How to Distribute Android Apps Beyond Google Play in 2026: Amazon, Samsung, Huawei, Direct APK

Why Fora Soft wrote this playbook

We’ve been shipping Android apps for 20 years — 625+ delivered projects, 100% Upwork success, and a portfolio packed with multi-store releases. From Franchise Record Pool (DJ music streaming on Play + direct-download for licensed catalogs) to BrainCert (global LMS shipping on Play + AppGallery for APAC and MENA customers), our engineers have wired every Android distribution channel that matters.

This playbook is not a vendor brochure. It’s the internal decision framework we use with clients scoping multi-store or sideloaded distribution — covering the technical work (signing, updater, Play-Asset-Delivery parity), the commercial math (fees, region mix, FX), and the regulatory angles (EU DMA, Russia, China, KSA, Brazil).

Because we deliver with Agent Engineering — specialist agents running in parallel on CI/CD, store builds, billing, crash analytics, and QA — a typical hybrid-distribution rollout lands in 3–6 weeks rather than the 10–14 weeks a traditional agency plans. That’s the lens behind every cost and timeline below.

Why distribute beyond Google Play in 2026

Four economic and regulatory forces make “Play only” a strictly worse strategy than it was five years ago.

1. Billing fees. Google Play charges 15% for the first $1M annual subscription revenue per developer and 30% above that. Amazon Appstore sits at 20%. Samsung Galaxy Store runs 0–30% depending on tier. Huawei AppGallery is 15–30%. Self-hosted direct-download keeps the full revenue minus payment processing (2–4%). For a $5M ARR app, the swing between Play-only and a balanced mix can be $300K–$900K/year.

2. Regional coverage. AppGallery is mandatory in China and dominant on Huawei’s 580M+ active devices — strong in MENA, CIS, and parts of LATAM. Samsung Galaxy Store is pre-installed on 400M+ active Galaxy devices globally. RuStore and other sovereign stores fill gaps in jurisdictions where Play is restricted.

3. Policy flexibility. Play has the tightest content and monetization rules. Amazon is permissive on crypto apps; Samsung is friendly to gaming hubs; AppGallery allows adult content in regions where laws permit. Several cast-iron use cases only ship outside Play.

4. Regulation is opening sideloading. EU Digital Markets Act forces Google to allow alternative billing and sideloading on Android. The US Epic v Google verdict is reshaping what Play can mandate. Ship-once-install-everywhere is becoming a real-world product strategy, not a theoretical one.

The realistic shortlist of Android distribution channels

Five channels cover >98% of legitimate Android installs globally. Everything else is either micro-niche or piracy-adjacent.

Regional stores worth considering on a case-by-case basis: RuStore and Yandex Store (Russia), ONE store (South Korea), APTOIDE (indie / LATAM), F-Droid (open-source / FOSS audiences), Itch.io and Epic Games Store on Android (indie games). None hit >1% global share but matter inside their niche.

Amazon Appstore: sleeper win on Windows 11

Amazon Appstore used to be the Fire Tablet store. Two things made it interesting in 2024–2026: Windows 11’s Subsystem for Android (WSA) ships Amazon Appstore as the default; and Amazon relaxed several content policies where Play is restrictive.

What ships well. Kids’ apps (Amazon FreeTime audience), utilities, productivity on Windows 11 tablets, crypto wallets, and streaming players that Play may reject under adjacent policy rules. Fire Tablet install base is still meaningful in North America and UK education markets.

Engineering work. Amazon accepts standard APK and AAB. If you use Google Play Services APIs, swap for Amazon Device Messaging (ADM) and Amazon In-App Purchasing (IAP) — the Amazon Appstore SDK handles both with thin wrappers. Signing can use your existing key; Amazon doesn’t mandate re-signing.

Submission. Binary upload, review in 1–5 days, rejections usually policy-related rather than technical. Fire Tablet compatibility requires additional testing (no Google services, different screen densities).

Samsung Galaxy Store: 400M devices, gaming-friendly

Samsung Galaxy Store ships pre-installed on 400M+ active Galaxy devices globally and is the primary store in Samsung’s own promotional surfaces (Bixby Home, Galaxy Themes, pre-boot).

What ships well. Gaming (Galaxy Store has a strong gamer audience and a GameLauncher promo program), edge-case categories Play restricts, Bixby skills, pen-first and foldable-first experiences, and Watch / One UI companion apps.

