What Makes Poki Web Games Keep Players Playing?

Poki web games keep players through visible progression, not fun alone: players stay when they know a new world, hat, or level is one more loop away. At Gamescom São Paulo 2026, Poki's team told one Brazilian indie developer that players average about 10 minutes per game, the core loop should last around 3 minutes, and the audience is mostly children aged 9 to 12. His own games averaged just 2 minutes 20 seconds. This article breaks down what that gap means and how to close it.

What happened at Gamescom São Paulo 2026?

A Brazilian indie developer went to Gamescom São Paulo 2026 after a devlog about making one game per week went viral on his channel, and the trip produced the most useful data of his game development journey so far. The video that started it proposed a simple experiment: ship a game every week, measure engagement through Poki, a web games platform that lets developers test titles with real players, and keep iterating until one game earns money.

That video reached roughly 160,000 views at the time of recording, and the Poki team noticed it. They contacted the developer, offered to cover his travel, and met him at the event in São Paulo, where they gave a talk about building games that engage on the web. The four-day event itself he found exhaustible in two: he visited every booth in that window and spent the remaining days in talks and meetings.

The trip overlapped with a deadline. He was mid-way through the Levelship IO game jam, also referred to in the community as the Levels IO jam, whose single rule was to build a game using artificial intelligence. He worked on it between booth visits, then stayed up until 4 a.m. on the last day to submit. The game did not win, and the main criticism was that it was not fun, a recurring weakness he traces to his strength: he has programmed since 2017, but the craft of making a game feel good is still one he is learning.

Who plays web games on Poki, and for how long?

The audience for Poki web games is mostly children between 9 and 12 years old, according to platform data shared at the event. That single fact reframes everything else. Games aimed at adults, or at gamers with years of Steam experience, carry complexity that this age group simply bounces off. The developer realized most of his own prototypes were built for the wrong player.

The numbers he took away form a compact design brief:

Players are predominantly children aged 9 to 12, so bright, colorful art and instant readability matter more than depth.
Fun must arrive within the first 10 seconds of gameplay; a menu screen is a leak, so send the player straight into the game.
An average player session lasts about 10 minutes on the platform.
The core gameplay loop should last roughly 3 minutes, with something new unlocked or revealed at each cycle.
His own games had never exceeded an average engagement time of 2 minutes 20 seconds, measured with the platform's playtesting tools.

For a developer who enjoys more complex games, designing for a nine-year-old is its own discipline. He describes it as mentally difficult: what delights a child, like unlocking a simple blue hat, does not delight him, and bridging that gap is a real skill rather than a matter of lowering standards.

How does Poki playtesting measure engagement?

Poki's playtesting tools give approved developers a measured average engagement time built from real play sessions. The devlog experiment was designed around exactly this: upload a game, let real players try it, and read the numbers instead of guessing. In the original video, the developer explained that the platform lets him test a game with 500 real players and compute an engagement average from those sessions, and he planned to repeat that cycle weekly until one game paid his bills. When Romy Halfweeg later recorded her closing message, she addressed it to what he estimated at 1,000 developers watching the channel, telling them to submit through the developer form to get access to the same playtesting tools.

Why does progression beat fun for retention?

Progression, not fun, was the diagnosis a Poki business development manager gave when asked why his games lost players after a minute or two. In a roughly 15-minute interview at the event, Romy Halfweeg, who works in business development at Poki and traveled from the Netherlands for the talk, told him she had played his games and found them already fun. The problem was that nothing told the player why to stay.

Her reasoning, in his paraphrase and hers: a player arrives at a web games platform wanting something quick and entertaining. They play one loop of a minute or two, enjoy it, and then ask themselves why they would not switch to another game. If there is no visible reward, no unlock, no next world, no teased content, there is no reason to answer that question in favor of staying.

The practical advice was to design the reason to continue before anything else, in this order:

Decide what reward the player earns: a new world, a new hat, or a new level.
Place each unlock so the player reaches one roughly every 3-minute loop.
Tease the next reward early, so the player knows what one more minute of play will earn.
Verify with playtesting data that average engagement passes the 3-minute mark, which he described as a gate for the platform.

Reward visibility early beats reward generosity late.

This maps directly onto his measured failure. Games with no progression system capped out at 2 minutes 20 seconds of average play, far below both the 3-minute loop target and the platform's 10-minute session average. The engagement gap was a motivation gap, and motivation is a design feature, not a byproduct of good code.

Is the web a good place for new game developers?

The web is a fast way to build real game development experience, and that was the first advice Halfweeg gave to beginners. Before joining Poki she made her own games, and she chose web games deliberately: short development cycles, small scopes that match player expectations, and a full release cycle in a couple of weeks instead of the two years a Steam-scale project can demand, especially as a side project next to a day job.

Her second recommendation was to study the market before building. She named the major web game platforms worth browsing to understand what popular games and audiences look like:

Poki, the platform she works with, focused on entertainment-first web games.
CrazyGames, one of the largest independent web game portals.
YouTube Playables, Google's playable experiences inside YouTube.
Y8 Games, a long-running web games portal.

