Roblox Analytics Changed on October 12, 2026: Why Your Impressions and Play-Through Rate Moved, and What to Do About It
Posted: Tue Oct 20, 2026 2:13 pm
Quick Summary
The update covers three surfaces: Home, Search, and Charts. The events in question are impressions (your game card shown to someone), clicks, and plays (someone who clicked actually loading in). Before this change, some of those events were lost when a client crashed or the app closed before the telemetry could be sent. After it, they come through a delivery pipeline that is less likely to drop them.[1]
That is the whole mechanical change. Nothing in the announcement describes a new ranking rule, a new signal, or a change to what the algorithm rewards. Roblox frames it purely as a measurement fix, and the phrasing matters: they say fewer events are lost, not that more people are playing your game.[1]
A note on dates. The changelog entry is dated October 12, and the developer forum announcement was posted the same day. If you are looking for a single cutover moment in your own charts, October 12 is it.[1][6]
Why your play-through rate went down
Play-through rate is plays divided by impressions. Roblox's explanation is arithmetic rather than behavioral: if impressions rise faster than plays, the ratio falls even though nothing about your game changed.[1]
Roblox says the shift is similar across games, so comparisons between your game and a benchmark, and your ranking relative to other games, should hold. Their FAQ goes further and states that before and after the update every game was measured the same way, so benchmarks and recommendations stayed consistent for everyone.[1][4]
The practical reading: a lower PTR on October 13 does not mean your thumbnail got worse. It means the denominator grew.
What the creators in the thread are reporting
The announcement thread has 5.5k views, 75 likes, and 13 users involved, and a good chunk of the replies describe something other than higher impressions.[1]
What this means for Recommended For You
RFY runs in two stages. Retrieval picks a candidate set of games for a user based on engagement, retention, and monetization signals. Ranking then orders that set and decides what to show. Crucially, only users who arrive through RFY count toward the ranking signals. Traffic from ads, friends, search, or social sharing can help you enter the candidate pool, but it does not count in ranking.[3]
The signals that matter most, in Roblox's stated priority order:[3]
Where this connects to the October 12 change: PTR is the top-weighted signal, and PTR is exactly the number that shifted. But Roblox is explicit that the update improves how events are recorded and does not change how RFY recommends games, and that they expect minimal impact on Home recommendations because the shift is similar across games.[1]
So the metric moved. The weighting did not.
How to read your analytics across the cutover
The table below separates the two kinds of action this change creates.
On the acquisition side, the source list worth knowing: Home recommendations, Continue play, Today's Picks, Home other, Friends, Search, Charts, Sponsored ads, Search ads, Portal ads, Teleport, and Other. The Acquisition dashboard also reports conversion rate, D7 retention, 7D playtime per user, 30D payer conversion, and 30D revenue per user, so you can see whether a source brings players who stick and spend, not just players who click once.[5]
One trap worth naming: if your game has Exclude from Recommendations enabled, Home recommendation metrics read as zero, because games excluded from recommendations cannot receive traffic from it. That is not a bug and it is not the October 12 change.[5]
What players should know
If you do not build games, none of this touches you directly. Your experience of Roblox does not change: the same games appear on Home, the same searches work, and the same charts rank the same way.
What can change is the health of the games you play. When a creator misreads a measurement artifact as a real decline, the common reaction is to panic and change a thumbnail, a title, or the first minute of the game. That is the failure mode to watch for, and it is the argument for the slow read Roblox recommends: compare periods before and after October 12 against each other instead of reacting to a single day.[1]
FAQ
Did Roblox change the algorithm on October 12, 2026?
No. Roblox changed how impressions, clicks, and plays from Home, Search, and Charts are recorded, moving them to a more reliable delivery pipeline. They state the update does not change how Recommended For You recommends games, and they expect minimal impact on Home recommendations because the shift is similar across games.[1]
Will my game rank worse because my play-through rate dropped?
Roblox says no, because the drop comes from impressions rising faster than plays and the change is similar across games, so comparisons and benchmarks hold. If your benchmark did not move the way your own PTR did, that is worth a bug report, since it contradicts the stated expectation.[1][4]
Should I re-baseline my analytics?
Yes. Roblox explicitly suggests comparing periods before October 12, 2026 against periods after it to understand the impact, rather than reading a single day or a single number.[1]
Does this change my historical data?
No. Data from before October 12, 2026 stays exactly as it was. Nothing gets rewritten.[1]
Is my PTR drop caused by worse thumbnails?
Probably not, on its own. Roblox attributes the PTR decrease to the measurement change, not to thumbnail quality. The useful test is your benchmark: if your benchmark moved the same way, the drop is measurement. If it did not, investigate further.[1][4]
Does this affect sponsored ads or search ads?
