What a soft launch is for
A soft launch is a controlled release in limited markets to answer one question: does the product work well enough to scale? Not 'does it get downloads' — downloads are cheap. Does it retain, does it monetize, does the funnel hold? The soft launch is your chance to measure these with real users in real conditions while the stakes are small. Skipping it and launching globally means learning the same lessons at global prices — with a global review rating you cannot reset. The discipline is simple: prove the funnel in small markets, then buy scale in large ones.
Choosing test markets
Pick test GEOs that predict your target markets, not just cheap ones. The classic logic: choose markets similar to your launch targets in language, device mix, payment behavior and monetization potential — but small enough that a failed test costs little. Testing in a market with completely different user behavior teaches you the wrong lessons. Equally important is volume: your test GEO must deliver enough daily users to make retention cohorts statistically meaningful within weeks, not months. Two or three comparable markets beat one big one — they let you see whether results are product truth or market accident.
The metrics that decide
Retention is the first gate: if users do not come back, nothing else matters. Day-1, day-7 and day-30 retention tell you whether the product earns a place on the phone. Monetization is the second: ARPU, ARPPU, conversion to purchase, and the trend — is LTV rising with cohort age or flat? Funnel health is the third: onboarding completion, tutorial drop-off, first-session length, and where exactly users abandon. And the synthesis: projected LTV versus your target acquisition cost. If LTV does not clear CAC with margin in the test market, no global launch fixes that.
Paid UA during soft launch
Paid acquisition in soft launch is for measurement, not growth. Its job is to deliver enough users, at representative quality, to make your metrics trustworthy. That means buying traffic that resembles your eventual scale traffic — same channels, similar targeting — because metrics from incentivized or off-target test users lie about your product. Keep budgets capped: you are buying data, and the data has a fixed price regardless of how much you spend beyond it. The mistake to avoid is judging creative performance at soft-launch budgets — small-spend ad metrics are noisy; measure product metrics precisely and ad metrics directionally.
Go, iterate, or stop
The decision framework is honest and brutal. Go: retention and monetization clear your thresholds with margin — scale into launch markets with the proven funnel. Iterate: one gate fails by a fixable margin — retention weak but monetization strong suggests onboarding problems; monetization weak but retention strong suggests pricing or offer problems — fix and retest. Stop: both gates fail, or the LTV-CAC gap is structural — the market is telling you something that more budget will not change. The teams that win are not the ones whose first launch works; they are the ones who read the soft launch honestly and act on it fast.