Engineering work. Standard APK/AAB with Samsung-specific manifest tweaks if you target the Galaxy Themes or Galaxy Watch surface. Samsung IAP SDK integrates via a 200-LOC wrapper; it coexists cleanly with Google Play Billing inside flavors.

Fee structure. 0–30% depending on monetization choice and volume. Apps using Samsung IAP exclusively sit at the lower end. Revenue share on gaming can drop further for volume titles participating in Samsung’s promotional programs.

Huawei AppGallery: mandatory for China, real upside elsewhere

Huawei AppGallery reports 580M+ MAU with a dominant position in China, growing strength in MENA (UAE, KSA), and traction in LATAM (Mexico, Colombia, Peru) and CIS markets. For any B2C app with China, MENA, or LATAM ambitions, it’s the second most important Android store after Play.

Engineering: HMS Core instead of Google Play Services. Huawei devices without Google services require HMS Core: Push Kit replaces FCM, Account Kit replaces Google Sign-In, Map Kit replaces Google Maps, Location Kit, ML Kit, and so on. HMS APIs mirror Google equivalents and Huawei provides conversion guides + a conversion tool (HMS Toolkit) that automates ~70% of the work for typical apps.

Build flavors. Ship three flavors: Google (GMS), Huawei (HMS), and a “unified” flavor for emerging markets where both are needed. Gradle product flavors plus agconnect-services.json vs google-services.json at build time.

Submission. Review typically 1–3 business days. Huawei requires localized metadata for China (simplified Chinese copy, Chinese screenshots, ICP license for apps targeting mainland). Outside China, English metadata is fine.

Commercial upside. AppGallery’s new-developer programs regularly drop fees to 0–15% for the first 6–12 months in priority categories (gaming, fintech, health). A sensible play: ship AppGallery as region-targeted for MENA and LATAM, ride the promotional fee window, and decide in year two whether to keep it or retire.

Direct APK / AAB: the highest-margin channel

Hosting the APK on your own site and accepting payment directly (Stripe, Paddle, Braintree) keeps the full 30% that would otherwise go to Play. For a $50 ARPU subscription app at 100K subscribers, that’s a $1.8M/year swing.

What ships well. SaaS with web-first signup funnels (Slack, Notion, Discord have all run sideload experiments), adult content, crypto wallets, grey-area categories, and any app whose audience is already on your marketing site.

Engineering: the updater problem. You lose Google Play’s silent update mechanism, so you must ship your own. Two clean patterns:

Pattern A — manifest-based. App periodically fetches a signed manifest (version, URL, SHA-256). If higher version: download APK, verify signature + hash, prompt user, launch PackageInstaller. Works on all Android 10+ with REQUEST_INSTALL_PACKAGES permission and verified installer identity.

Pattern B — owner-updater. Apps you control (kiosks, enterprise deployments) can use device owner / profile owner APIs to silently update without user prompt. Requires MDM enrollment or work-profile deployment.

Signing and integrity. Sign with a V4 + V3 + V2 APK signature scheme. Host on HTTPS with HSTS. Include SHA-256 in a signed manifest served from a different origin than the APK. Verify signatures on device before install.

Android AAB reality. Play Asset Delivery is Play-only. For direct distribution you ship a universal APK (larger) or implement your own delivery (split APKs by ABI / density / language). Bundletool generates these from your AAB; most apps keep universal APK for simplicity and pay the 30–60MB size tax.

Hybrid distribution — how to wire it correctly

The goal is one source-of-truth codebase producing N store-specific artifacts from one CI job — not N forks slowly drifting.

1. Gradle product flavors. Model each distribution as a flavor: google, amazon, samsung, huawei, direct. Flavor-specific source sets override billing, push, and analytics. Flavor-specific BuildConfig fields expose the current store to runtime checks.

2. Billing abstraction. Implement a thin IBillingClient with swappable Google Play Billing / Amazon IAP / Samsung IAP / Huawei IAP / Stripe backends. The app code never sees the concrete implementation.

3. Push abstraction. IPushClient with FCM / ADM / HMS / self-hosted (Gotify, UnifiedPush) backends. Device registration tokens flow through a single backend endpoint that routes send requests to the right service.

4. Crash analytics abstraction. Firebase Crashlytics works on GMS devices; on HMS devices use AppGallery Crash or self-hosted Sentry. Sentry is increasingly the portable default.

5. CI/CD pipeline. GitHub Actions or GitLab CI: single workflow builds all 5 flavors, runs flavor-specific tests, signs each with the right keystore, and uploads to each store via their API (Play Developer API, Amazon Developer API, Samsung Seller Office API, AppGallery Connect API). Failed uploads alert Slack; successful uploads open a release ticket for manual approval.

