Android Game Source Code Is Cheap. The Obligations That Come With It Are Not.

Android game source code is cheap to buy and expensive to own. Google Play's repetitive-content policy, annual API deadlines and install costs decide the rest.

A finished Android game with full Unity source, art and a build guide sells on the code marketplaces for less than the cost of a decent laptop. That price is real. It is also the smallest number in the whole exercise, and treating it as the cost of entry is how people end up with fourteen half-finished builds, a suspended developer account and no revenue.

The reskin business — buy source, swap the art, publish, repeat — worked for a stretch. It does not work the way it used to, and the reasons are specific, dated and published. Worth walking through them before the next purchase, because a source-code licence is still a good buy for some businesses. Just not the ones who think they are buying a shortcut to distribution.

🗑️ Google Play Deleted Roughly Half Its Catalogue

Start with the number that reframes everything else. According to Appfigures data reported by TechCrunch in April 2025, Google Play went from about 3.4 million apps at the start of 2024 to about 1.8 million — a 47% decline. Over the same window Apple's App Store went from roughly 1.6 million to 1.64 million. This was not an industry-wide contraction. It was one store enforcing.

Google confirmed the causes to TechCrunch: the July 2024 ban on apps with "limited functionality and content", expanded verification requirements, mandatory testing for new developer accounts, more human review, and AI-assisted threat detection. In 2024 alone Google says it blocked 2.36 million policy-violating apps from being published and banned more than 158,000 developer accounts.

Note what did not happen: new releases kept coming. TechCrunch reported 10,400 releases on Play by late April 2025, up 7.1% year over year. The store is not closed to newcomers. It got a lot less tolerant of a particular kind of newcomer.

📋 The Policy Clause That Ends the Reskin Farm

If you run a portfolio strategy on Android, read Google Play's spam policy properly rather than the summary somebody posted in a Telegram group. Two clauses do the damage.

The first prohibits "copying content from other apps without adding any original content or value." The second is the one that catches reskin portfolios specifically: Google prohibits "creating multiple apps with highly similar functionality, content, and user experience," and adds that "if these apps are each small in content volume, developers should consider creating a single app that aggregates all the content." The overarching standard on Google's spam and minimum functionality policy is that apps must "provide value to users through the creation of unique content or services" rather than "merely provide the same experience as other apps already on Google Play."

Read that as a business rule, not a legal one. Ten reskins of one match-3 template, published under one account, with the same loop and different sprites, is the textbook description of what that clause exists to catch. The art swap is not the differentiator the policy is asking for. It never was — enforcement just used to be thin enough that it didn't matter.

There is a legitimate version of this. If each title genuinely has different content volume, different progression, different modes, you have a portfolio. If the only diff is a texture atlas, you have a policy exposure and a shared account risk: suspensions land on the account, not the app.

🔁 Every APK You Own Is a Rebuild Obligation, Annually

This is the cost that never appears in the purchase decision and always appears later.

Google Play's target API level rules mean a shipped APK is not a finished asset. Per Play Console Help's target API level requirements, from 31 August 2026 new apps and app updates must target Android 16 (API level 36) to be published, and existing apps must target Android 15 (API level 35) or higher to stay available to new users on devices running a newer Android version than the app targets. Miss it and the app does not vanish — it quietly stops being installable by most of the people you are paying to reach. An extension to 1 November 2026 is available on request.

That cycle repeats. Every title you own needs a toolchain, an engine version that still builds, a signing key you have not lost, and someone who can fix whatever the new API level broke. Multiply by portfolio size. Ten reskins is ten rebuilds a year, forever, whether or not any of them earned anything.

Then there is the toolchain bill itself. Unity cancelled the Runtime Fee on 12 September 2024 and went back to seat-based subscriptions — good news for anyone who was modelling per-install charges — but the same announcement raised Unity Pro pricing by 8% and Unity Enterprise by 25% from 1 January 2025, and set the Unity Personal eligibility ceiling at $200,000 in annual revenue and funding. If your source-code purchase is built on an engine version you cannot legally or practically build with, the licence you bought is a folder of files.

The account tax on top

Publishing itself now has a floor. Per Play Console Help, personal Google Play Console accounts created after 13 November 2023 must run a closed test with a minimum of 12 testers opted in continuously for at least 14 days before applying for production access, and must summarise real tester feedback as part of the application. Testers who opt out and rejoin restart the clock; the 14 days have to be consecutive.

Two weeks and a dozen real humans per title is survivable for one game. As a per-SKU cost across a volume strategy, it is the thing that quietly kills volume strategies.

🧰 What a Source-Code Licence Actually Hands Over

Buyers check the price and the screenshots. The parts that determine whether the purchase is usable are further down.

• Engine and plugin versions. Which exact editor version does this build in? Are the third-party plugins included, or are they licences you must buy separately and re-purchase per seat?
• Asset provenance. Art, audio and fonts often arrive under stock licences that permit use in the seller's product. That does not automatically transfer. Ask for the asset manifest and the licence for each item.
• Ad and analytics SDKs. Some packages ship with the seller's ad IDs still embedded, which means your revenue goes to them until you find it. Search the project for hardcoded keys before you build.
• Exclusivity and resale count. Non-exclusive source is sold to everybody. If forty other buyers publish the same build, Google's similarity detection is looking at all forty of you.
• Support scope and duration. "Includes support" usually means email replies for a fixed window, not a rebuild when the next API level lands.