The announcement covers Home, Search, and Charts, and the acquisition dashboard lists Sponsored ads, Search ads, and Portal ads as separate sources, each with its own metrics. Roblox does not state that ad measurement was reworked here, so treat ad numbers as their own thing and compare them on their own timeline.[1][5]
Why do some creators see higher impressions and others see lower?
Roblox predicted higher impressions and play counts. Multiple creators in the announcement thread report lower counts and PTR drops larger than the small decrease described, and at least one reported a bounce rate spike. The thread had not produced a staff answer explaining the gap when this was written, so both readings are documented here rather than picking one.[1]
Sources
- On October 12, 2026, Roblox moved how impressions, clicks, and plays from Home, Search, and Charts get recorded. The reason Roblox gave is a more reliable delivery pipeline, so fewer events vanish when the app closes or crashes.[1]
- Expect higher impression, click, and play counts from those surfaces, plus a small drop in play-through rate (PTR), because impressions can rise more than plays.[1]
- Historical data does not change. Roblox's advice is to compare the period before October 12 against the period after it, not to read a single daily number.[1]
- Your Recommended For You (RFY) ranking does not change with this update. Play-through rate, first play bounce rate, play days per user, and playtime per user carry the most weight there, and they still measure behavior after a player finds your game.[3][4]
- Plenty of creators on the announcement thread report the opposite of what the FAQ predicted, including lower impressions and bounce rate spikes. Treat your own before and after comparison as the only number that matters.[1]
The update covers three surfaces: Home, Search, and Charts. The events in question are impressions (your game card shown to someone), clicks, and plays (someone who clicked actually loading in). Before this change, some of those events were lost when a client crashed or the app closed before the telemetry could be sent. After it, they come through a delivery pipeline that is less likely to drop them.[1]
That is the whole mechanical change. Nothing in the announcement describes a new ranking rule, a new signal, or a change to what the algorithm rewards. Roblox frames it purely as a measurement fix, and the phrasing matters: they say fewer events are lost, not that more people are playing your game.[1]
A note on dates. The changelog entry is dated October 12, and the developer forum announcement was posted the same day. If you are looking for a single cutover moment in your own charts, October 12 is it.[1][6]
Why your play-through rate went down
Play-through rate is plays divided by impressions. Roblox's explanation is arithmetic rather than behavioral: if impressions rise faster than plays, the ratio falls even though nothing about your game changed.[1]
Roblox says the shift is similar across games, so comparisons between your game and a benchmark, and your ranking relative to other games, should hold. Their FAQ goes further and states that before and after the update every game was measured the same way, so benchmarks and recommendations stayed consistent for everyone.[1][4]
The practical reading: a lower PTR on October 13 does not mean your thumbnail got worse. It means the denominator grew.
What the creators in the thread are reporting
The announcement thread has 5.5k views, 75 likes, and 13 users involved, and a good chunk of the replies describe something other than higher impressions.[1]
- CenturyofGaming wrote that they saw lower impressions and play counts since October 12, with PTR down as the only consistent change. EmptySet replied that a lot of their stats dropped suddenly right after the change and said, and I am quoting them directly here, Its clearly not improving discovery from what I have seen.[1]
- EmptySet also reported that bounce rate spiked at the same time, with no game changes on their end, and wondered whether new-player server teleports were being counted as bounces, or whether teleport failures were suddenly tracked more accurately. They noted the 61 to 180 second rate did not move the same way.[1]
- HatlessBanjo, who owns a game averaging around 7k CCU, posted a chart showing a 26 percent drop in PTR and a drop in QPTR from personalized thumbnails, while the benchmark showed no decline at all. They asked the obvious question: if the change is happening to everyone, why did the benchmark not move? Their other game, One Tap, which runs a higher CCU, barely changed.[1]
- Ghaucho, a game designer, used the thread to ask for the reach and highly engaged players analytics to be put back where they were, after the game management page tabs were rearranged.[1]
What this means for Recommended For You
RFY runs in two stages. Retrieval picks a candidate set of games for a user based on engagement, retention, and monetization signals. Ranking then orders that set and decides what to show. Crucially, only users who arrive through RFY count toward the ranking signals. Traffic from ads, friends, search, or social sharing can help you enter the candidate pool, but it does not count in ranking.[3]
The signals that matter most, in Roblox's stated priority order:[3]
- Play through rate
- First play bounce rate, split into under 60 seconds and 61 to 180 seconds
- Play days per user, split into Day 1, Day 2-7, and Day 8-28
- Playtime per user, capped at 60 minutes per user per game per day
Where this connects to the October 12 change: PTR is the top-weighted signal, and PTR is exactly the number that shifted. But Roblox is explicit that the update improves how events are recorded and does not change how RFY recommends games, and that they expect minimal impact on Home recommendations because the shift is similar across games.[1]
So the metric moved. The weighting did not.
