I once spent a weekend making TikToks for an app I'd built. Filmed six of them, edited them, added captions, posted them across two days. The following week, downloads went up by about 40%. Great, right? Except I'd also posted in a Reddit thread on Saturday, and someone had shared my App Store link in a Discord server I didn't even know about.
So which one worked? I genuinely had no idea. That's the problem attribution solves.
Attribution in one sentence
Attribution connects a purchase (or install) in your app back to the specific marketing activity that caused it. Not the channel in general. The specific link, post, or campaign.
When someone taps a link you shared on Instagram, downloads your app, and later buys the premium upgrade, attribution is the system that says: this purchase came from that Instagram post on March 3rd.
Without it, you're making decisions based on timing and gut feeling. Sometimes you'll guess right. Often you won't.
Why guessing breaks down fast
With one marketing channel, you can sort of tell what's working. If the only thing you're doing is posting on TikTok, and downloads go up when you post and drop when you don't, that's a reasonable signal.
But nobody stays on one channel forever. You start posting on Reddit. You try Apple Search Ads. You send your link to a few influencers. Maybe you run a newsletter. Now downloads are coming from five places at once, and you've lost any ability to guess which channel is pulling its weight.
This is where most indie developers stay stuck. They're spending real time (and sometimes real money) on marketing, but they're allocating effort based on vibes rather than data. The Reddit post that took ten minutes might be outperforming the TikTok that took three hours. You'd never know.
How it actually works
The mechanics are simpler than you'd think.
You create a tracked link for each campaign or source. Each link has a unique identifier. When someone taps it, the system records a click with a timestamp and the link ID.
Later, when that person opens your app and makes a purchase, the app's SDK fires an event. The attribution system matches the purchase back to the original click. Done. You now know that this user came from that link.
The matching is usually deterministic: the device that clicked is the same device that purchased. No fingerprinting, no probabilistic guessing, no privacy headaches. It works reliably and it works within Apple's privacy rules.
What you get from proper attribution
Real numbers change how you make decisions. Here's what I mean:
You find out that your Instagram posts drive lots of clicks but almost no purchases. Meanwhile, Reddit threads drive fewer clicks but higher conversion. So you shift your time toward Reddit and stop agonising over Instagram content.
Or you discover that one particular influencer drove 30 purchases while another drove zero. Now you know who to work with again and who to drop.
Or you realise that your Apple Search Ads are costing you £2 per install but only £0.40 of those installs ever convert to a purchase. The maths doesn't work, so you turn them off.
None of those decisions are possible without attribution. With it, they're obvious.
The enterprise problem
Traditional attribution tools were built for companies with big budgets. AppsFlyer, Adjust, Branch, Kochava. They're powerful, well-established, and designed for apps doing millions of installs.
They're also expensive. We're talking hundreds or thousands of pounds per month, with contracts, onboarding calls, and dashboards full of metrics that don't mean much when you're getting 50 downloads a week.
If you're a solo developer or a small team, these tools are overkill. You don't need multi-touch attribution models or cross-device graphs. You need to know: did this link lead to a sale, yes or no?
What smaller tools look like
A simpler attribution tool gives you three things:
Tracked links. You create a link for each campaign, influencer, or channel. Each one redirects to your App Store listing but records the click first.
Purchase matching. An SDK in your app (or a webhook from your payment provider) reports purchases. The system matches them to clicks.
A dashboard. You see which links drove clicks, installs, and purchases. That's it. No complex funnels, no probabilistic modelling, no month-long onboarding.
The pricing is different too. Instead of a flat monthly fee regardless of usage, some tools charge per attributed event. If you make five sales this month, you pay for five attributions. If you make five hundred, you pay for five hundred. Your attribution cost scales with your revenue instead of eating into it.
When attribution starts mattering
There's a stage before attribution matters, and it's worth being honest about it. If your app has ten downloads and zero marketing activity, setting up attribution is premature. You've got bigger problems.
Attribution becomes useful the moment you're actively marketing across more than one channel. That could be as simple as posting on TikTok and also having a link in your Instagram bio. Two channels, two possible sources. Already worth tracking.
It becomes essential when you're spending money. If you're paying for ads, paying influencers, or investing serious time into content creation, you need to know what's paying off. Otherwise you're just burning resources and hoping for the best.
Setting it up isn't hard
Most attribution tools take under an hour to integrate. You install an SDK (usually a few lines of code), create your first tracked links, and start sharing them. Purchases start showing up in your dashboard as they happen.
If you're using RevenueCat for billing, some attribution tools hook directly into it via webhooks. You don't even need to write purchase-tracking code in your app. RevenueCat tells the attribution system about every transaction, and it matches them to clicks automatically.
The whole process is less work than building a single feature in your app.
What to do with the data
Once attribution is running, check it weekly. Look for three things:
Which channels are driving actual purchases (not just clicks)? Double down on those.
Which channels are driving clicks but no purchases? Either fix the conversion problem or redirect your effort.
Which individual links or campaigns performed best? Try to figure out why. Was it the hook? The audience? The timing? Then replicate it.
That feedback loop is the real value of attribution. It's not a report you look at once. It's the thing that tells you whether your marketing decisions are working, every single week.
Getting started
If you want to try attribution without committing to an enterprise contract, LinkOwl charges 5p per attributed purchase with no monthly fee. Create a tracked link, share it, and see which of your marketing activities are actually driving sales.
You only pay when you make money. Which, if you think about it, is how it should work.