Steam Next Fest has exactly one hard entry requirement that a store page can fail on its own: a publicly playable demo. So I counted. Of 286 unreleased Steam store pages captured through our analyzer between March 7 and August 18, 2026, 110 listed a demo on the store page. That is 38.5%. The other 176 would have failed the demo requirement on the day I pulled the data.
The rest of this post is the tally behind that number, split by how close each game claims to be to release, plus the trailer, screenshot, and language counts for the same 286 pages. Every figure below states its sample size and comes from raw Steam store data, not from our scoring model. The official deadline table for the October fest lives in our Steam Next Fest October 2026 dates post; this post is the population that fest draws from.
Six numbers from the sample
- Of 286 unreleased Steam store pages captured between March 7 and August 18, 2026, 110 (38.5%) listed a publicly visible demo, the one thing Steam Next Fest requires that a store page can supply on its own.
- 131 of those 286 pages (45.8%) carried no release timing more specific than “Coming soon” or “To be announced” as of August 18, 2026.
- Pre-launch pages that named an exact release day had a 77.8% demo rate (35 of 45), against 29.0% (38 of 131) for pages showing only “Coming soon” or “To be announced”.
- 27 of 286 unreleased pages (9.4%) had no trailer or video of any kind, compared with 2 of 242 released pages (0.8%) in the same capture.
- The median unreleased page carried 8 screenshots and 3 supported languages, against 10 screenshots and 10 languages for the 242 released pages in the same dataset.
- Only 70 of 286 pages (24.5%) cleared four basic pre-fest bars at once: a demo, at least one video, at least five screenshots, and a release date more specific than “Coming soon” or “To be announced”.
How the sample was built
The sample is every Steam app someone ran through steampageanalyzer.com that we still hold a stored store-page capture for. Deduplicated to one record per Steam App ID, keeping the most recent capture, that is 528 apps of type game, captured between February 10 and August 18, 2026. Splitting on Steam’s own coming_soon flag gives 286 unreleased pages (captured March 7 to August 18, 2026) and 242 released ones. The unreleased set is the population Next Fest recruits from: Valve’s Next Fest documentation states that “only unreleased games are eligible to participate” and that a title may join one fest, ever.
Every field I report comes straight out of Steam’s public store data for each app: the demos array, the count of entries in movies and screenshots, the release_date string and its coming_soon flag, the supported-language list, controller support, platforms, and price. Nothing here uses our own scoring, our issue list, or our recommendation text. Those outputs changed shape more than once in 2026, so pooling them across a six-month window would compare different rubrics against each other. Raw store fields did not change.
Four things I deliberately excluded:
- Store tags. Our stored tag counts are not comparable across the capture window (an earlier capture path double-counted, a later one dropped live tags), so any tag prevalence number here would be an artifact of when the page was captured rather than what the page contains. Dropped entirely.
- Prices on unreleased pages. Only 2 of 286 unreleased pages exposed a price at all, because Steam does not publish a price through its store API until a game is purchasable or taking pre-orders. There is no pre-launch pricing finding to report from this data.
- Discounted prices in the released set. 76 of the 234 priced pages (32.5%) were on an active discount at capture time, so the visible price was not the game’s price. Where I quote a price I use the pre-discount base price: the median released game in this sample has a base price of $13.99 (n=232).
- Anything about outcomes. This data contains no wishlists, no demo downloads, no sales, and no confirmation that any of these games registered for a fest. It describes pages, not results.
This is a self-selected sample. These are pages whose developers went looking for a store page audit, which if anything biases the sample toward teams paying attention. The real Steam-wide numbers are unlikely to be better than these.
If your page is in the unreleased group and the demo line above made you check something, run it now rather than after registration closes.
Only 38.5% of unreleased pages carry a demo
110 of 286. The comparison that surprised me is the released set: 62 of 242 released games (25.6%) still had a demo listed, which means a live demo is not some rare pre-launch artifact that vanishes at launch. Plenty of shipped games keep one up.
The demo gap matters more for Next Fest than any other page element because it is the only requirement you cannot negotiate or fix late. Valve publishes these deadlines per edition rather than as a fixed number of weeks before the opening, so the only safe source is the page for the fest you are entering. For the October 2026 edition, the demo build and store page must be submitted for review by September 14 to be live when the Press Preview starts, and all required items are due September 28. Registration itself closed on August 31.
One caveat on the count: this measures demos that are publicly visible on the store page at capture time. A demo app that exists in Steamworks but is still hidden pending review will not appear, so “no demo listed” is not the same as “no demo built”. It is, however, exactly what a shopper or a journalist sees.
The release date predicts the demo better than anything else
I bucketed all 286 unreleased pages by how precise the release date string is. This is the strongest split in the dataset, and it is not close.
| Release date shown | Pages | Share of 286 | Demo listed | Median screenshots | No video at all | Median languages |
|---|---|---|---|---|---|---|
| Exact day (e.g. “Nov 12, 2026”) | 45 | 15.7% | 78% | 8 | 0% | 6 |
| Month (e.g. “November 2026”) | 20 | 7.0% | 40% | 9 | 10% | 2.5 |
| Quarter (e.g. “Q4 2026”) | 43 | 15.0% | 33% | 9 | 5% | 3 |
| Year only (e.g. “2026”) | 47 | 16.4% | 32% | 8 | 13% | 5 |
| “Coming soon” | 82 | 28.7% | 28% | 8 | 13% | 2 |
| “To be announced” | 49 | 17.1% | 31% | 7 | 12% | 2 |
Naming a day does not cause a demo to exist. The two travel together because both are downstream of the same thing: a team that has decided when it is shipping. Every other bucket sits in a tight band between 28% and 40%, and the jump only happens at the point where a date goes on the page.
