How to Soft Launch a Mobile Game Without Burning $20K Learning the Hard Way
How to soft launch a mobile game the disciplined way: which countries to test in, the D1/D7 and CPI numbers to hit, what it costs and when to scale, iterate or kill.
A soft launch is a limited release of your mobile game in one or more small test countries, with a modest paid budget, so you can measure real retention, monetization and acquisition cost before you commit to a global launch. It isn't about revenue. It's about buying the data that tells you whether to scale, iterate or kill the project.
We've run soft launches for our own titles and for publishers, and the pattern doesn't change: teams that skip or rush this step find their problems in the US at ten times the cost. Done properly, a soft launch takes 6 to 12 weeks, costs $5,000 to $40,000 in media depending on genre, and ends with a go or no-go decision backed by numbers instead of hope.
Quick answer: soft launch in the Philippines first for cheap retention data, then in Canada or Australia for monetization and CPI that resemble the US. Buy 2,000 to 5,000 installs per round, measure D1, D7 and early LTV against your genre benchmarks, and make a go, iterate or kill call at each gate. Budget $5,000 to $40,000 over 6 to 12 weeks.
Why soft launch a mobile game at all?
Internal playtests and friends-and-family TestFlight builds tell you the game runs. They can't tell you whether a stranger who tapped a Meta ad will come back tomorrow, or what that stranger costs. Those two numbers, retention and CPI, decide whether a free-to-play game is a business.
A soft launch also protects your global launch. Store featuring, ratings and the first wave of reviews are all tied to your worldwide release. If the first US cohort hits a broken tutorial or an aggressive interstitial, that rating stays with you for years. Fixing it in a test market where almost nobody will remember is far cheaper. Our QA and testing team treats soft launch as the last and most important QA phase: real devices, real networks, real players.
Which countries should you soft launch in?
Pick test markets for what you want to learn, not for convenience. Cheap markets answer retention and technical questions. Expensive, US-like markets answer monetization and CPI questions. Most studios run two phases.
| Country | Typical CPI (casual) | Best for | Watch out for |
|---|---|---|---|
| Philippines | $0.10 to $0.40 | Early retention, crash rates, FTUE drop-off, big English-speaking sample | Monetization far below US; low-end Android skews performance data |
| Canada | $1.50 to $4.00 | US-like monetization and behavior, iOS share, CPI proxy | Small population; audiences saturate after a few weeks |
| Australia / New Zealand | $1.50 to $4.50 | High ARPDAU, strong iOS, close US proxy | Time zone makes live monitoring from the US awkward |
| Nordics (Sweden, Norway, Denmark, Finland) | $1.50 to $5.00 | High-spending players, strong iOS and Android mix, English fluency | Expensive; use to confirm monetization, not to tune retention |
| UK or Ireland | $1.00 to $3.50 | Larger English sample before the US | Many publishers count the UK as a launch market, so you burn a real audience |
A typical sequence looks like this. Philippines (sometimes with Brazil or Indonesia for Android scale) for 2 to 4 weeks to fix retention and stability. Then Canada and Australia for 3 to 6 weeks to confirm monetization and CPI. Then a Nordic market if you want a second look at high-value players. Don't soft launch in the US unless you have a big budget and a specific reason; you'll pay far more per install and poison your most valuable market with an unfinished product.
What KPIs should you track during a soft launch?
Every KPI needs a threshold written down before the launch starts. Otherwise you'll rationalize whatever number comes back. We've watched teams do exactly that. These ranges reflect what we typically see for reasonably polished games; your genre and audience will shift them.
| KPI | Hyper-casual | Casual / puzzle | Mid-core | Why it matters |
|---|---|---|---|---|
| D1 retention | 35 to 50% | 35 to 45% | 30 to 40% | Is the first session good enough to come back? |
| D7 retention | 8 to 15% | 12 to 20% | 12 to 20% | Does the core loop hold beyond novelty? |
| D30 retention | 2 to 5% | 5 to 10% | 6 to 12% | Will anyone still be here to monetize? |
| CPI (Canada / Australia) | $0.30 to $1.00 | $1.50 to $4.00 | $3.00 to $10.00 | What does a player cost in a US-like market? |
| Tutorial completion | 85%+ | 80%+ | 70%+ | Where the first-session drop happens |
| Crash-free sessions | 99%+ | 99%+ | 99%+ | Stability before anything else |
| Payer conversion (D7) | n/a | 1 to 3% | 2 to 5% | Does the economy give a reason to buy? |
Two numbers decide most soft launches: D1 and D7. D1 is about the first session and onboarding. D7 is about whether the loop has depth. A game with 45 percent D1 and 8 percent D7 has a great hook and nothing behind it. A game with 30 percent D1 and 18 percent D7 has a slow start and a loop that works, and that's usually the easier problem to fix.
Monetization KPIs (ARPDAU, ARPPU, early LTV) only mean something in the second phase, in Canada, Australia or the Nordics, with at least 2,000 installs per cohort. Read monetization from a Philippines cohort and almost any game looks unprofitable.
How much does a soft launch cost?
Most of the cost is media. The part teams forget to budget is the engineering and design time to react to what the data says.
- Phase 1 media (retention): 2,000 to 5,000 installs in the Philippines at $0.10 to $0.40 each is roughly $500 to $2,000 per iteration. Plan for 2 to 3 iterations.
- Phase 2 media (monetization and CPI): 2,000 to 5,000 installs each in Canada and Australia at $1.50 to $4.00 is $6,000 to $20,000 per round. Mid-core games cost more.
