Position-based attribution is a rule for sharing credit for a conversion among the marketing touchpoints that led to it. The usual version gives 40% of the credit to the first interaction, 40% to the last and divides the remaining 20% equally among everything in between. Because a chart of the credit looks like a U, it is also called U-shaped attribution. It is one of several attribution models.
How position-based attribution works
Take a customer who first finds a kitchen fitter through a Facebook ad, later reads a guide reached from organic search, opens a newsletter email and finally books after clicking a Google search ad. Under position-based attribution, Facebook receives 40%, Google Ads 40%, and organic search and email 10% each.
Short journeys behave differently. With one touchpoint, it takes all the credit. With two, most tools split it evenly, 50% each. The 40/20/40 split only appears with three or more touchpoints.
Which touchpoints count depends on the lookback window. A model that looks back thirty days will miss a first touch from two months earlier, so for considered purchases such as a new kitchen or a solicitor, set the window long enough to reach the real start of the journey, or the “first” touch will simply be the earliest one still in range.
The model sits between two extremes. First-click attribution gives everything to whatever started the journey, last-click gives everything to whatever finished it, and linear treats every step equally. Position-based assumes that introducing a customer and closing the sale are the most important jobs, while still acknowledging the middle.
Where it is available
Google removed position-based, first-click, linear and time-decay models from both GA4 and Google Ads in 2023. At the time of writing (October 2026), GA4 offers data-driven attribution and last-click variants only. Position-based models remain in some CRM and marketing automation platforms, and you can calculate one yourself from GA4’s BigQuery export or a spreadsheet of customer journeys.
Why it matters
Many UK businesses run one channel to find new customers and another to convert them: paid social or YouTube to introduce the brand, search and email to close. Judged on last click alone, the introducing channel looks like it loses money. A position-based view makes its contribution visible and can stop a business cutting the very spend that fills the pipeline.
Its weakness is that the 40/20/40 split is an assumption, not a measurement. The model does not know whether the first touch actually caused anything.
Common mistakes
- Treating the split as truth. It is a lens for comparing channels, not proof of what each one caused.
- Losing the first touch. If consent was declined, or the journey crossed devices, the real first touch may be invisible, and the model credits whatever it saw first.
- Counting direct visits as touchpoints. Direct sessions often mean returning customers or untracked links, and giving them 40% says more about tracking gaps than marketing.
- Switching models to flatter a channel. Choose the model before looking at the results, not after.
How to act on it
Use position-based attribution as one of several views. Compare it with last click and data-driven results; channels whose credit changes a lot between models are the ones worth testing properly. Where budgets allow, an incrementality test is the best check of whether a channel really drives sales.
Make sure the inputs are trustworthy first: consistent campaign tags, working conversion tracking and a reasonable lookback window. Choosing how to weigh channels when setting budgets is part of the performance marketing work I do, and the comparison with time-decay attribution is a useful next read.