Attribution decides which touchpoint gets credit for a conversion. Change the model and the same campaign can look like a winner or a loser — without a single thing changing about the traffic itself.
Almost no conversion comes from one click. A user discovers you through a native ad, comes back a week later from search, then converts on a third visit — and only one rule decides which of those touches earns the sale. That rule is your attribution model, and misunderstanding it is how operators cut the campaign that was actually feeding the whole funnel. This guide walks the common models, explains why affiliate networks lean on one of them, and shows how to read credit without being fooled by it. If the underlying plumbing is new, start with affiliate tracking explained.
Attribution is not about whether a conversion happened — your tracking already confirms that. It is about assigning credit when more than one touch was involved. Every model answers the same question differently: of the clicks that touched this user, which one (or which combination) deserves the sale? Because your optimisation decisions follow the credit, the model quietly shapes which sources look profitable, which look worthless, and therefore which ones you scale or kill.
Single-touch models hand the entire conversion to one click; multi-touch models spread it. The table shows who each one credits and where it earns its keep.
| Model | Who gets the credit | Best for |
|---|---|---|
| First click | The first touch that started the journey | Judging top-of-funnel discovery |
| Last click | The final touch before converting | Simple, closing-focused payout |
| Linear | Split evenly across all touches | A balanced view of the full path |
| Time-decay | More to the most recent touches | Short, recency-driven sales cycles |
| Position-based | Weighted to first and last, less to the middle | Valuing both discovery and close |
| Data-driven | Modelled from real conversion paths | High-volume accounts (GA4 default) |
Each carries a built-in bias. First click over-credits discovery and ignores everything that closed the deal; last click over-credits the closing touch and erases the sources that created the demand; linear is the most balanced but can flatter touches that barely mattered; and time-decay leans on recency, which suits fast sales cycles but punishes long consideration windows.
Most affiliate networks credit the last click before conversion, because it is simple, hard to dispute, and maps cleanly to a postback that fires on the final action. The consequence is real money: an upper-funnel source that started the journey — the native placement or content piece that first put the offer in front of a user — can receive zero credit under last-click, even though the sale would not exist without it. Know the model before you judge a channel's value, or you will cut the wrong campaigns and starve the top of your own funnel.
An assisted conversion is one where a source helped along the way but was not the final click. These are exactly the touches last-click makes invisible, and they are where discovery channels quietly earn their keep. Analytics tools surface them in an assisted-conversions or path report, and the gap between a source's last-click credit and its assisted credit tells you whether it is a closer or an opener. A channel that looks weak on last-click but heavy on assists is often carrying the campaign — cutting it is how a whole account slowly goes cold.
There is no single correct model; there is a correct model for the question you are asking. To judge what finally paid, last-click is fine — it is what most networks settle on anyway. To understand what created demand, look at first-click or assisted data. For a fuller picture across many paths, GA4 now defaults to data-driven attribution, which is why its numbers rarely match a network that credits last-click — the two are answering different questions on purpose. The discipline is to pick a model deliberately, hold it steady, and judge every source against the same rule rather than switching models to flatter a decision you have already made. How this connects to the wider measurement picture is covered in analytics for beginners.
Credit assignment eventually shows up as profit or loss, because you scale what looks like it pays. If your model over-credits closing touches, you will pour budget into retargeting and bottom-funnel placements while the discovery sources that fill the funnel wither — then wonder why the retargeting pool keeps shrinking. Read attribution alongside the actual return, not in isolation: a source with modest last-click credit but strong assisted value can be worth more than its headline number, which is the same logic behind judging campaigns on ROI, not vanity totals. Attribution numbers are only as trustworthy as the tracking underneath them, so a clean server-to-server postback setup is the precondition for reading any model at all.
Last click. The final touch before conversion gets the whole payout, because it is simple and maps directly to the postback that fires on the sale. It also means any source that only started a journey can receive nothing, which is why you should never judge an upper-funnel channel on last-click alone.
They use different models on purpose. GA4 defaults to data-driven attribution, which spreads credit across the path, while most networks credit last click. Neither is wrong — they answer different questions, so expect the totals to diverge rather than reconcile perfectly.
A conversion where a source contributed a touch along the way without being the final click. Assisted-conversion reports reveal the discovery and mid-funnel channels that last-click hides, so you can see which sources are opening journeys rather than just closing them.
No. Pick a model deliberately and hold it steady so every source is judged by the same rule. Switching models to justify a decision you have already made is a way to fool yourself — use different models to answer different questions, not to flatter a favourite campaign.
If you run campaigns on gut feel, you are flying blind with your own money.
Beginner · 6 min readHow affiliate tracking really works: clicks, click IDs, postbacks, server-to-server flows and why clean attribution is the foundation of a profitab...
Core · 10 min readWhat a tracking pixel is, how it fires in the browser, how it differs from server-to-server tracking, and where the Meta Pixel, the Conversions API...
