Definition
Attribution answers the question: which touchpoint gets credit for a deal? A customer sees an ad, later reads a blog, searches the brand weeks on and buys. Which step 'caused' the sale? Attribution is the model that distributes those shares.
The simplest model is last-click: 100% of the conversion goes to the final contact before purchase. It's easy to measure and systematically wrong - it ignores everything that built attention and trust beforehand. First-click, linear and data-driven models distribute value across the journey more realistically.
With the demise of third-party cookies and journeys spread across devices, exact attribution gets harder. Modern teams therefore complement platform attribution with incrementality tests (hold-out groups) and marketing-mix models - measuring the real contribution rather than just the last click.
Why it matters
Trusting last-click blindly cuts exactly the channels that create demand and over-funds the ones that merely collect it. Wrong attribution leads straight to wrong budget decisions.
In practice
- 01Last-click gives brand search 100% - even though the Instagram ad two weeks earlier created the demand.
- 02A hold-out test switches off ads in one region and measures how many conversions still arrive - showing true incrementality.
- 03Linear attribution spreads value evenly across all touchpoints of the customer journey.
