Google Put Your Android Catalogue on Windows by Default. “Untested” Means Nobody Will Find It.
Google Play Games on PC lists Android games on Windows by default, and a licensed catalogue lands there badged “untested” — hidden from browse, and unbudgeted. The default flipped in March 2025. Most licensees have never opened the app to see what their titles look like on a desktop, and the licence they signed almost certainly does not use the word Windows anywhere.
This is not a hypothetical channel you might one day pursue. If your APK titles are live on Google Play, they are already enrolled unless somebody in your Play Console went and turned it off. The question is not whether to enter the channel. It is what your catalogue currently looks like inside it, and whether the answer is embarrassing.
🪟 The Default Flipped, and the Licensee Was Not in the Room
At GDC in March 2025, Google wrote the change plainly on the Android Developers Blog: "To help maximize your discoverability, we're making all mobile games available on PC by default with the option to opt out anytime."
Google's Play Console programme page for Google Play Games on PC puts the reach at 140+ markets. The app has run on Windows since 2022 and, as of August 2026, still carries a beta label — which tells you something about how much weight to put on the channel, but not that you can ignore it.
The mechanics matter for anyone who licenses rather than builds. The Play Console account holding the listings is the one with the opt-out switch. On a licensing deal, that account belongs to whoever publishes — which may be you, may be the licensor, and on a reskin arrangement may be a third party neither of you thinks about weekly. Nobody negotiated that switch, because at the time the deal was drafted it did not exist.
🏷️ Three Badges, and Only Two of Them Are Findable
Google's playability review process documentation describes badges that range "from untested (not reviewed yet, but may work) to playable (works well) to optimized (designed for the best PC experience)."
The discovery consequence is the whole story, and it is one sentence in the docs: an untested game "will only appear if a user specifically searches for it in the Google Play Games on PC search menu." Playable and optimized titles turn up in browse, in text search, and on game details pages. Untested titles turn up only if someone already knows the name and types it.
For a catalogue business, that is a distinction without a difference. Nobody types the name of a casual title they have never heard of. A 1,000-title catalogue at the untested tier has an addressable audience on PC of approximately zero, while still occupying a store listing that a player can land on and bounce from.
Worth noting for accuracy: Google's player-facing help page on Windows compatibility for games describes five states, not three — it adds unsupported (known technical issues that "may prevent this game from running correctly on PC," though users can launch anyway) and not available. The same page carries a caveat that badges reflect one-time testing rather than continuous evaluation, so a badge earned on one build does not follow your updates forever. Developer docs and player docs disagree on the taxonomy. Plan against the player-facing one, because that is what your audience sees.
📋 What "Playable" Actually Costs, Per Title
The playable tier is not a high bar in isolation. Per Google's review documentation, a game needs to be stable, avoid significant ANRs, hold a stable frame rate at 30fps or better, and support mouse and/or keyboard. On top of that it must disable unsupported Android features and permissions, disable unsupported Google APIs, and use scoped storage for file access.
One detail helps licensees more than any other: ARM-only builds are accepted at the playable tier. x86-64 is strongly encouraged, not required. If you were assuming a full recompile per title, you were assuming a bill that Google does not currently charge you.
The optimized tier is a different animal. It adds the Input SDK, high-resolution assets, PC aspect ratios, UI scaled for large screens, 60Hz rendering, and a cloud save implementation that works across PC, Android and iOS with account binding to Play Games Services v2. That is a per-title engineering project inside source you did not write. For a licensed catalogue, optimized is a decision you make about five titles, not a thousand.
The manifest lines that quietly filter you out
Google's PC compatibility guidance lists hardware features that must be marked android:required="false" for a game to work in the environment — among them android.hardware.touchscreen, android.hardware.telephony, the whole android.hardware.sensor.* family, camera, NFC, WiFi and Bluetooth. A stack of permissions has to go too, including location, camera, external storage read/write, FOREGROUND_SERVICE and SYSTEM_ALERT_WINDOW.
A casual game that declares a touchscreen requirement is not badly built. It is honestly built, for a phone. It just fails on a desktop for a reason that never appears in your QA notes, because you tested on phones.
Every touch idiom becomes a defect
The same documentation asks for pinch-to-zoom and multi-touch gestures to be replaced with keyboard and mouse equivalents, for on-screen text-entry keyboards to be removed, for scrollbars on scrollable lists, for clickable elements to accept clicks anywhere in their visible bounds, for dialogs to have visible close buttons, and for the word "tap" to become "click" in user-visible strings.
Read that list again as a licensee. Every item is a change inside a compiled build, in a codebase governed by an agreement that says what you may and may not modify. If your Android game licence grants distribution but not modification rights, the entire playable checklist is work you cannot legally do yourself.
💸 The Part That Decides Whether Any of This Is Worth Doing
Most licensed Android catalogues run on ads. So the channel lives or dies on one question: do ad requests from Google Play Games on PC fill, and at what rate?
Google's ad guidance for Google Play Games on PC answers the technical half cleanly. Existing SDKs "should continue to function in Google Play Games on PC without changes." The requests are identifiable: the user agent carries an HPE device build string, the Play Integrity API can verify the instance, and — the line to underline — "the Platform will be reported as Windows."
