Mobile App Growth Blog

A creator posts, your installs rise, and nothing in your dashboard connects the two. The gap is almost never the creator — it is that the link, the listing and the redemption path were designed after the deal was signed instead of before it. This is the mechanics side: what the stores actually hand back to you, and what no amount of briefing can recover.

Uninstall rate is the retention metric everyone quotes and almost nobody defines. Google Play folds 30 days of device inactivity into the same number as a genuine uninstall, Firebase counts a different population entirely, and on iOS the metric you think you are reading does not exist. Here is what each source actually measures, and the one Play threshold that carries a documented consequence.

Every list of the cheapest CPI countries traces back to another list, and none of them trace back to a source you can check. The more useful question is why a market is cheap in the first place, because the same conditions that suppress the install price usually suppress the revenue behind it. This is how to judge a market from data you actually own.

Community distribution fails in a predictable way: the account gets limited, the server gets reported, or the store flags the reviews you asked for. Almost every one of those outcomes is described in a published rule you can read before you write the first message. This is what those rules actually say, and where they cannot be quoted at all.

You doubled the budget and the cost per install went up, so someone concluded the campaign is broken. It usually is not. Audience depth is finite, a tighter target narrows the pool you are allowed to buy from, and the edits people make in a panic cost more than the saturation they were meant to fix.

Somebody has sent you a slide with a push opt-in rate and an open rate on it, and asked why your app is below both. The honest answer is that the two numbers were almost certainly not measured the same way yours are. This piece explains what makes cross-app push benchmarks structurally incomparable, and what to build instead.

Every performance team has been told that raising the budget resets the algorithm. Google does not document it that way — the triggers it publishes for the Learning status are bid strategy events, and the budget cautions live in a different place with different conditions attached. This is the operational method for stepping spend up without paying for the same calibration twice.

Every Indian app team eventually asks the same question: TikTok is gone, so where does the short-video budget go? Most answers you will find name a list of domestic platforms without showing you a single line of their advertising documentation. This one is built the other way round, from the ad managers whose rules we could actually read.

AI apps can grow revenue while losing money on their heaviest users. This guide shows founders how to cost every AI action, model the usage tail, set honest limits and price for durable contribution margin.

Firebase says one install count, your MMP says another, and Play Console says a third. Your event has vanished into a row marked "(other)". Nothing is broken. Firebase is a batched, aggregated reporting system and you are reading it like a database — this is what it is actually doing, and which of the three numbers belongs in a board pack.

You uploaded the build, your colleague installed it in minutes, and the fifty external testers you promised a link to have nothing. Nothing is broken. TestFlight has a documented set of build states, internal and external distribution follow different rules, and almost every "TestFlight is stuck" complaint is one of five states with a named fix.

Revenue fell off a cliff overnight, the Policy centre says ad serving is limited, and Google will not tell you why. There are exactly three documented causes, they call for completely different responses, and one of them is not a punishment at all. Here is how to tell which one you have — and what carries on paying while it runs.

The budget is set, the campaign is enabled, and Google is spending nothing — or it was spending fine until Tuesday. These are not the same problem, and neither is the campaign that spends happily while your CPI climbs every week you scale it. Four distinct failures, each with its own diagnosis, and a triage order that stops you from changing settings that were never wrong.

Money left the account, the campaign is live, and two dashboards disagree by 35%. Somebody is going to ask which number is true, and the honest answer is that all of them are — because Meta, your MMP, Play Console and Google Ads each count a different event and call it an install. Here is what each one actually measures, how much gap is normal, and how to reconcile them in half an hour.

The chart says you lost more users than you gained, and your app is not dying. Play Console counts a device that has simply gone quiet for 30 days as a loss, and counts an acquisition far more strictly than it counts a departure. Both definitions are published. Once you read them, the crossing lines stop being alarming and start being useful.

You submitted on Monday, it is Thursday, and the status has not moved. Apple publishes no review-time commitment, so nobody can honestly tell you how long it will take — but Apple does publish every status your app can be in, and most of them are asking you for something specific. Here is what each one actually means and which ones you are supposed to act on.

Nobody puts crash rate on the growth dashboard, so nobody sees it charging rent. Google reduces the visibility of apps that cross its published thresholds, can put a warning on the store listing visitors are reading, and says outright that a larger app hurts install success. Every one of those raises what you pay for a user — and Google Ads itself tells advertisers to go and check vitals.

Sign-in works on your device, works on the tester’s device, and fails for every real user the day you go live. Nothing in your code is wrong. Google Play re-signs your app with a key you have never seen, and the fingerprint you registered with Google Cloud belongs to a different certificate entirely. Three keys are in play, and knowing which is which resolves this in ten minutes.

Your rating fell further than the new reviews should allow, and you cannot work out why. Google Play does not show users a lifetime average — it shows a figure weighted toward recent ratings, which is why a bad fortnight moves it so hard. The same mechanism is why recovery is faster than the arithmetic suggests, and there is a documented way to accelerate it.

Google will not accept new apps or updates on Billing Library 7 or below from 31 August 2026, and there is a documented extension route to 1 November if you need it. Your live app keeps working either way — what you lose is the ability to ship. Here is exactly what version 8 removed, why restore purchases is the hard part, and the order to migrate in.

Installs fell off a cliff, the app is still live, nothing was released, and nobody changed a campaign. There are seven documented causes, two of which restrict your distribution without removing your app, and one of which quietly makes you unavailable on newer devices. Here they are in the order that costs you the least time to check.

Your payback period is six months and your LTV:CAC looks healthy, so the model says spend more. The model is not wrong — it is answering a different question. Ad spend leaves your account in days, store revenue arrives up to 45 days after the month it was earned, and the platform keeps its share first. That gap, not your payback period, sets your ceiling this month.

Every article about buying installs says "you could get banned" and none of them describes what being banned is. Google publishes four distinct enforcement actions with sharply different consequences — one keeps your ratings, one destroys them permanently, and one takes every account you own. This is what each costs and what survives.

Your CRM reports 94% delivered. Your session data says almost nobody opened the app. Both are correct, because "delivered" in a push dashboard does not mean what founders think it means. This is the whole Android delivery chain, the point at which each link breaks, and how to prove where yours is failing.