Marketing attribution is the science of crediting the right channels for conversions. It sounds straightforward. In practice, it’s one of the most contested and poorly-implemented areas of marketing analytics. Most teams are using a model they know is wrong (last-click) because the alternatives feel too complex. Here’s a more pragmatic path.

Why Last-Click Is Broken

Last-click attribution assigns 100% of conversion credit to the final touchpoint before purchase. This systematically over-credits bottom-funnel channels (branded search, direct, retargeting) and under-credits awareness and consideration channels (content, social, display, video). The result: budgets get shifted toward channels that close deals, and away from channels that create the conditions for closing deals. This feels rational in the short term and becomes a growth limiter over time.

The Attribution Models Worth Understanding

First-click: Credits the first touchpoint. Overcredits top-of-funnel. Useful as a complement to last-click to see what’s initiating journeys.

Linear: Distributes credit equally across all touchpoints. Simple, fair, but doesn’t reflect the reality that some touchpoints matter more than others.

Time decay: Weights touchpoints more heavily as they get closer to conversion. Reasonable for short-cycle purchases.

Data-driven: Uses ML to assign credit based on the actual conversion paths in your data. The most accurate but requires sufficient conversion volume (typically 1,000+ conversions per month per channel to be reliable).

A Pragmatic Three-Step Approach

Step 1: Run both first-click and last-click reports side by side. Channels that appear high in both are clearly important. Channels that appear in first-click but not last-click (content, social, display) are assisting conversions but not getting credit. This simple comparison surfaces most of the hidden contribution.

Step 2: Add incrementality testing for major channels. Turn off a channel entirely (or reduce spend significantly) for a defined test period and measure the conversion impact. This is the most direct measure of a channel’s true contribution. Difficult to run frequently, but worth doing annually for your largest channels.

Step 3: Use media mix modelling for top-down validation. MMM uses statistical regression to attribute revenue to marketing inputs based on historical data, without needing individual-level tracking. Useful for validating channel mix decisions at a portfolio level, especially as cookie-based tracking degrades.

The Metric to Report to Leadership

Revenue per dollar of marketing spend, by channel, measured over a 90-day window with first-touch attribution blended with last-click. Imperfect, but much more honest than pure last-click, and explainable to a non-technical audience. Accompany it with the caveat that this model still undervalues brand and content investments whose effects compound over time.