Understanding attribution
An introduction guide to UTM vs in-platform attribution and how to interpret these numbers
What is UTM-based attribution
UTM-based conversions are conversions based on UTM-tracking parameters that you add to links to your website. As an example, you might tag the link to your website from one of your Meta ads with a utm_source=facebook UTM hyperparameter.
This hyperparameter makes sure that when a tracking pixel such as Converge on your website starts tracking a user in a session, it knows that this session came from Meta.
Because users can have multiple sessions before they purchase, the same purchase can be attributed to different channels based on the attribution model that you pick. In the flowchart below, based on a first-touch model, the purchase should be attributed to Meta. Based on a last-touch model, the purchase should be attributed to Google.
What is in-platform attribution
In-platform conversions are conversions as calculated by the attribution model of the ad platform itself. As an example, when you mark in Meta a “7-day click, 1-day view” model and look at the Facebook Ads Manager, you see all conversions that Facebook believes came from a click on the ad in the last 7 days or a view of the ad in the last day.
This is typically the model that the ad platform will use itself to optimize its ad targeting.
Which attribution method should I choose to make decisions?
You should typically include both numbers in your decision-making, as they both have their own strengths and weaknesses. You can find a summary of both methods below.
| Attribution Model | Strengths | Weaknesses |
|---|---|---|
| UTM-based method | Great for relative comparisons: UTM metrics allow you to compare all channels based on the same attribution model | May be less accurate: UTM-based attribution methods cannot take into account view-through conversions or cannot stitch sessions across devices |
| In-platform based method | Better accuracy: in-platform metrics are usually more accurate because ad platforms have more data to attribute (they know which users viewed ads). | Hard to compare across traffic sources: Every ad platform employs different attribution models, making it infeasible to compare conversions. Some channels (e.g. Organic) do not have in-platform attribution. |
As a general rule of thumb, UTM-based methods are better for relative comparisons across traffic sources and in-platform-based metrics are better for checking performance within a traffic source (e.g. comparing the number of different Meta campaigns against each other).
Known limitations of UTM-based attribution in relative comparisons
Although the UTM-based attribution method is better for relative comparisons, it comes with some known limitations.
UTM-based attribution typically underreports top-of-funnel channels (e.g. Meta and TikTok) Because UTM-based tracking does not work across devices and does not have any information on view-through conversions, it tends to get worse at reporting things top-of-funnel (in sessions early in the customer funnel). The longer an average customer takes to convert the more underreported these channels will be.
UTM-based attribution typically overreports conversions to search and brand traffic Because UTM-based tracking might miss earlier sessions until conversion, it will tend to overreport channels that are later in the customer journey such as brand or search traffic.