What is marketing attribution and why do platforms all claim the same conversion?

Attribution is the process of deciding which marketing touchpoints deserve credit for a conversion. It sounds administrative and it is one of the hardest problems in the discipline.

Why the numbers never add up. A customer might see a Facebook ad, search your brand on Google, click a Google ad, read a blog post, receive an email, and then buy. Each platform reports using its own tracking, its own rules and its own self-interest. Facebook counts it. Google counts it. Your email tool counts it. Add up the platform-reported conversions and you routinely get more conversions than you actually had — sometimes far more.

This is not usually fraud. Each platform genuinely observed a touchpoint and genuinely applied its model. They are all describing the same event from different vantage points.

The common attribution models:

Last click gives all credit to the final touchpoint. Simple, and systematically overvalues bottom-of-funnel channels — especially brand search, which often just captures demand another channel created.

First click gives all credit to the first, overvaluing awareness channels.

Linear splits credit evenly; time decay weights recent touches more; position-based weights the first and last most heavily.

Data-driven models use the platform's own observed data to assign fractional credit. Better in principle, opaque in practice.

Attribution windows matter enormously. A platform crediting a conversion up to 28 days after a view with no click will claim vastly more than one counting 7-day click-through only. Comparing two platforms on different windows is meaningless.

What practitioners increasingly do instead: treat platform numbers as directional, watch blended metrics — total spend against total revenue — and run incrementality tests or geographic holdouts, which measure what would have happened anyway. That last question is the one attribution cannot answer.