Sample reports

This is what
you get.

Five reports in the format I send to clients. The companies and numbers are invented. The structure, the checks and the reasoning are the same as in my real work. Every report opens with a one-page executive summary: the answer, what was done, and what to do next.

Cover: Ad revenue kept falling. Three weeks to fix it.First page: Ad revenue kept falling. Three weeks to fix it.
Investigation + implementation · 3 weeks

Ad revenue kept falling. Three weeks to fix it.

The full engagement in two reports. The initial report found three causes and argued against adding ad slots. The implementation report shows what was built, how it was measured against a holdout, and what it earned: +32% ad revenue per user.

Related work: Case 011, Case 016
Cover: Registration loses 75%. Which step should we redesign?First page: Registration loses 75%. Which step should we redesign?
Investigation · 4 days

Registration loses 75%. Which step should we redesign?

The funnel was rebuilt from raw events as real people. About 25% were really lost. The losses were at the entrances, existing users and shared links, not in the step the team planned to rebuild.

Related work: Case 001, Case 007
Cover: The test said +19%. After rollout, nothing changed.First page: The test said +19%. After rollout, nothing changed.
Experiment review · 3 days

The test said +19%. After rollout, nothing changed.

A winning onboarding test shipped to everyone, and paid subscriptions didn’t move. Nothing broke: a metric that counted trials, an outside traffic spike and a second test reading a different audience made it look like a win. Why rolling back wouldn’t help, and what to do instead.

Related work: Case 006, Case 010, Case 014
Cover: Registrations nearly doubled. Paying users didn’t.First page: Registrations nearly doubled. Paying users didn’t.
Investigation · 5 days

Registrations nearly doubled. Paying users didn’t.

Paying users rebuilt month by month. A referral promo that bought sign-ups, a lost App Store keyword, and a price increase that turned out not to be the problem.

Related work: Case 009
Cover: The paywall went blank. Do we need an emergency release?First page: The paywall went blank. Do we need an emergency release?
Urgent · 1 day

The paywall went blank. Do we need an emergency release?

Purchases fell by a third after an iOS release. Three separate bugs, found and contained on the server the same day. No emergency release, and the code fixes shipped in the regular release.

Related work: Case 016
Working with your data

Questions before you grant access.

What access do you need?

Read-only access to where the answer lives: analytics, your data warehouse, App Store Connect or Google Play, logs. Through your VPN if you prefer. Before we start, I send a short list with the exact roles, so your team can set it up quickly.

Why read-only access and not exports?

Investigations don’t follow a script. One finding points to the next table. With direct access I can follow it the same hour instead of waiting days for the next export.

How do you handle our data?

We sign a data processing agreement before I see anything. Data I export for analysis stays on my encrypted laptop, is used only for your project, and is deleted when the work ends. Reports go to you through your own channels.

Do you use AI?

No AI provider ever gets your users’ data. I use AI as a tool to check queries and research markets. Code is mine: I write and review every line, and AI assists only if you agree.

Let’s talk

Bring me
the problem.

An app that needs building. A migration that needs someone to own it. Numbers that don’t make sense. A product decision your team can’t resolve. Tell me what’s happening. I’ll tell you whether I’m the right person, how I’d approach it, and what the work would cost.

Email Alex ↗alex@x91tech.com
You bring the context. I take the work seriously.
Not sure what you’d get? See sample reports →