How to Run a Steam Playtest Before Launch: Choosing Between Playtest, Demo, and Beta, Recruiting Testers, What to Measure, and a Four-Week Pilot Plan

Steam Playtest is close to the only official tool that lets you put an unreleased game in front of real Steam users for free while routing that interest into wishlists and follows on your main store page. Valve ran the feature as a beta with selected games starting in November 2020, during which more than one million players took part across 161 games, and opened it to all developers in April 2021. Valve's Steamworks documentation now describes Steam Playtest as "the preferred method" of enabling closed betas on Steam.

Yet many teams treat a playtest as a QA milestone for the dev team rather than designing it as a stage in the marketing funnel. Consider what the feature actually does: the signup button sits on your main store page, signups and participation are tracked separately from wishlists and reviews, and you decide when it opens and closes. That makes a playtest the cheapest place to build and validate pre-launch interest. Based on Valve's official documentation as of October 2026, this post covers when to use Playtest versus a demo, a beta branch, or Next Fest, how to set it up, how many testers you need and where to find them, what to measure, and how to structure a four-week pilot.

We covered how wishlists shape launch performance in Steam Wishlists Decide Your Launch. Here Is How to Build Them Before Day One, and demos and pre-event preparation in Steam Next Fest Marketing. This post fills the stage between them: the Steam Playtest you run before, or alongside, a demo. The October 2026 Steam Next Fest runs October 19-26, and the next editions open on February 22 and June 14, 2027, so if you are aiming for a first-half 2027 launch, now is when to put the playtest on the calendar.

How Is Steam Playtest Different From a Demo, a Beta Branch, or Next Fest?

Steam Playtest is a free test app that runs on a separate "child" App ID tied to your main game, with its signup point on the main store page. A demo is a polished sample meant to help players decide to buy, a beta branch opens a specific build to people who already hold a license for the game, and Next Fest is an event open only to unreleased games with a demo. They serve different purposes, so the right move is usually to place them at different stages rather than pick one.

These are the core Playtest characteristics Valve states. The playtest has no separate store page and appears as a section on the main game's page, designed, in Valve's words, "so that customers can sign up and access the playtest but still Wishlist or Follow the main game." Joining or leaving the playtest does not affect a customer's wishlist for your game, a customer who has only played the playtest cannot review your actual game, and playtime, wishlists, reviews, and refunds are tracked separately from the main game. It is free for both customers and developers, and selling access or in-game transactions are not allowed. On the other hand, signups carry no expectation of secrecy, so it is not a tool for confidential testing.

Turned into a decision, the table says this. If the build is still rough and the goal is feedback and server validation, use Playtest. If the first ten minutes are finished and can carry a purchase decision, use a demo. If you need to show a game you have not announced yet to a small private group, use Release Override keys and a beta branch. Next Fest requires a demo, so a playtest cannot stand in for it. On demos, Valve says "we generally recommend waiting until launch," while adding that a pre-release demo helps if players need to touch the game to understand it. In practice, the natural order is to refine the build through playtests and build your Next Fest demo from what you learn.

How Do You Set Up a Steam Playtest?

Valve's documentation lays out seven steps, and the playtest app goes through its own simplified review, separate from the main store page review. If you do not build review time into the schedule, you may miss the opening date you announced.

The sequence is as follows. First, create the Playtest app from the Associated Packages & DLC page in Steamworks. Second, provide the library capsule and community assets; the playtest name can be changed until release and can be localized. Third, complete the release process for the playtest, which includes a simplified store page and build review; the store review checklist consists only of capsule images and icons. Fourth, enable signups from the Special Settings tab of the main game's store page editing tools. Once you publish the change, customers see the signup option, and you can show or hide signups at any time. Fifth, upload depots and builds as you would for any game or demo. Sixth, once the build is live, add players from the signup pool. Seventh, deactivate when you are done.

