Framework
Cheap installs and expensive paying users are the same problem.
The diagnostic below tells you which of three things is broken, in the order to check them: tracking, creative, then onboarding and paywall. It uses the numbers you already have in your ad accounts and your analytics. Read it here, or take the PDF.
Pull these numbers
For the last 7 and the last 30 days, per platform (Meta, Google App campaigns, Apple Search Ads). If any column is empty, that is the first finding.
| Number | Meta | Google App campaigns | Apple Search Ads |
|---|---|---|---|
| Spend | 7 d / 30 d | 7 d / 30 d | 7 d / 30 d |
| Installs reported by the platform | 7 d / 30 d | 7 d / 30 d | 7 d / 30 d |
| Installs reported by your MMP (AppsFlyer, Adjust or Tenjin) | 7 d / 30 d | 7 d / 30 d | 7 d / 30 d |
| Installs reported by the stores | 7 d / 30 d | 7 d / 30 d | 7 d / 30 d |
| App opens | 7 d / 30 d | 7 d / 30 d | 7 d / 30 d |
| Sign-ups | 7 d / 30 d | 7 d / 30 d | 7 d / 30 d |
| Trial starts | 7 d / 30 d | 7 d / 30 d | 7 d / 30 d |
| Subscription starts | 7 d / 30 d | 7 d / 30 d | 7 d / 30 d |
| Revenue | 7 d / 30 d | 7 d / 30 d | 7 d / 30 d |
Tracking
Can the numbers be trusted?
| Check | Passes when |
|---|---|
| Install counts agree | The three install counts (platform, MMP, store) agree within 15% |
| Events fire correctly | The four events (App Open, Signup, Start Trial, Subscription Start) fire once per user, in order, on both iOS and Android |
| iOS postbacks | SKAdNetwork postbacks are arriving for iOS |
| Optimisation event | The optimisation event on each campaign is one you actually receive at readable volume |
Creative
Are the installs the right people?
| Read | Starting point we use, not a benchmark |
|---|---|
| Install-to-sign-up rate by creative and by placement | A sign-up rate that differs by more than 2x between creatives on the same audience is a creative problem, not an app problem |
| Cost per install against cost per sign-up | |
| Frequency and CTR trend on the top three creatives | Falling CTR with rising frequency is fatigue |
Onboarding and paywall
Do the right people become payers?
| Read | Starting point |
|---|---|
| Sign-up-to-trial rate | If sign-up rates are healthy across creatives and trial starts are low, the problem is in the app |
| Trial-to-paid rate | If trials start and do not convert, the problem is the paywall or the first week of value |
| Where in onboarding users drop, screen by screen | |
| Paywall placement, timing, trial length and price |
Decision table
| Symptom | Most likely broken | Check first |
|---|---|---|
| Dashboards disagree by more than 15% | Tracking | Step 1 |
| Cheap installs, almost no sign-ups | Creative | Step 2 |
| Sign-ups fine, trials low | Onboarding | Step 3 |
| Trials fine, paid low | Paywall | Step 3 |
| Everything low and spend is small | Sample size, not a funnel problem | Step 0, wait for readable volume |
| One platform much worse than the others | Tracking on that platform first, then creative | Step 1, then 2 |
The order matters. Fixing creative on top of broken tracking optimises to noise; fixing the paywall before the creative sends the wrong users to a better screen.
Get the PDF
Email only. No other questions. We do not add you to a newsletter.
Run it with us
A 30-minute scoping call with a senior engineer, designer or growth lead.
Book a scoping callBook a callQuestions founders ask us
Which numbers do I need before I start?
Spend, installs from the platform, the MMP and the stores, and four events: App Open, Signup, Start Trial, Subscription Start. If you cannot get all of them, that is the first finding.
Why is tracking checked first?
Because a creative or paywall change on top of broken tracking optimises to noise. You cannot read steps two and three until step one passes.
What if the dashboards disagree by more than 15%?
Stop and reconcile. Usually one platform is claiming installs another already counted, or an event fires twice on one OS.
How much spend do I need before the diagnostic means anything?
Enough for readable volume on the optimisation event: as a starting point, around 50 sign-ups a week per campaign on Meta and around 30 a month on Google before moving the event. Below that, the finding is "wait", not "fix".
Can Algoace run this on my account?
Yes. It is week one of every app growth engagement, and the result decides what we fix first.
Growth case studies
See all work →.png)
Ready when the work has to be right the first time.
A 30-minute scoping call with a senior engineer, designer or growth lead. You will leave with a written view of scope, risks and a realistic timeline, whether or not you hire us.
12 years40+ shipped products6 industries
Karachi and Sharjah