Hey, Ivan here!
Today I'll cover:
why developers turn off profitable campaigns without knowing they're profitable
how to build analytics on par with $20M+ companies
how to increase revenue with simple changes
Let's dive in.
Quick wins from the Adapty report.
Adapty released their subscription apps report. Key takeaways:
Weekly > Yearly by LTV. Weekly subscriptions generate more revenue per user.
And weekly paywalls with trials outperform every other setup.
Higher prices = more money. High-price apps earn 3× more. Conversion doesn't compensate for lower prices.
What to do: raise prices, add weekly subscriptions, and test trials.
Lots of useful data in there – check it out.
In partnership with Adapty
How to build corporate-level analytics
I built analytics for my app in just one week – and it's a game changer. Now I know exactly:
how trial conversions flow into paid subscribers
how long each cohort takes to pay back
what ROAS and CAC I need on days 3, 7, and 30 to hit my payback targets
how many days until a campaign breaks even
which countries are top performers
how payback KPIs differ by country
At my previous job, building this took 6-12 months, 2 analysts, and 1 data engineer.
Today I did it in 1 week by myself with Claude Code.
Step 1. Collect subscription data
Set up webhooks from RevenueCat / Adapty to your server. Store: user_id, event_name, revenue, date.
Also export raw event data for your entire history – you'll need it for cohort analysis.
What goes into the database: user_id, event type (trial started, subscription started, renewal, cancellation, refund), product_id, revenue, date.
Ask Claude Code to set up the DB and load everything.
Step 2. Connect Apple Ads API
If you're running Apple Search Ads, connect directly via API. If you have other traffic sources and an MMP – connect to the MMP.
You need: spend, impressions, taps, installs – broken down by campaigns, ad groups, and keywords. By day.
Step 3. Link ads to subscriptions
Adapty / Superwal / RevenueCat has attribution in webhooks – campaign_id, adgroup_id, keyword_id from Apple Search Ads.
Important: store attribution separately from events – one record per user.
Step 4. Build ROAS curves and target KPIs
This is the key step – without it you won't know if a campaign is profitable or not.
4.1. Build historical ROAS curves
Ask Claude Code to build a payback chart for monthly cohorts based on historical data – both paid and blended (with organic). The goal is to see how fast old cohorts paid back.

4.2. Calculate target KPIs
Ask Claude Code to generate a KPI curve from historical cohorts based on your target payback period.
Say you need payback in 180 days (ROAS 1.0x by d180). From my historical curves, ROAS grows like this:
d7 = 0.28x
d30 = 0.48x
d60 = 0.67x
d180 = 1.0x
So if a campaign shows ROAS 0.28x on d7 – it's on track. If it's 0.15x – it's a candidate for shutdown.

4.3. Compare current cohorts to KPIs
For each fresh cohort (and campaign) calculate ROAS d7/d30 and compare to target. Green – above KPI, red – below.
4.4. Predict for fresh cohorts
For cohorts younger than 60 days, predict final ROAS based on historical curves. This lets you avoid waiting 6 months to find out whether a campaign will pay back.

Step 5. Build a dashboard
Ask Claude Code to build a React dashboard. Mine shows:
Overview – KPI cards (revenue, spend, CAC, ROAS for current month), trend charts
Marketing – ROAS by campaign, top countries
Cohorts – LTV curves by month, retention matrix
ROAS Evolution – how each cohort's ROAS grows from d7 to d180
What this gives you in practice
Many developers turn off ad campaigns, thinking they're unprofitable, when in reality they're not.
This analytics helps you:
build payback KPIs
know within a week if a cohort will pay back or not
stop turning off profitable campaigns
scale winning campaigns (even if they don't pay back on day 1)
If I'd had this analytics 2 years ago, my business would be 10× bigger now.
Reply to this email if you need more details – I'll do a deep dive.
Worth checking out
🍎 Apple Ads Insights – Apple launched new dashboards and visualizations. Looks nice, but without revenue data – you're only seeing half the picture.
💰 From failure to $22k/mo – Max struggled with one failed product for 5 years. Now he has 30 apps and $22k/mo.
Before you go:
I need more hands-on experience with different apps at different scales, so I started doing audits and consulting (not cheap).
Questions? Reply to this email.
See you next week,
Ivan.