There are two signup types. Limited, the default, collects requests and lets you grant access in batches as you have capacity. Open automatically adds players who request access from the main store page. Under Limited, players are selected randomly from the signup pool, and you can pick specific countries, mainly to limit latency when server capacity is tight. Players receive an email from Steam when they are added and can see their status on the main store page. Switching from Limited to Open starts a process that accepts pending playtesters, which Valve says will generally start within a few minutes. If you turn on the experimental Friend Invites feature, participants can invite one to three friends each, and invitees must have been friends with the inviter for at least 30 days.

Secondary sources disagree on keys, so here is what Steamworks says. You can request playtest keys even before your store page is live, only Standard Release keys can be requested, and Release State Override keys cannot be used for a playtest. While the main game typically has a limit of 2,500 Release State Override keys, playtest keys can generally be requested in "much larger quantities." However, Valve treats a request for more than 50,000 keys as a sign that your playtest is effectively an open beta and suggests opening signups on the store page instead, and it reserves the right to reject key requests. Keys must never be sold or included in paid bundles or subscriptions, and keys already granted are not revoked when the playtest is set to Not Playable.

To end a playtest, set it to "Not Playable" in the Steam Playtest partner settings. That option is hidden while signups are open, so turn off store signups first. Once deactivated, Steam will not launch the playtest app, and if you set it to Visible again, players who already had access keep it without signing up again. One thing to note: Steam does not email players when a playtest becomes playable or not playable. Valve warns that you might frustrate or disappoint your community if you take the experience away without communicating it beforehand, and recommends using Steam Events and Announcements. There is also a reset function that removes all participants, but it cannot be undone and disables every playtest key including unactivated ones, so use it carefully and only when moving to a new phase.

When Should You Open a Playtest? A Pre-Launch Timeline

Valve considers the best, and probably most common, time to use Playtest to be before release, after your Coming Soon page is up. Because the signup button lives on the main store page, a playtest does not work as a funnel without one. Valve also says that keeping a playtest up after release probably only makes sense if your game is already free-to-play.

Below is a suggested layout starting 12 months before launch. It is not a Valve schedule but an example combining the documentation above with public cases, so adjust it for your game's scope and genre.

The core idea is to use the playtest as a lab before the demo. With the next Next Fest on February 22-March 1, 2027 and the one after on June 14-21, a team aiming at February would realistically run its first playtest in November-December 2026 and release its demo by early January. The principle of deciding in advance what to validate before a full launch, as with a mobile soft launch, is covered in Soft Launch Strategy: What to Validate, and When to Go Global. Playtests benefit from the same discipline: write down a pass criterion for each stage.

How Many Testers Do You Need, and Where Do You Find Them?

The number you need changes by orders of magnitude depending on what you are validating. Qualitative testing to find usability problems can start at around five people per group, metrics you judge by the numbers, such as playtime or where players drop off, need dozens or more, and server load testing is sized to your target concurrent player count.

UX research firm Nielsen Norman Group says it is good practice to start qualitative tests with five users and to test distinctly different user groups separately. For quantitative studies, it says these usually need more than 30 participants, and that confidence intervals are narrow only with 40 or more. Applied to a playtest, that means finding the friction points in the tutorial and first session through small observed sessions with about five players per genre-experience group, and judging median playtime or drop-off at a specific point on data from at least 40 active testers. Not everyone you grant access to will play, so grant more people than your target number of active testers. Valve does not publish a benchmark for how many signups actually play, so the accurate approach is to measure it in your first batch and size the next batch from that.

The paid ads figure comes from the life sim This Grand Life 2, reported by WN Hub in May 2024, where a mostly Reddit-based campaign produced $0.80 to $2.50 per wishlist depending on the ad group. It is a cost per wishlist rather than per playtest signup, and it is one game's result, so treat it only as a starting point for measuring your own cost per signup and per wishlist with a small spend over the first one to two weeks. How to use a Discord server as the hub of playtest operations is covered in Discord Marketing for Game Studios: What a Server Does From Pre-Launch Hype to LiveOps Retention.

