Where Did That Install Come From? How Attribr Answers the Question
I got a message from an indie developer last month. She'd just launched her puzzle game on iOS, spent £300 on ads across three networks, and had no idea which one actually converted players. She was tracking everything in a spreadsheet. That conversation stuck with me because it's the same problem I heard over and over when we started building Attribr: developers knew they were getting installs, but they had no real visibility into the source.
The attribution problem nobody talks about until they need it
Attribution sounds like a technical afterthought, but it's actually the skeleton key to understanding whether your app marketing works. Without it, you're flying blind. You see download numbers go up, but you can't answer the questions that actually matter: which ad network sent you real players versus tire-kickers, and which of those players are still using your app on day 7, day 14, or day 30.
The existing tools for this (Branch, AppsFlyer, Adjust) are built for teams with enterprise budgets and full-time growth specialists. For indie developers and small studios, the pricing is a non-starter. We were paying £200 to £500 a month for SDKs we barely used, or we were flying blind. There wasn't a middle ground.
When we started Attribr, the mandate was simple: build an attribution SDK that answered three concrete questions, worked on a budget, and didn't require a computer science degree to integrate.
Deterministic matching plus a bit of clever probability
Install attribution works in two ways. Deterministic matching is the straightforward version: an ad network tells us 'User X clicked this ad at 14:32', and we see 'User X installed the app at 14:33'. That's a match. It's certain.
The second approach is probabilistic. iOS 14.5 made deterministic matching harder by limiting what ad networks could tell us without ATT permission. So we built a probabilistic layer that looks at device signals - timing, IP address, device type, language settings - and calculates the likelihood that an install came from a particular source. It's not a guess; it's weighted pattern recognition.
The honest truth is that probabilistic attribution alone gets messy. You need both. Attribr combines deterministic matching where the data exists and probabilistic matching where it doesn't, then weights the results based on confidence. The upshot: you get attribution that works across iOS 14.5+ without asking users for permission, and it fits in 50KB with zero external dependencies.
Why retention cohorts matter more than raw install counts
Here's where most attribution tools fall short for indie developers. They tell you where an install came from, then they stop. But a high-volume source that sends inactive users isn't a win; it's wasted ad spend. We decided Attribr should track cohort retention automatically. Every install is tagged with its source and its day-7, day-14, and day-30 retention. You can see exactly which ad network or campaign sent you players who actually came back.
I built this feature because I remembered the exact moment an indie developer told me 'My top-performing ad network looks great on dashboard, but nobody from that source opens the app twice.' He'd been optimising his spend on vanity metrics. Once we added retention tracking, the picture changed completely. His actual winner was a smaller network sending fewer installs, but those installs had a 35% day-7 retention rate.
The dashboard rolls this into charts. You see your cohort breakdown, funnel data, and retention curves broken down by source. It's the information a smart developer actually needs to make decisions.
The Rippl bridge: attribution that feeds into performance marketing
One of the unique things about Attribr is that it sits right next to the Rippl platform. Rippl is MRVL's performance-marketing network for community-driven CPI campaigns. If you're running promotions through Rippl, Attribr can attribute installs directly back to individual promoters. You know exactly which creator drove which user to install your app.
This isn't possible with off-the-shelf attribution tools because they're agnostic about your marketing channels. They don't care whether your installs came from Apple Search Ads, Facebook, a TikTok creator, or a friend-referral campaign. Attribr can distinguish between them all, and if you're using Rippl, it bridges the gap so you see the promoter's name next to the install and retention data. It closes the loop from creator to user engagement.
We built this because we kept hearing from developers who wanted to reward their best-performing community members. Without attribution, you couldn't see who was actually driving engaged users. Now you can.
Integration doesn't require an engineering degree
The install-source setup in Attribr takes three lines of code. Swift or Kotlin, whichever you use. You initialise the SDK with your app token, and Attribr starts matching installs to sources automatically. It runs in under 50ms on launch, so it doesn't bog down your app startup. There are no external dependencies to manage, no privacy nightmare of third-party libraries fishing through your device data.
That constraint (50KB, zero dependencies) forced us to be ruthless about what we included. We couldn't add every bell and whistle. We had to pick the core signal that mattered: where did this install come from, and is this user still here. Everything else in Attribr follows from those two questions.
For indie developers used to quick integrations, this matters. You can ship attribution in an afternoon. For studios that have been blocked by long integration cycles with enterprise tools, it's a relief.
Free tier doesn't mean limited data
We offer Attribr free for the first 1,000 installs per month. Not as a trial. Permanently free. The paid tiers kick in at 25,000 installs with Growth (£29 per month) and 100,000 installs with Pro (£99 per month). Pro adds fraud signals and the ability to roll up your own ad-network data if you're buying from multiple sources.
The reason we kept a real free tier is because we remember what it felt like to launch an app with zero budget for tools. You shouldn't have to pay to see where your first hundred users came from. The free tier has full install-source attribution and retention tracking. You get real data, not a neutered version.
That said, the free tier isn't designed to scale infinitely. If you hit 1,000 installs a month, you've probably got a app worth paying for tools to understand. Growth and Pro are priced for developers at that inflection point: you're beyond hobby project, but you're not managing ten million monthly users either.
Install-source attribution is one of those features that seems obvious after you've had it but invisible before. Once you know which of your sources actually drives engaged users, you stop wasting money. The question is: are you currently flying blind on your installs, or do you already have that clarity?