 12 like an ordinary day

Roblox now records Home, Search, and Charts impressions, clicks, and plays more reliably; counts may rise and PTR may fall without a real performance change.

Updated on10/21/2026·ELRobloxianoIA
Ejemplo oficial de métricas de adquisición por fuente; los valores son ilustrativos.

Official example of acquisition metrics by source; values are illustrative.

Available · Researched: October 21, 2026 (UTC). Sources: [official measurement announcement](https://devforum.roblox.com/t/improved-discovery-measurement-in-creator-analytics/4930122), [official weekly recap](https://devforum.roblox.com/t/weekly-recap-october-12-18-2026/4938638), and [Creator Updates](https://create.roblox.com/updates). AI-assisted writing.

Quick answer

Since October 12, 2026, Roblox has used a more reliable delivery pipeline to record impressions, clicks, and plays from Home, Search, and Charts. Creator Analytics may therefore show more events and a small decrease in play-through rate (PTR) even when player behavior has not worsened.

What changed

The new pipeline loses fewer events when the app closes or crashes. Roblox applies the change to every experience and did not change the system that decides which games appear in Recommended For You.

How to read your charts

  1. Mark October 12 as a measurement break.
  2. Compare pre-change periods with one another and post-change periods with one another.
  3. If impressions and plays rise together, do not automatically attribute the jump to a campaign or update.
  4. If PTR slips slightly, check whether impressions rose more than plays before deciding that the thumbnail became less effective.

What not to do

Do not combine a week before the change and a week after it as if they belonged to one homogeneous series. Roblox did not rewrite historical data, so the pipeline difference remains visible.

You also do not need to change code or settings to receive these measurements. The announcement says no creator action is required and expects minimal impact on relative recommendation position or benchmarks.

Success check

Add “Discovery measurement change: October 12, 2026” to your report and separate Home, Search, and Charts from other sources. This prevents a technical jump from being presented as confirmed growth.

Checklist: note the cutoff → separate surfaces → compare like periods → interpret PTR with impressions and plays → do not claim growth from the measurement change alone.

Related reading: a custom thumbnail is required for Home recommendation eligibility.

Was this helpful?

Keep exploring