- Creative production: at least 10 to 20 video and playable variants to find a workable CPI. Budget $1,000 to $5,000; our guide to making ad creatives with AI shows how to keep this lean.
- Analytics and MMP: an attribution partner such as AppsFlyer, Adjust or Singular, plus product analytics. Free tiers cover most soft launches; expect $0 to $1,000 per month.
- Development during soft launch: two developers and a designer reacting to data for 2 to 3 months. At our dedicated-developer rate of $25 per hour that's typically $12,000 to $25,000. At US agency rates, multiply by 3 to 5.
- Total: a casual game usually lands between $15,000 and $40,000 all-in. Hyper-casual can be done for $5,000 to $10,000. Mid-core and strategy often run $40,000 to $100,000 because of higher CPI and longer measurement windows.
How long should a soft launch run?
Long enough to see D7 for at least two cohorts after your last meaningful change, and D30 for at least one cohort if you're judging monetization. That puts the minimum around 6 weeks and the typical range at 8 to 12. Hyper-casual studios often decide in 2 to 3 weeks on D1 and CPI alone, because their economics don't depend on D30.
The usual failure isn't running too short. It's running too long. A soft launch that drags on for 6 months of small improvements almost always means the team is avoiding the kill decision. Set an end date before you start and hold to it.
How do you decide: go, iterate or kill?
We use a simple gate at the end of each phase. Write the numbers down before you start and share them with everyone involved, including whoever is funding the game.
Gate 1: after the retention phase (Philippines, weeks 2 to 4)
- Go to phase 2 if D1 is within 5 points of your genre benchmark, D7 is within 3 points, and crash-free sessions are above 99 percent.
- Iterate if D1 is close but D7 is well below benchmark. The loop needs depth: more content, better progression pacing, clearer goals. Fix it, launch a fresh cohort, measure again.
- Kill or pivot if D1 is more than 10 points below benchmark after two iterations. The hook isn't working, and no economy tuning fixes a game people won't play twice.
Gate 2: after the monetization phase (Canada or Australia, weeks 5 to 10)
- Go global if projected D90 LTV is at least 1.2 to 1.5 times blended CPI in a US-like market, retention held, and you have a 3 month content plan ready.
- Iterate if retention is fine but LTV is short. That's an economy, pricing or placement problem, and it's usually fixable with a LiveOps pass: a better first offer, rewarded placements at the right moments, a starter pack, a battle pass.
- Kill if LTV is below 0.7 times CPI after two monetization iterations and creative testing hasn't moved CPI. The market is telling you this game can't be bought profitably.
Don't change several things at once between cohorts. If you rework the tutorial, add a starter pack and change interstitial frequency in the same build, you won't know which change moved D1. One or two changes per iteration, measured against a clean cohort, feels slower and is much faster overall.
What should be ready before the soft launch starts?
A soft launch only produces useful data if the instrumentation is complete and the build is stable. This is the pre-flight list we run with every client.
- Analytics events for install, tutorial steps, level or session completion, first purchase, ad impressions per placement and session length, verified end to end on both platforms.
- An MMP installed and tested with a real campaign, so you can see retention and revenue per source, campaign and creative.
- Remote config for at least the tutorial flow, ad frequency, first-offer pricing and difficulty curve, so you can test without a store resubmission.
- Crash reporting (Firebase Crashlytics or similar) with alerts.
- A store listing, screenshots and a short video that match the ad creative.
- One dashboard the whole team looks at daily: installs, CPI, D1, D7, ARPDAU, crash-free rate.
- Written go, iterate and kill thresholds, agreed by the people funding the game.
Test retention in a low-CPI market like the Philippines, then confirm monetization and acquisition cost in Canada, Australia or the Nordics. Measure D1, D7 and early LTV against benchmarks you set before launch. Give it 6 to 12 weeks and $5,000 to $40,000 in media, and be willing to make the kill call. The games that reach global launch after a disciplined soft launch are the ones that survive the first 90 days.
Frequently asked questions
What is a good D1 retention for a soft launch?
Typically 35 to 45 percent for casual and puzzle games, 30 to 40 percent for mid-core and 35 to 50 percent for hyper-casual. Below 25 percent at D1 usually means the first session or the tutorial needs rework before anything else.
Which country is best for soft launching a mobile game?
The Philippines is the most common first market for retention testing because installs are cheap and the audience speaks English. Canada and Australia are the most common second markets because player behavior and spending resemble the US. Nordic countries are used when you need high-value monetization data.
How many installs do you need for a soft launch?
At least 2,000 installs per cohort to read D1 and D7 with reasonable confidence, and 5,000 or more to read payer conversion. Smaller samples swing by several points from day to day and lead to bad decisions.
Can you soft launch on iOS only or Android only?
Yes. Many studios start on Android in the Philippines because installs are cheaper and review cycles are faster. Still run a smaller iOS cohort in Canada or Australia before global launch, because iOS monetization and CPI are usually different and iOS often drives most of the revenue.
What happens after a successful soft launch?
Global launch, a push for store featuring, a scaled UA plan built on the creatives that won in testing, and a LiveOps calendar for the first 90 days. The soft launch build should already have the analytics, remote config and event systems LiveOps will rely on.
Tell us your genre, platforms and budget and we'll come back within 48 hours with a soft launch plan: test markets, KPI thresholds, an instrumentation checklist and a media estimate.
PLAN MY SOFT LAUNCH→