The same due-diligence discipline applies here as anywhere in game licensing: the grant is a list of permissions, and anything not written down was not granted.

💸 The Install You Now Have to Buy

Assume you clear policy, clear testing and ship. Nobody knows the game exists.

Adjust's Gaming App Insights Report, 2026 edition, published March 2026, puts the global blended median gaming cost per install at $0.56 — up 30% year over year — with global average CPM up 20% to $4.34 and North American CPM at $15.98. Adjust's framing is that acquisition costs rising this fast is precisely why retention became the 2026 priority rather than volume.

Treat published CPI benchmarks with care generally. Trade blogs quote genre and platform figures several times higher than Adjust's blended median, because they are measuring different cohorts, different geographies and sometimes different definitions of an install. The honest read is a wide range whose direction everyone agrees on: up. What that means for a reskin is simple arithmetic. A generic title with no brand, no franchise and no organic pull is buying every single user, at a price that rose 30% in a year, against retention that a template-derived game rarely earns.

🏪 Where Source Code Genuinely Earns Its Price

None of this makes source-code deals bad. It makes Play Store arbitrage bad. Source ownership is worth real money in the places where you are not competing for organic discovery at all, because distribution is contractual rather than algorithmic.

• OEM and carrier preloads. Preinstallation is a negotiated slot, not a ranking. Epic's December 2024 deal with Telefónica to preload the Epic Games Store on Android devices across Spain, the UK, Germany, Mexico and Spanish-speaking Latin America is the visible end of a market that operates on the same logic further down: a device or operator partner wants content on the handset, and wants it customised. You cannot customise what you do not have the source for.
• Operator app stores and portals. Carrier catalogues need titles adapted to local billing, local language and local device profiles. See telecom and operator game portals for how that channel is structured.
• Offline and low-connectivity markets. An APK that runs without a connection is a different product from a web game, and the markets where that matters are large.
• Enterprise and kiosk fleets. Managed device estates, in-store units, training deployments. No store, no CPI, no ranking — a build spec and a purchase order.
• Rebranding for a client. Agencies and brands paying for a custom build need the code, the assets and the right to modify. That is a services business with a source dependency, not a portfolio play.

Notice the pattern. Every one of those is a case where somebody already has the audience and needs the content. That is the opposite of publishing a reskin and hoping.

⚖️ APK or HTML5 Is a Question About Which Obligations You Want

The comparison usually gets framed as reach or performance. The more useful frame is maintenance and control.

An APK gives you the Play surface, offline play, device APIs, push, and preload eligibility. It costs you an annual target-API rebuild per title, store review on every update, account-level suspension risk, and a publishing process with a two-week testing floor for new personal accounts.

An HTML5 build gives you instant play, one deployment across web, portals, smart TVs, in-flight systems and embedded surfaces, and updates that ship when you push them. It costs you the platform-specific capabilities and, on the open web, a harder discovery problem.

Most operators who run both end up using them for different jobs: HTML5 for reach, embedding and partner placement; Android builds for store presence, offline markets and preload deals. Picking one because it is cheaper to acquire is how you end up paying for the other one twice.

🚫 Five Ways Buyers Get Source-Code Deals Wrong

1. Budgeting the purchase, not the ownership. The licence fee is one line. Rebuilds, engine seats, store compliance and UA are the actual P&L.
2. Buying volume before proving one. Ten titles multiplies every recurring cost by ten before you know whether the first one retains anybody.
3. Publishing near-identical builds from one account. The similarity is the violation. Account suspension takes the whole portfolio, not the offending SKU.
4. Skipping asset provenance. Third-party art and audio licences are the most common hidden defect in a source package and the one that surfaces at exactly the wrong moment.
5. Assuming the store is the distribution plan. It is a shelf. Everyone with $2,000 and a template is on the same shelf.

🎮 Where Forestry Games Fits

Forestry Games has licensed games since 2017 and maintains a catalogue of 1,049 titles spanning HTML5 and Android APK, with titles also published on Google Play and the App Store. The company develops HTML5 games in-house, works with branded IP, and holds brand partnerships including Disney, Nickelodeon, Cartoon Network and Warner Bros. If you are weighing a source-code route against a licensed catalogue, the source code options are worth comparing against a straight licence side by side — including which maintenance obligations sit with you and which do not.

🧭 Before the Next Source-Code Purchase

Work through this in order, and stop at the first honest no.

1. Name the distribution. Who already has the audience — an operator, an OEM, a client, a portal you own? If the answer is "Google Play," you do not have a distribution plan.
2. Price three years of ownership. Annual target-API rebuilds, engine seats, store compliance work, and the testing floor per title.
3. Audit the package before payment. Editor version, plugin licences, asset manifest, embedded SDK keys, resale count.
4. Model the acquisition honestly. Use a benchmark with a named methodology, assume it rises, and check whether the game's retention could ever repay it.
5. Ask whether you need the source at all. If nobody is going to rebrand, re-skin or re-engineer the title, a licence to run it is cheaper and carries fewer obligations.

The businesses still making money on Android source code in 2026 are almost all in the second category from that list of channels — they sell into somebody else's distribution, and the source is what lets them say yes to a customisation request. The ones that struggled were running an arbitrage that Google spent two years and 1.6 million delisted apps closing. Buy the code because you have a customer for what you will do with it. That is the only version of this that has ever held up.