Who Really Earned the Sale?
Attribution is the most consequential analytics question almost nobody gets right, because it decides where your money goes. When a customer buys, multiple channels usually touched them along the way — a search ad, a blog post, an email, a remarketing banner. Attribution is how you decide which of those gets the credit. Get it wrong and you’ll defund the channels quietly creating demand while pouring money into the ones that merely show up at the finish line. This piece explains the models, why the default is dangerous, and how to think about credit honestly.
The problem: journeys, not single clicks
The root issue is that buying is rarely a single event. Someone discovers you through a blog post, comes back via a search ad weeks later, gets nudged by a remarketing banner, and finally converts after clicking an email. Five touchpoints, one sale. Attribution is the unavoidable, imperfect act of dividing one outcome among several causes — and how you divide it determines which channels look successful and therefore which ones get funded. There’s no perfectly objective answer, but some answers are far less wrong than others.
The trap is that the easiest answer is also one of the worst.
Why last-click is so dangerous
The default in most tools — and the instinct in most heads — is last-click attribution: whatever the customer clicked last gets all the credit. It’s appealing because it’s simple and unambiguous. It’s dangerous because it systematically overvalues the channels that close and undervalues the channels that create demand. The blog post that introduced you and the social content that built trust get zero credit; the branded search or remarketing click at the end gets everything.
Follow last-click faithfully and you’ll reach a destructive conclusion: cut the “underperforming” top-of-funnel channels and double down on the closers. Do that and the closers soon have nothing to close, because you defunded everything that filled the pipeline. Last-click doesn’t just misreport history; it actively steers you toward cutting the wrong things. This is the single most common way attribution leads businesses astray.
The range of models
Between last-click and a fuller picture sits a range of approaches, each a different theory of how credit should be shared.
- First-click credits the channel that started the journey — the mirror image of last-click, overvaluing discovery and ignoring everything that closed.
- Linear splits credit evenly across every touchpoint — simple and more balanced, though it pretends every touch mattered equally, which is rarely true.
- Time-decay gives more credit to touchpoints closer to the conversion — a reasonable compromise that still respects earlier influence.
- Data-driven uses your actual conversion patterns to estimate each touchpoint’s real contribution — the most sophisticated, and usually the closest to honest when you have enough data.
None is perfect. The point of knowing them is to stop accepting the default unthinkingly and to choose a model whose biases you understand.
How to think about it practically
You don’t need a perfect model; you need to stop being actively misled. A few practical principles get most of the way there. Move beyond last-click as your only lens — even glancing at first-click or linear alongside it reveals which channels last-click is hiding. Where you have the data and volume, data-driven attribution gives the most honest division. And crucially, treat attribution as directional, not exact: its job is to stop you from defunding the wrong channels, not to produce a precise ledger. The goal is better decisions about where money goes, and even an imperfect multi-touch view beats a confidently wrong single-touch one.
It also helps to hold attribution loosely against business reality. If your numbers say a channel contributes nothing yet sales fall whenever you pause it, trust the sales — attribution is a model, and the territory wins.
The bottom line
Attribution decides where your budget flows, which makes the lazy default — last-click — genuinely costly: it overcredits the channels that close, starves the ones that create demand, and steers you toward cutting exactly the wrong things. Understand the range of models, lean toward data-driven where you can, and treat the whole exercise as directional guidance rather than precise truth. Done that way, attribution stops you from optimising the wrong number and starts pointing your money at what actually earns the sale.
Want help crediting your channels honestly so you fund the right ones? Get in touch.
