Why do 99% of Roblox games fail?

Is it really 99%? What Roblox publishes

No Roblox document gives a failure rate. These are the figures it does publish, read on 22 October 2026. The RDC 2026 rows are Roblox’s conference figures as reported by Dexerto; Roblox’s own release of them was not found.

Millions of games against tens of thousands of creators in DevEx, with a median DevEx payout around $1,500 a year, is the measured shape behind the saying. The DevEx figures are measured in Robux spent, per the report’s footnote.

What does Roblox’s Home feed rank?

Home’s Recommended for You sort works in two stages. Retrieval picks a set of games each user might enjoy from signals such as engagement, retention and monetisation. Ranking then orders them, and Roblox states that how far a game goes “depends entirely on the engagement and retention of users who come to your game through Recommended for You”. The signals, in Roblox’s priority order:

The signals are averages per user, so a small game is not ranked down for being small. Since April 2026 the feed looks at a 28-day window instead of 7 days, and Roblox’s Q2 letter says it “intentionally provides more impressions for highly retentive games”. Sources: Roblox discovery docs and the Q2 2026 shareholder letter.

Why do most Roblox games stall?

1. No first players, so no signal. Ranking only counts players who arrived from Recommended for You. Roblox says ads, search, friends and social sharing “can accelerate your consideration”, and that even a small number of players can signal a game is worth showing to more. A game nobody plays never enters that loop.
2. Players leave in the first minute. First play bounce rate under 60 seconds and at 61 to 180 seconds is one of the four most important signals, and it counts against the game.
3. No reason to come back. Play days are measured on day 1, days 2 to 7 and days 8 to 28. A game that is finished in one session scores on the first and nothing after.
4. The listing disqualifies it. A game with only the default thumbnail is not eligible for Home recommendations. Titles that lead with giveaways, metadata that does not match the gameplay, and copies of existing games are not prioritised.
5. Something else holds players better. Roblox lists it as one of the four factors behind Home impressions: a game’s distribution can fall when other games engage players more, even if its own signals hold.

Is it the game or the distribution?

Both, in that order of discovery. Distribution decides whether a game is measured at all; retention decides how far it goes once it is. A good game with no first players is never tested; a game that gets players and loses them in a minute is tested and stopped. Roblox’s own example: a game with 10,000 daily players from Home adds 5,000 from ads and keeps its Home traffic as long as the Home players behave the same. Ads add players; they do not buy ranking.

Where do the first players come from?

• Your own clips and community: short gameplay clips on TikTok and Shorts, a Discord server and a Roblox group. Free, slow. The Discord guide covers the server side.
• Sponsored ads: Roblox’s Ads Manager “Plays” goal is built to send initial traffic that helps “qualify your game for Recommended For You”. It is an auction with no published price; the cost guide has what Roblox does publish.
• Creators: a YouTuber or TikToker playing the game for a flat fee. The YouTuber guide covers finding and paying them.
• Clipping campaigns: many creators post short clips of the game and you pay per 1,000 verified views. On MetaClip campaigns start at $150; the campaign guide walks through one.
• Curation: Roblox’s Standout Games sort features novel games picked by its curation team; nominations are open.

The players guide turns this into a checklist, and the channel guide compares all seven channels on cost.