Why we built install-source attribution for indie developers
Three months into running MRVL, we got a message from a solo iOS developer in Berlin. She'd just launched a puzzle game, watched 2,000 users download it in week one, and had no idea which of her five marketing channels had actually driven those installs. The attribution tools she'd looked at were either £500 a month or refused to talk to her because her install volume was too small. That conversation stayed with me.
The attribution blind spot
Here's what most indie developers do: they spend weeks polishing a game, launch across iOS and Android, push it to a handful of ad networks and app store optimisation channels, then watch the install graph spike. Then nothing. They have no real idea which channel is worth spending on next, which cohort of users is actually coming back, or whether their CPI is profitable at all.
The big attribution platforms make sense if you're moving millions of pounds a month in user acquisition budget. But if you're a small studio shipping your first title, or an indie dev splitting time between coding and marketing, paying £2,000 annually for attribution feels like buying a Ferrari to drive to the corner shop.
What gets lost is the signal. Without knowing your install sources and which users stick around past day 7, every marketing decision becomes a guess.
Why we couldn't just use what existed
We spent a month looking at what was available. The market had two options: expensive enterprise solutions built for publishers spending six figures on UA, or lightweight SDKs that gave you basic install tracking but buried the retention data behind paywalls. Neither solved the actual problem.
There was also the iOS 14.5 problem. Apple's changes to ATT (App Tracking Transparency) meant that probabilistic matching alone was increasingly unreliable. But the indie developers we spoke to couldn't afford to build their own deterministic network. They needed something that worked without special permissions, without guesswork, and without integration overhead.
The other thing that struck us: nobody had really connected install attribution to performance marketing in a way that made sense for communities and smaller budgets. We saw Rippl building this genuinely useful CPI marketplace where developers could tap into performance marketers directly, but there was no clean bridge between knowing an install came from Rippl and actually tracking that user's value over time.
Building something that actually fits
We started with constraints. The SDK had to be small enough that it wouldn't bloat an indie dev's app. We landed on 50KB. It needed to work without dependencies, because every third-party library is another point of failure and another thing to update. It needed to launch in under 50 milliseconds, which meant no phone-home delays or startup hangs. And the integration had to be genuinely simple - not marketing-speak simple, but actually three lines of code in Swift or Kotlin.
The retention cohort piece came from listening to what developers actually wanted to know. Not vanity metrics. Not total installs. Just: how many of my users came back on day 7, day 14, and day 30 from each channel? That single question, answered accurately, changes how you allocate your marketing budget.
And because we knew Rippl was launching a real platform for community-driven user acquisition, we built the Rippl bridge directly into Attribr. Now when a performance marketer brings you installs through Rippl, you can see exactly which marketer drove which install and track those users through your retention funnel. It's the only attribution SDK that does this.
What we left out on purpose
Attribr doesn't try to be everything. It's not a fraud detection service masquerading as attribution. It's not a fingerprinting engine that relies on probabilistic magic and calls it accuracy. It's not built to handle campaigns at the scale of 10 million monthly installs, because that's not who we're building for.
What it does do: it tells you with real accuracy where each install came from, which users came back, and whether those users have any fraud signals attached to them. For indie studios and small teams, that's the lever that matters. For enterprise teams running millions in spend, there are other tools.
We added fraud signals and ad-network roll-up reporting in the Pro tier, because developers at that scale (25,000 to 100,000 monthly installs) need to see patterns. But the core - deterministic plus probabilistic attribution, retention tracking, iOS 14.5+ compatibility without ATT - that's in the free tier. A thousand installs a month, completely free.
A year later
We've been shipping Attribr for just over a year now. The developer from Berlin ended up using it to figure out that her CPI from TikTok ads was roughly half the cost of her Google App Campaigns cohort, but her Google cohort retained 40 percent better at day 30. That one insight completely changed where she spent her marketing budget in month two. She still uses Attribr.
What's become clear is that indie developers don't need flashier features or more integrations. They need accuracy they can trust, retention data they can actually use, and a tool that doesn't get in the way. If your SDK adds latency to your launch, you've lost before you've started. If your dashboard takes 20 clicks to answer a simple question, nobody's going to use it.
The most rewarding feedback we've had isn't about growth or scale. It's from developers who say: for the first time, I actually know whether my marketing is working. And that changes everything.
If you're shipping an indie game or app and you've wondered where your installs actually came from, what's stopped you from finding out?