How Do You Collect Feedback and What Should You Measure? A Four-Week Pilot

Valve does not provide feedback tools, so you have to design collection yourself, and the yardstick is not signup volume but conversion: from signup to actual play, and from play to wishlists and follows. Valve suggests a feedback button in the main menu, a forum, Discord, a bug tracker, or a website, and recommends using announcements to point players to them. There is no built-in survey feature.

One more thing to know: as of October 2026, the Playtest documentation does not describe a report that links individual participants to main-game wishlists. Because participation does not affect wishlists and is tracked separately, wishlist conversion has to be read through indirect metrics. If you send store page traffic through UTM links per external channel, Steam UTM Analytics breaks out wishlists and purchases made within 72 hours of the click by channel, though only visits from users logged into Steam in that browser are counted and conversions are finalized four days after the visit.

Packaged as a four-week pilot, it looks like this. In week 0, write down pass criteria (for example, targets for activation rate, median playtime, and lift in average daily wishlists), prepare per-channel UTM links, a survey, a Discord feedback channel, and a crash reporter, and lock the average daily wishlists of the previous two to four weeks as the baseline. In week 1, open Limited signups and grant a small batch that includes core Discord fans, running observed sessions with around five players per genre-experience group. In week 2, confirm crashes are under control, then expand batches by country, and for multiplayer games, open a timed weekend session announced in advance through Steam Events. In week 3, run the survey, ship a fixed build, and ask for wishlists and follows of the main game on the end-of-session screen and in announcements. In week 4, announce the end in advance, set the playtest to Not Playable, compare wishlist lift, activation rate, playtime, and survey results against the week 0 criteria, and decide the next step: a second playtest, building the demo, or adjusting the schedule.

Ask for wishlists, but it is safest not to tie the ask to rewards. Valve prohibits soliciting reviews in exchange for games, DLC, money, or other rewards, and prohibits asking customers to review your product from within your application. Valve's review and wishlist documentation does not explicitly address rewards for wishlists, so if you need a ruling, ask Valve directly. Start with the asks you can make without any reward, such as end-of-session copy and the store page itself. Store page conversion design is covered separately in how to turn Steam store page visits into wishlists.

What Do Public Cases and Common Mistakes Teach?

Public cases show both that collecting signups first and opening the doors in stages is an effective way to build interest, and that opening at scale before you are ready exposes server and first-impression problems in full view.

The most specific numbers come from publisher tinyBuild. In a company blog post on June 19, 2024, CEO Alex Nichiporchik said the portfolio had generated over three million wishlists since the start of 2024, more than one million of them in the past month alone. He called Steam Playtest "the single most powerful tool to get early hype for your game, and to actually get people to play." The extraction shooter Level Zero: Extraction got close to 100,000 signups before its March playtest (March 15-19), and Duckside, revealed as an April Fool's announcement, drew over 70,000 signups. Duckside's playtest surfaced server problems, and the team says stability was improved afterward. For Kingmakers, signups opened right after the announcement, the test stayed closed to the wider audience for a few weeks, and the doors opened once critical mass had built. Nichiporchik's advice: "Do not let everyone in straight away," and "a few weeks is usually enough." The post does not disclose per-game wishlist counts or how many playtest signups went on to wishlist, so the causal link between signup volume and wishlist results cannot be confirmed in numbers.

For a timed stress test right before launch, Embark Studios' ARC Raiders is an example. During its "Server Slam" on October 17-19, 2025, which was open to everyone without registration or access codes, it reached 189,668 concurrent players on Steam alone (per SteamDB, as reported by MP1st), and the game launched on October 30. Reports do not confirm whether the test used the Steam Playtest feature, so it is more accurate to treat it as an example of an operating pattern, a weekend-only opening about two weeks before launch, than as a Playtest case. It is also a multiplatform game from a large studio, so rather than using it as a benchmark for indie scale, take from it that a pre-launch test can aim at server validation and pre-launch buzz at the same time.

