Attribution model

What it means

An attribution model is the rule by which credit for a conversion, such as a purchase or an enquiry, is split among the marketing channels through which the customer reached the site. Few people buy on their first visit: they often click an ad first, then read an article found through search and finally arrive through an email.

The simplest is the last-click model, which gives the whole conversion to the final channel. The first-touch model credits the channel that brought the customer in first. The linear model splits credit evenly, and the data-driven model distributes it according to the statistics of many customers’ behaviour.

No model is the truth; each simplifies the customer journey. Their value is that by comparing models a company sees which channels first introduce people to it and which merely close the deal, and can set its budget accordingly.

How it works in practice

The model is set in web analytics and in ad platforms, which often each use a different one. That is why figures from different tools never match and why the sum of conversions reported by the platforms usually exceeds the actual number of orders in the online store.

For attribution to work, measurement must be clean: consistent UTM parameters on campaigns, correctly configured conversions and cookie consent handled by the rules. Where cookies are limited, server-side tracking helps capture more data.

For financial decisions we recommend comparing attribution with hard figures from the accounts: revenue, margin and the number of new customers in the period. If an ad platform reports an excellent ROAS but revenue is flat, the channel is probably claiming customers who would have come anyway.

What to watch out for

Last-click attribution systematically undervalues channels that build awareness, such as content, social media or display, and overvalues branded search and remarketing. A company cutting budgets on that basis can gradually switch off exactly what brought customers in at the start of their journey.

The second risk is trusting a single platform’s numbers. Every ad platform has an incentive to report as many conversions as possible. An independent view comes from your own analytics, comparison with booked revenue and occasionally a test in which a channel is paused to see what happens.