Intempt Docs
GuidesAnalytics

Using attribution models

Which touch gets the credit when a deal closes, and why the answer changes the number. Set the attribution model on an Insights report and compare two.

Covered in the Academy. Intempt Academy covers seeing which channels really drive revenue step by step.

Overview

A deal closes after several touchpoints, an ad click, an email open, a demo call. Which one gets the credit? There's no single correct answer, different attribution models answer different questions. Attribution lets you pick a model and apply it to an Insights report's breakdown, crediting a conversion to a specific touchpoint instead of just counting it as a raw event.

Use cases

  • Seeing what brings people in. Use First touch.
  • Seeing what closes the deal. Use Last touch, the default.
  • Splitting credit across the whole journey. Use Linear, U-Shaped, J-Shaped, Inverse J-Shaped, or a Custom split instead of giving one touchpoint all the credit.
  • Seeing every touchpoint that was present, not just who gets credit. Use Participation to count every touchpoint in full, instead of dividing credit between them.
  • Reconciling two different numbers. When two reports on the same conversion disagree, check whether they're using the same model before assuming one is wrong.

How to use it

Open an Insights report and open Breakdown. Click Apply attribution, then set:

  • Model. Each model splits credit differently across a touchpoint's position in the journey, first, middle, or last:
ModelFirst / Middle / Last
Last touch (default)0% / 0% / 100%
First touch100% / 0% / 0%
Linear33% / 33% / 33%
U-Shaped40% / 20% / 40%
J-Shaped20% / 20% / 60%
Inverse J-Shaped60% / 20% / 20%
Participation100% / 100% / 100%, every touchpoint counted in full rather than split, so totals across touchpoints don't add up to 100%
CustomSet the first/middle/last split yourself

attribution model dropdown percentage splits - attribution - Intempt documentation

  • Lookback window. How many days back a touchpoint can be and still receive credit.

lookback window 30 days highlighted - attribution - Intempt documentation

Reading the results

Running the same data through First touch and Last touch produces two different winners, and neither is wrong, they answer different questions: what brought people in, versus what got them over the line.

💡 Good to know

The real failure mode isn't picking the wrong model, it's not saying which one you used. If you're reporting an attributed number, note the model alongside it so it can be reconciled against anyone else's number.

Where to go next

On this page