She also argued the future of the format is expanding. Web games require no download and run instantly, which makes them attractive to platforms that are not game platforms at all. She cited the Chinese ride-hailing app Didi, which shows players an HTML 5 game while they wait for a taxi, and noted that more ecosystems are adding HTML 5 games to their non-gaming products while dedicated platforms like Poki also report growth. Both trends point toward more demand for web games in general, not less.

Developers interested in the platform can apply through Poki's developer page, which includes a submission form; approved developers get access to the same playtesting tools used in the original viral experiment.

Web or Steam: which should an indie developer choose?

The middle path won: validate on the web, then bring the standout to Steam. The developer is explicit that there is no universally right answer, because it depends on the goals and profile of whoever is building. But for his situation, the economics lined up cleanly, and the two destinations reward nearly opposite behavior:

	Web (Poki and peers)	Steam
Development cycle	A couple of weeks per game	Often two years, especially part-time
Scope	Small, matching player expectations	Larger, deeper experiences
Best use	Fast validation with real players	Destination for proven concepts
Risk	Low per experiment, measured in data	High, one long bet

The web rewards speed. Short cycles reveal what works and what does not, without betting years on an unproven idea. Steam rewards the opposite: depth, longer development, and a finished experience players pay for. Using the web as a validation layer and Steam as a destination means the expensive, multi-year project only happens for concepts already proven with real players.

That is also how his channel's experiment changed shape rather than ending. The new question he poses for upcoming videos: can an indie developer who never found a hit on the web find one on Steam? The games that perform above the 3-minute engagement bar are the candidates. The rest become tuition.

This journey sits inside a wider Brazilian creator community doing the same kind of public experimentation, alongside channels like Gustavo Dev Doido and educational projects such as Crazystack Typescript and the Bootcamp do Dev Doido, which cover the programming side of shipping web projects like those documented at crazystack.com.br.

Is Gamescom São Paulo worth it for indie developers?

For the networking alone, yes; for the length of ticket he bought, no. The developer's verdict after the trip is split. The conversations, the Poki interview, and the chance to play Brazilian indie games without long lines were clearly worth the flight. What he would change is the ticket itself.

He exhausted the exhibition floor in two of the four days he had paid for, and found the remaining talks less valuable than expected. The big stands, anchored by titles like League of Legends and AAA releases, drew crowds that the indie booths did not, but that same imbalance left indie games easy to reach and test. His advice in effect: go, meet people, play the indie section thoroughly, and buy fewer days than enthusiasm suggests.

He also worked through the event, which is a fair warning for anyone attending with an active deadline. Between booths he found quiet spots to keep building his game jam entry, and skipped the final day entirely to submit by 4 a.m. An event trip and a jam deadline can coexist, but only with the expectation that neither gets full attention.

FAQ

Who is the audience for web games on Poki?

According to platform data shared at Gamescom São Paulo 2026, the player base is mostly children aged 9 to 12. Games aimed at this group need instant readability, bright visuals, and gameplay that delivers fun within the first 10 seconds, with no menu between the player and the game.

How long should a web game's gameplay loop be?

The guidance from Poki's team was roughly 3 minutes per loop, with something new unlocked or revealed at each cycle. Average sessions on the platform run about 10 minutes, so a well-structured game can hold a player across several loops if each one advances progression.

Why did the developer's games average only 2 minutes 20 seconds?

The games were fun but offered no visible progression, so players had no reason to stay after one loop. The recommended fix was to define rewards and unlocks early and tease them from the start, so the player knows what one more minute of play will earn.

Is the web better than Steam for a first game?

For a new developer, the web offers a much shorter cycle: a small game in a couple of weeks instead of a two-year Steam-scale project. The strategy proposed in the video is hybrid: use the web to validate ideas quickly, then develop the concepts that stand out into Steam releases.

What was the Levelship IO game jam?

It was a game jam whose only rule was to build a game using artificial intelligence. The developer built his entry during Gamescom São Paulo 2026, submitted it at 4 a.m. after skipping the final event day, and lost; the main criticism was that the game was not fun. He plans a dedicated video on the results and the debate they caused.

How can developers apply to publish on Poki?

Developers submit through Poki's developer page at developers.poki.com, which hosts a submission form describing the developer, the game, and the plan for it. Approved developers receive access to the platform's playtesting tools, which measure engagement with real players.

Turn Your Own Devlog Into a Written Deep Dive

The lesson from this trip is that a platform's own data can overturn a developer's assumptions, and the same is true of your own catalog of videos. If you have interviews, postmortems, or lessons sitting inside YouTube videos, they are worth more as searchable, quotable articles than as 12-minute timelines a viewer has to scrub through.

That is exactly the flow Skala Blog was built for: paste a YouTube URL at Skala Blog, get a transcription, and generate a structured article from it, the way this piece was built from one devlog.

Source video