6. Feature flags for store-specific policy. A single ENABLE_EXTERNAL_PAYMENTS flag, flipped per flavor, disables the direct-payment links required by EU DMA on Play builds outside EU and enables them on direct / Huawei / Samsung builds everywhere. This saves you from 2024–2026 Play rejection loops.

Billing and compliance — the landmines

1. Play’s anti-steering rules. Google Play restricts linking to external payment from inside the app. The EU DMA carved out exceptions for EU users since March 2024; the US Epic v Google verdict (December 2023) is forcing similar changes. If your app uses direct billing, ship flavor-specific copy and screens: Play EU/US gets an “Also available on our site” CTA; Play in other regions doesn’t.

2. Tax compliance on direct sales. When Play charges a user in Germany, Google handles VAT. Sell directly and you collect and remit VAT yourself (or via Paddle/Chargebee/LemonSqueezy as merchant-of-record). Similar story for Brazilian CNPJ + ICMS, Canadian GST/HST, Indian GST. Merchant-of-record services charge 5–7% and handle it all — usually worth it vs a 30% Play fee.

3. China: ICP license. Any app distributed inside mainland China needs an ICP filing, a local business entity, and (for gaming) a Banhao license. Non-trivial; most foreign developers ship AppGallery only to non-China regions and skip mainland until revenue justifies the entity.

4. Russia: RuStore and payment routing. Since 2022, Russian users can’t pay Play subscriptions with Russian cards. RuStore is pushed as replacement. Most Western developers either exit Russia or ship free-tier only; a few with Russian subsidiaries monetize via RuStore.

5. Subscription migration. A user subscribed on Play and later downloads the direct APK shouldn’t pay twice. Entitlement sync via your backend (account-bound, not device-bound) is essential. Most apps we’ve migrated used Google Play Billing’s purchase receipts plus Stripe’s customer IDs, reconciled nightly.

Security, tamper-protection, and store fraud

Once an APK is hosted anywhere but Play, copies appear on piracy sites within days. Three mitigations work.

1. Play Integrity / SafetyNet replacement. Google’s Play Integrity API issues device attestations that your backend verifies. On non-Play flavors, use Firebase App Check (HTTPS token) plus your own device-fingerprint layer (hardware IDs, SafetyNet on GMS devices, HMS Safety Detect on Huawei).

2. Root / emulator detection. Libraries like RootBeer or commercial tools (Appdome, Promon SHIELD, Guardsquare DexGuard) detect rooted devices, known emulators, and patched binaries. Ship these everywhere but tune thresholds — false positives alienate legitimate enthusiasts.

3. Receipt validation server-side. Never trust client-reported entitlement. Every IAP goes through a backend verification call to Google Play Developer API, Amazon RVS, Samsung IAP receipt server, Huawei IAP verification, or Stripe webhook. Entitlement flips only after verification success.

4. Watch for fake stores. Third-party aggregators like APKMirror, Aptoide, Uptodown index your APK automatically. Some add malware. Monitor DNS typosquats and APK aggregators monthly; issue DMCA if tampered copies appear.

Analytics and attribution across stores

Play Install Referrer attributes Play installs. Amazon has InstallReferrerClient equivalent. Samsung provides install_referrer via manifest metadata. Huawei provides agconnect-installreferrer. Direct APK links can carry UTM parameters via Intent extras or a first-launch deep link.

Attribution aggregators like AppsFlyer, Adjust, and Singular normalize across stores. They add 1–3% of revenue in fees but save months of custom wiring. For most apps above $1M ARR, the ROI is positive.

Report store performance as a single dashboard: MAU by store, LTV by store, crash-free rate by store, IAP conversion by store. Each store has its own seasonality and promotional calendar that affect the numbers.

Cost model: multi-store wiring for an existing app

Starting from an existing Google-Play-only Android app, typical scope and cost with Agent Engineering:

A typical subscription app with $1M ARR recovers the cost of full hybrid rollout in 4–12 months through lower fees on direct and regional channels. Above $5M ARR the ROI is absurdly positive — often >10x.

Mini case: +22% installs and +19% net margin in 5 weeks

Situation. A video streaming app with 180K MAU and $2.1M ARR, Play-only, saw EU DMA go live and realized they could start steering checkout off-Play. They wanted a hybrid plan live in a quarter.