How to read your analytics across the cutover
- Split every trend chart at October 12, 2026. Compare the 30 days before against the 30 days after. A line that jumps on the 13th is measurement, not player behavior.[1]
- Watch PTR against your benchmark, not in isolation. Roblox's FAQ says benchmarks and comparisons hold, so if your benchmark did not move the way your number did, that is a real discrepancy worth a bug report rather than a thumbnail change.[1][4]
- Check bounce rate by time segment. The Discovery docs split first play bounce into under 60 seconds and 61 to 180 seconds, so a spike in one and not the other tells you which part of your onboarding is actually losing people.[3]
- Do not touch retention numbers yet. Day 1, Day 2-7, and Day 8-28 measure behavior after discovery, and the announcement does not claim those were re-measured. If retention moved, look for another cause first.[1][3]
- File a bug report if the drop is large. The thread shows creators with PTR drops well past the small decrease Roblox described, and a published chart is a stronger report than a forum post.[1]
The table below separates the two kinds of action this change creates.
| What changed | What to do | Why |
| Impression, click, play counts from Home, Search, Charts | Re-baseline your expectations, then compare pre and post October 12 | Counts can be higher because fewer events are lost, not because more people played[1] |
| Play-through rate | Judge it against your benchmark, not the raw number | Impressions can rise faster than plays, so the ratio shifts on its own[1] |
| Historical data | Leave it alone | Roblox states data from before October 12, 2026 stays as it is[1] |
| Recommended For You ranking | No action | The update does not change how RFY recommends games[1] |
| Thumbnails and title | Only change if PTR against benchmark really fell | A raw PTR drop with a flat benchmark is the signal to investigate[1][4] |
One trap worth naming: if your game has Exclude from Recommendations enabled, Home recommendation metrics read as zero, because games excluded from recommendations cannot receive traffic from it. That is not a bug and it is not the October 12 change.[5]
What players should know
If you do not build games, none of this touches you directly. Your experience of Roblox does not change: the same games appear on Home, the same searches work, and the same charts rank the same way.
What can change is the health of the games you play. When a creator misreads a measurement artifact as a real decline, the common reaction is to panic and change a thumbnail, a title, or the first minute of the game. That is the failure mode to watch for, and it is the argument for the slow read Roblox recommends: compare periods before and after October 12 against each other instead of reacting to a single day.[1]
FAQ
Did Roblox change the algorithm on October 12, 2026?
No. Roblox changed how impressions, clicks, and plays from Home, Search, and Charts are recorded, moving them to a more reliable delivery pipeline. They state the update does not change how Recommended For You recommends games, and they expect minimal impact on Home recommendations because the shift is similar across games.[1]
Will my game rank worse because my play-through rate dropped?
Roblox says no, because the drop comes from impressions rising faster than plays and the change is similar across games, so comparisons and benchmarks hold. If your benchmark did not move the way your own PTR did, that is worth a bug report, since it contradicts the stated expectation.[1][4]
Should I re-baseline my analytics?
Yes. Roblox explicitly suggests comparing periods before October 12, 2026 against periods after it to understand the impact, rather than reading a single day or a single number.[1]
Does this change my historical data?
No. Data from before October 12, 2026 stays exactly as it was. Nothing gets rewritten.[1]
Is my PTR drop caused by worse thumbnails?
Probably not, on its own. Roblox attributes the PTR decrease to the measurement change, not to thumbnail quality. The useful test is your benchmark: if your benchmark moved the same way, the drop is measurement. If it did not, investigate further.[1][4]
Does this affect sponsored ads or search ads?
The announcement covers Home, Search, and Charts, and the acquisition dashboard lists Sponsored ads, Search ads, and Portal ads as separate sources, each with its own metrics. Roblox does not state that ad measurement was reworked here, so treat ad numbers as their own thing and compare them on their own timeline.[1][5]
Why do some creators see higher impressions and others see lower?
Roblox predicted higher impressions and play counts. Multiple creators in the announcement thread report lower counts and PTR drops larger than the small decrease described, and at least one reported a bounce rate spike. The thread had not produced a staff answer explaining the gap when this was written, so both readings are documented here rather than picking one.[1]
Sources
- Roblox Developer Forum: Improved Discovery Measurement in Creator Analytics (October 12, 2026)
- Roblox Developer Forum: Weekly Recap, October 12 to 18, 2026
- Roblox Creator Documentation: Discovery and the Recommended for You signals
- Roblox Creator Documentation: Discovery FAQ
- Roblox Creator Documentation: Acquisition analytics and attribution
- Roblox Creator Hub Changelog