Then the commercial half, in Google's own words to ad providers: "you can choose to simply refrain from serving ads to requests from games on Google Play Games on PC." The same page notes that for ads promoting mobile apps, "users may not be able to install these apps on their PC."
Put those together. The single largest ad category funding casual Android games is app-install advertising, and its entire mechanism is broken on a Windows desktop: the click cannot end in an install. Buyers know it. Networks are explicitly permitted to skip this inventory. Nothing about this is a bug — it is a correctly functioning market pricing a surface where the primary conversion event does not exist.
So do not model PC revenue from your Android eCPM. Serve it, measure it separately by platform, and treat whatever comes back as found money. If your monetization stack cannot split reporting by platform, you will fold a low-fill surface into your blended numbers and quietly drag your own averages down.
🖥️ The Hardware Floor Is Not a Phone Floor
Google's Play Help page on PC requirements sets the minimum at Windows 10 (v2004), four physical CPU cores, 8GB of RAM, an SSD with 10GB free, and an Intel UHD Graphics 630 GPU or comparable. Recommended moves to eight logical cores and a gaming-class GPU. Both configurations require a Windows admin account and hardware virtualization enabled.
Those last two are the interesting ones, and they define the channel's shape. Admin rights plus virtualization rules out most managed corporate laptops, most school fleets, and most locked-down machines of any kind. This is a consumer-at-home channel. If you were sketching a B2B angle — games on staff desktops, a portal for an enterprise client — that door is closed by the install requirements, not by the catalogue.
It also means the audience skews toward people who deliberately installed an Android game runner on a decent PC. That is a narrower and more intentional population than "Windows users," and it behaves like a gaming audience, not a casual-arcade audience.
📅 What GDC 2026 Added, and Who It Was For
The March 2026 round of announcements, reported by TechCrunch on 11 March 2026 and set out on Google's own Android Developers Blog, pushed further in one direction: paid games. "Buy once, play anywhere" pricing lets a developer set a single price covering mobile and PC versions. Game trials let players try a paid title before buying and continue where they left off. The full purchase journey now completes inside Google Play Games on PC for PC-only titles. Google's Play Console programme page describes the trial as a 60-minute window.
Read the direction of travel honestly. Every headline feature of the last cycle serves premium and paid content. None of it does anything for a free, ad-funded arcade catalogue. Google is building a PC storefront that competes for paid game spend, and free casual titles are the volume that fills the shelves rather than the reason the shelves exist.
🚫 Five Ways Operators Get This Wrong
- Discovering the channel through a support ticket. A player emails about broken controls on a game you did not know was on Windows. That is a bad way to learn your distribution footprint.
- Opting the whole catalogue out on reflex. Untested availability costs nothing and occasionally converts. The correct response to an unmeasured channel is measurement, not deletion.
- Chasing the optimized badge across a large catalogue. Cloud save, Input SDK integration and PC asset sets, multiplied by hundreds of titles, is a budget with no revenue case behind it.
- Blending PC ad revenue into mobile reporting. Different fill, different formats, different conversion. If you cannot see it separately, you cannot decide anything about it.
- Assuming the licence covers Windows. "Android" in a 2019 or 2022 contract meant phones and tablets. Whether it now includes a Windows runtime that reports its platform as Windows is a question your agreement did not anticipate.
✍️ The Clause to Add Before Your Next Renewal
None of this is exotic contract drafting. Four points settle it:
- Name the surface. State whether the grant covers Android-compatible runtimes on non-Android operating systems, including Google Play Games on PC, and say so explicitly rather than relying on the word Android to carry it.
- Assign the opt-out. Whoever holds the Play Console account controls this. Write down who decides, and how the other party requests a change.
- Get modification rights for compatibility work. Not creative changes — permission to adjust manifest flags, input handling and UI strings for platform compatibility, plus the obligation on the licensor to supply builds that pass the playable checklist if they would rather do it themselves.
- Split reporting by platform. If revenue share is calculated on net receipts, PC receipts should be visible as their own line before anyone argues about them.
🧭 What to Do This Week
Install Google Play Games on PC on one machine. Search your own titles by name. Look at what badge each one carries, then open two or three and try to play them with a mouse. You will know within twenty minutes whether your catalogue is embarrassing on a desktop or merely unoptimised — and those two findings lead to completely different decisions.
Then pull your Play Console list and check which titles have PC availability on. Pick your five strongest performers, put them through the playable checklist, and leave the rest untested. Measure PC separately for a quarter. If it does not clear the cost of the checklist work, you have learned that cheaply, and the catalogue is still there.
🌲 Where a Licensed Catalogue Fits
Forestry Games has licensed games since 2017 and maintains a catalogue of 1,049 titles spanning HTML5 and Android APK builds, with titles also published on Google Play and the App Store. Questions of platform scope, build modification rights and per-surface reporting come up in every serious catalogue conversation, and they are easier to settle in the agreement than to litigate afterwards.
The practical takeaway is smaller than the channel sounds. Windows is not a new market for a casual Android catalogue — it is an existing default with poor discovery and uncertain fill. Treat it as a measurement exercise on five titles and a paragraph in your next contract, not a roadmap item. The operators who get hurt here are the ones who find out about the channel from a player, six months after Google turned it on for them.