The same cohort is cleaner on every other measure I checked. Zero of the 45 exact-day pages lacked a video, against 12% to 13% for the “Coming soon” and “To be announced” groups. Their median language count is 6 against 2. Read the other direction, the finding is uncomfortable: 131 of 286 pre-launch pages (45.8%) still show no timing at all, and that group is where most of the missing demos, missing trailers, and English-only listings sit.
A vague date is not neutral. Steam’s coming soon page surfaces upcoming releases partly by date, and “To be announced” opts you out of the date-driven lists entirely while telling the fest-browsing shopper nothing about when they can play.
Where the 286 pages fall short
Five failure modes, counted on the same sample:
Screenshots are the quiet one. Steamworks requires “at least 5 screenshots of your product” per the graphical assets documentation, and 46 of 286 pages (16.1%) sit at exactly that minimum. Only 104 of 286 (36.4%) carry 10 or more. Filling the shelf is free work compared with everything else on a pre-launch to-do list, and our screenshot guide covers what the extra five should contain.
The video number is small but it is the harshest one on the list. 27 pages, 9.4% of the pre-launch sample, had no video on the store page at all. In the released set, that number is 2 out of 242.
Unreleased pages against released ones, field by field
Same capture, same fields, split on the coming_soon flag. This is a snapshot comparison of two different sets of games, not a before-and-after of the same pages, so read it as “what a finished page tends to have that a pre-launch page does not yet”.
| Store page field | Unreleased (n=286) | Released (n=242) |
|---|---|---|
| Demo listed | 38.5% | 25.6% |
| No video at all | 9.4% | 0.8% |
| Median videos | 1 | 2 |
| Median screenshots | 8 | 10 |
| Median supported languages | 3 | 10 |
| English-only language list | 35.9% (n=284) | 22.0% (n=236) |
| Full controller support | 41.9% (n=284) | 50.4% (n=236) |
| Windows only, no Mac or Linux | 75.2% | 59.1% |
| External website linked | 47.2% | 69.0% |
The language row is the widest gap in the table, and it is the one with a clean Next Fest reading. A pre-launch page listing 3 languages against a released median of 10 tells you localization is treated as a launch-window task. Next Fest, though, is the moment a game gets pushed to a worldwide audience browsing a single event page. If the store page is English-only that week, the demo download is a coin flip for a large share of the people who see it. Which languages to add first is a separate call, covered in our Steam page localization guide.
The second row worth noting is videos: median 1 before release, median 2 after. Pre-launch pages tend to have the announce trailer and nothing else, which is fine right up until the week your page is competing against a few thousand others on one grid.
What this data cannot tell you
Being explicit, since these numbers are going to get quoted:
- It does not say demos cause wishlists. For that argument and the survey evidence behind it, see do Steam demos increase sales and Chris Zukowski’s work on why demos do not hurt your visibility.
- It does not identify Next Fest registrants. Steam publishes no list of who registered before a fest opens. These are unreleased pages, which is the eligible population, not a confirmed entrant list.
- It is a snapshot per page, not a time series. Each app appears once, at its most recent capture. A page that had no demo in April may have one now.
- The sample is small and self-selected. 286 unreleased pages is enough to separate a 78% demo rate from a 29% one; it is not enough to make claims about individual genres. Genre-level demo rates in this sample sit between 36% and 50% across the seven genres with 50 or more pages, which is inside the noise for these sample sizes, so I am not publishing a genre table.
- It says nothing about demo quality. A demo that exists and a demo worth playing are different objects. Demo best practices covers the second one.
Check your own page against these numbers
Six checks, in the order the data says they matter, with the benchmark from this sample next to each:
- Is a demo publicly visible on the page today? 61.5% of pre-launch pages fail this. It is also the only Next Fest requirement with a hard build deadline attached, and for the October 2026 edition that deadline is September 14.
- Does your page name a release day, month, or quarter? 45.8% of pre-launch pages show nothing but “Coming soon” or “To be announced”. A date puts you in the cohort where 78% have their demo ready.
- Is there at least one video? 9.4% of pre-launch pages have none. Released pages run a median of 2.
- Are you above 10 screenshots? Only 36.4% of pre-launch pages are. 16.1% sit at Steam’s 5-screenshot minimum.
- Is the language list still English-only? 35.9% of pre-launch pages are, against 22.0% of released ones.
- Would your page clear all four of demo, video, five screenshots, and a real date? 24.5% of the sample does. That is the bar this study measures, and it is a low one.
Our Next Fest prep tool walks the same checks against your App ID with the deadlines attached, and the free page analysis reports the raw counts above for your page so you can see which side of each number you are on.
Cite this study: “Steam Next Fest Data: Only 38% of Pre-Launch Pages Have a Demo”, Steam Page Analyzer, September 8, 2026. Sample of 286 unreleased Steam store pages captured March 7 to August 18, 2026. Canonical URL: https://steampageanalyzer.com/blog/steam-next-fest-demo-data