Four mistakes come up repeatedly. The first is opening too early. Open signups on a build that still crashes or on servers that cannot hold set the first impression during the playtest, and since playtests carry no expectation of secrecy, that impression gets shared as is. The second is running indefinitely. Valve limits the case for keeping a playtest after launch to free-to-play games, and a playtest left open forever gets consumed like a free demo and erodes anticipation for launch. Set an end date before you start and announce it. The third is a test with no wishlist ask. Playtest participation is not reflected in wishlists automatically, so ask clearly for wishlists and follows of the main game via the wishlist button next to the signup, the end-of-session screen, and the closing announcement. The fourth is planning to use a playtest in place of Next Fest. Only unreleased games with a playable demo can take part, so connect playtest results to your demo production schedule. The first image players see on your store page also needs to convey the genre accurately so that signups and wishlists come from the right people, which we covered in detail in What Steam Capsule Art Has to Do: 2026 Sizes and Valve's Rules, Choosing and Briefing an Artist, Testing Without an A/B Tool, and Wishlist Conversion Data.

How Should the Funnel After the Playtest Be Designed? The Playio Perspective

If the playtest is the stage where you validate interest inside Steam, the next job is bringing gamers with matching tastes from outside Steam to your main store page and checking whether that interest turns into purchases and play after launch. That is the stretch Playio's Steam product is built for.

Playio is a gaming community of five million gamers, where AI analyzes genre preferences, play history, and in-game behavior to connect games with players whose tastes match. For PC Steam games, the Steam product can set up each stage of the funnel as its own action: wishlist and follow, add to cart, purchase, playtime milestones, and achievement completion. The playtest app is a separate App ID and is not itself a target of these actions, but gamers who arrive at the main store page during a playtest announcement can sign up and wishlist or follow on the same screen. After launch, the same campaign structure carries through to purchase, playtime, and achievements, which is the difference from campaigns that stop at the wishlist count. Reviews are never designed as a campaign action, because Valve prohibits soliciting reviews in exchange for rewards.

If you run the two together, keep the order straight. Lock the wishlist baseline in week 0 of the pilot and route the external campaign only through dedicated UTM links, read separately in Steam UTM Analytics, so the effect of the playtest itself and the effect of the external campaign do not blend.

As a concrete next step, send your game's genre, Steam store page link, current wishlist count, planned playtest dates and target tester numbers, and your launch date or the Next Fest you plan to join to [email protected], and we will send you a playtest-to-launch funnel pilot plan first: per-channel UTM structure, target actions and billing stages for each step, and a decision rule based on purchases and playtime rather than wishlist count.

You can find more details here. (https://playioadsen.oopy.io/bizdeck)

Key Takeaways

As of October 2026, Steam Playtest is a free testing tool that runs on a separate child App ID tied to your main game. Its signup sits on the main store page, so you can collect wishlists and follows on the same screen, and participation is tracked separately from main-game wishlists and reviews. Valve describes it as the preferred method of enabling closed betas on Steam. As a baseline, use Playtest for feedback on rough builds and server validation, a demo for purchase decisions and Next Fest, and Release Override keys with a beta branch for private tests of unannounced games. Grant access in batches under the default Limited signup type; keys are available in large quantities, but requests above 50,000 are treated as an effective open beta; and end the test by announcing it and setting it to Not Playable. Plan around five testers per group for qualitative work and at least 40 active testers for quantitative judgments, and run a four-week pilot that compares activation rate, playtime, and average daily wishlist lift against a baseline rather than counting signups. Collecting signups first and opening in stages works, as tinyBuild's case shows, but avoid opening too early, running indefinitely, forgetting to ask for wishlists, and planning to use a playtest in place of Next Fest.

