From visit to conversion

Return on ad spend is the revenue credited to an ad divided by what the ad cost. This page follows one visitor through three visits and a purchase and shows how Attribution arrives at that number: what each visit records, how a visit is tied to an ad, and how the credit for the purchase is shared out.

An organic visit

Imagine I am a potential Attribution customer browsing Google. I search for "attribution app" and click the first result, the organic listing for attributionapp.com, rather than an ad.

I land on the attributionapp.com homepage and the snippet on that page records a visit. What is on the page does not matter to Attribution; where I came from and the URL I arrived at do.

The visit records a referring domain of google.com and a destination URL of attributionapp.com/ with no advertising parameters. A visit from a search engine with no ad parameters is organic, so this one lands in the Organic channel under its Google Organic filter. Visits & Visitors explains what else a visit records, and Default channels lists the sources every project starts with.

A paid Google visit

Two days later I search again and this time click a Google ad. I land on the page with a gclid parameter in the URL.

Something in the landing URL always ties a paid visit to the ad that produced it; what that something is depends on the platform.

• Google Ads adds its own click ID, the gclid, to every ad click. Attribution reads it from the visit and matches it to the click through the Google Ads integration, which can take up to a day.
• Meta Ads has no such ID that Attribution can match, so Attribution adds its own fbaid parameter to your ad URLs through the Meta Ads integration and looks for it on the visit.
• Other platforms, Microsoft Advertising for example, are matched by the UTM parameters on the ad: utm_source, utm_medium, utm_campaign and utm_term, through a filter that looks for those values.

Either way the visit lands in the filter for that campaign, here Google Ads, and the same integration pulls the campaign's spend each day, so the filter knows what the click cost: $5 in this example.

A paid Meta visit and the purchase

Two days after that I click a Meta ad, a visit Attribution credits to the Meta Ads filter, with $10 of spend behind it. During this visit I submit the demo form and then buy. Those are two track() calls from the site, Demo Requested and Order Paid with revenue: 120, and the form submit also carries an identify() call with my email, so Attribution knows who I am. The install page shows how these calls are sent.

One visitor, three visits

All three visits belong to one visitor: the snippet recognised the same browser each time, and the identify call tied the purchase to a known person. Open the visitor from the dashboard and the sidebar shows the whole story, first touch, path and a timeline of the three visits with the events that happened during them.

This is why the question "which ad gets the credit for the $120" exists at all. The organic visit, the Google ad and the Meta ad all happened before the purchase, and Attribution has to decide how much each one earned.

How the credit is split

The attribution model is that decision. Under the Linear model every visit before a conversion earns an equal share, so the one conversion and its $120 are split three ways: $40 to Google Organic, $40 to Google Ads, $40 to Meta Ads. The conversion itself is shared the same way, which is why each row shows 0.33 conversions rather than 1: a third of one Order Paid event. The dashboard counts whichever event is selected as the Conversion Event in the control bar, Order Paid here; pick Demo Requested instead and the same three visits share one demo request and no revenue. The visitor sidebar shows the split for this one person.

On the dashboard the same $40 appears on each row, and now spend enters the picture. Google Ads earned $40 for $5, a return of 800%; Meta Ads earned $40 for $10, a return of 400%; the organic visit earned $40 and cost nothing. All Traffic sums it up: $15 spent, $120 earned.

The model is a setting, not a fact. Switch the view to First Touch and the whole $120 moves to the visit that started the journey, Google Organic, while both paid rows show $0 against their spend.

The five models and when each fits are on Attribution models, and Traffic options covers what to do with direct visits in a journey. Every number on the dashboard is the sum of stories like this one, one per visitor, split by the model you have chosen.

In short

Real journeys have more touches, more visitors and more conversions, and the same rules apply: each visit is recorded with where it came from, each paid visit is tied to its ad and its spend, and each conversion's revenue is shared out by the model. How it works explains each of these pieces in turn.

If you have any questions on this please feel free to contact [email protected].