5-week plan. Agent Engineering ran specialist tracks in parallel: flavors and billing abstraction, Amazon Appstore integration, AppGallery + HMS conversion for MENA, direct APK with self-updater and Stripe MoR via Paddle, CI/CD pipeline for all four flavors, and content-design work for Play-EU and Play-non-EU steering copy.

Outcome. Week 5 launch across Play + Amazon + AppGallery + direct. Over the next 90 days: +22% total installs (AppGallery drove MENA growth that Play was missing), +19% net margin on Play EU users who switched to direct checkout, and the $42K total rollout cost paid back by month 3. The same team now ships monthly to five channels from one codebase. Want a similar plan for your app?

Decision framework — pick your mix in five questions

1. What’s the monetization model? Subscription or premium IAP — big upside from direct + Play billing mix. Ads-only — multi-store mostly helps reach, not margin. Free tool — stick to Play + one regional.

2. Which regions matter in 18 months? China: AppGallery mandatory. MENA/LATAM/CIS: AppGallery worth the work. US/EU: Play + direct. Windows 11 tablet users: add Amazon.

3. What’s your audience’s tech savvy? Sideload-friendly audiences (gaming, crypto, power users) take direct APK easily. Consumer mainstream needs Play-dominant with “also available” gentle steering.

4. Which Play policies constrain you? If Play rejects or restricts your category (crypto, adult-adjacent, certain gambling), non-Play is not optional — it’s your entire distribution. Plan accordingly.

5. What’s engineering capacity? 3-person team with one Android engineer — ship Play + direct first, add Amazon + Samsung in Q2. 6+ engineers — ship all five channels in one quarter.

Pitfalls we see kill multi-store rollouts

1. Forking the codebase. “Let’s make an AppGallery version” becomes a parallel fork that drifts. Always use Gradle flavors + shared source; never hard-fork.

2. Hidden Google Play Services. Libraries like Firebase Crashlytics, Firestore, or Google Maps silently pull GMS and crash on HMS devices. Audit all transitive dependencies; swap to multi-provider abstractions early.

3. Ignoring the updater on direct. Users install the APK and never see v1.1. Ship a manifest-based updater on day one or the experience is broken.

4. Double-billing subscribers. Same user re-subscribes via Stripe because your backend didn’t recognize the Play entitlement. Ship account-based entitlement across all stores before you launch any new channel.

5. Missing the review-policy shift. Play updates billing / steering policy every 3–6 months. Subscribe to the Play Developer blog and Android Police; keep a compliance agent reviewing changes quarterly.

KPIs to track per-store from day one

Acquisition KPIs. Impressions, install conversion, CPI per store / region, organic share, ASO ranking for top 5 keywords.

Monetization KPIs. ARPU by store, trial-to-paid conversion by store, net-of-fee revenue per user, churn by store. The net-of-fee metric is what tells you whether Play at 15% is really beating direct at 3% processing + 6% MoR = 9% all-in.

Reliability KPIs. Crash-free rate per store (often differs because of device mix), update-adoption rate per store (slow adoption on direct means broken updater), store-review rating + response latency.

When NOT to leave Google Play

We tell clients to stay Play-only when three conditions hold.

First, when ARR is under $500K and growth is constrained by product, not fees. A 30% reduction on thin revenue doesn’t fund engineering; product work is the better use of capacity.

Second, when the audience is US/EU mainstream consumer with no strong regional or policy drivers. Play covers 98%+ of your addressable market; the extra work doesn’t earn its keep.

Third, when Play policy risk is minimal and you don’t need the hedge. Some apps genuinely do not benefit from having an escape hatch; ours isn’t religion.

A 6-week hybrid-distribution delivery roadmap

The plan below is how we ship a Play + Amazon + AppGallery + direct rollout under Agent Engineering. Traditional teams typically run this in 12–16 weeks.

FAQ

What to Read Next

Ready to ship your Android app on more than just Google Play?

Multi-store Android distribution in 2026 is a pragmatic lever — not a religious war. Keep Google Play as your baseline, add Amazon for Windows 11 + Fire, Samsung for Galaxy-specific features, AppGallery for China / MENA / LATAM, and direct APK for high-ARPU subscriptions. One Gradle-flavors codebase, one CI/CD pipeline, five targeted channels.

If you’re scoping a hybrid rollout, our 6-week plan is battle-tested across video streaming, e-learning, music, and SaaS apps. If you just need an honest build-or-skip call for AppGallery or direct APK, we’ll tell you that too — 30 minutes, no sales deck.