For inquiries about Playio's advertising solutions, reach out at: [[email protected]](mailto:[email protected]?subject=Playio%20inquiry%20from%20blog%20-%20Steam%20Playtest%20guide)

Sources

• Steamworks Documentation, Steam Playtest (separate child App ID; signup on the main page alongside Wishlist and Follow; seven setup steps; Limited and Open signup; random and country-specific grants; email notification; key policy: Standard Release keys only, large quantities, 50,000+ treated as an effective open beta, no selling; Release State Override keys typically 2,500; Not Playable deactivation and advance communication; wishlists, reviews, refunds, and playtime tracked separately; free, no monetization; Friend Invites 1-3 per player and 30-day requirement; no expectation of secrecy; reset; feedback channel examples; best used before release, after release only for free-to-play): https://partner.steamgames.com/doc/features/playtest

• Steamworks Documentation, Testing On Steam (Steam Playtest is the preferred method of enabling closed betas; Release Override key activators own the game forever by default; key banning tool): https://partner.steamgames.com/doc/store/testing

• Steamworks Documentation, Demos (separate App ID; can appear anywhere a free game can; demo user reviews; one wishlist notification within 14 days of first release; generally recommend waiting until launch): https://partner.steamgames.com/doc/store/application/demos

• Steamworks Documentation, Branches (definition of beta branches, password protection, testing use): https://partner.steamgames.com/doc/store/application/branches

• Steamworks Documentation, Upcoming Events (Next Fest October 19-26, 2026, February 22-March 1, 2027, June 14-21, 2027; only unreleased games with a playable demo are eligible): https://partner.steamgames.com/doc/marketing/upcoming_events

• Steamworks Documentation, User Reviews (no soliciting reviews in exchange for rewards; no in-app review requests; only purchases count toward the review score): https://partner.steamgames.com/doc/store/reviews

• Steamworks Documentation, Wishlists (no explicit rules on wishlist rewards or incentives; demo release notification): https://partner.steamgames.com/doc/marketing/wishlist

• Steamworks Documentation, UTM Analytics (wishlist, purchase, and activation conversions; 72-hour attribution; logged-in visits only; finalized four days after the visit): https://partner.steamgames.com/doc/marketing/utm_analytics

• Game World Observer, Steam releases Playtest feature (April 22, 2021; beta from November 2020; over one million players across 161 games; opened to all developers; country and region limits): https://gameworldobserver.com/2021/04/22/steam-releases-playtest-feature-allowing-developers-to-invite-players-to-test-games

• tinyBuild, 3 million wishlists (June 19, 2024; Alex Nichiporchik; three million wishlists in 2024 and one million in the past month; Level Zero: Extraction close to 100,000 signups, March 15-19; Duckside over 70,000 signups; staged opening for Kingmakers; "Do not let everyone in straight away"): https://www.tinybuild.com/single-post/3-million-wishlists

• Game World Observer, How tinyBuild games generated over 3 million wishlists in 2024 (June 20, 2024; coverage of the post above; Duckside server problems and later improvements): https://gameworldobserver.com/?p=25081

• MP1st, ARC Raiders Server Slam Draws Nearly 190K Players on Steam Alone (October 18, 2025; October 17-19 open without registration; 189,668 concurrent players on Steam per SteamDB; launch October 30): https://mp1st.com/news/arc-raiders-beta-draws-190k-players-steam

• WN Hub, How solo dev spent $4,365 on paid ads to reach Popular Upcoming on Steam (May 31, 2024; This Grand Life 2; $4,365 in ad spend, 4,000 wishlists, $1.10 average, $0.80-2.50 range; Reddit CTR): https://wnhub.io/news/investment/item-43597

• Nielsen Norman Group, Why You Only Need to Test with 5 Users (March 18, 2000; Jakob Nielsen; test distinct user groups separately): https://www.nngroup.com/articles/why-you-only-need-to-test-with-5-users/

• Nielsen Norman Group, 5 Test Users: Qual vs. Quant (July 11, 2021; Raluca Budiu; start qualitative tests with 5; quantitative usually more than 30, narrow confidence intervals with 40 or more): https://www.nngroup.com/articles/5-test-users-qual-quant/