Marketing Attribution: Deciding Who Gets Credit for a Conversion

Why it exists

A customer might see a social media ad, click a search ad a week later, and then buy from a promotional email. Attribution modeling exists to solve the problem of fairly distributing credit for that single sale across all those contributing marketing touchpoints. Without it, you can't accurately measure the ROI of each channel.

The mental model

Think of it like a soccer team scoring a goal. Does only the player who kicked the ball in get credit? Or do you also credit the midfielder who made the key pass and the defender who started the play? Attribution models are the different rulebooks for assigning that credit. "Last Touch" gives all credit to the scorer, while a "Linear" model gives everyone who touched the ball equal credit.

How it works

An attribution model applies a rule to distribute credit for a conversion (like a purchase or sign-up) among all the marketing interactions a user had within a set "lookback window," such as the last 30 days. Common models include: First Touch (100% credit to the first interaction), Last Touch (100% to the last), and Linear (equal credit to all). More complex models like U-Shaped give 40% to the first touch, 40% to the last, and distribute the remaining 20% among the middle interactions. Time Decay gives more credit to touchpoints closer to the conversion.

When to use it

Use attribution modeling to justify marketing budgets and optimize your channel mix. For example, use a First Touch model to understand which channels are best for customer acquisition and brand awareness. Use a Linear model for products with long consideration periods, like B2B software or subscriptions, where multiple touchpoints are necessary to nurture a lead. Use a U-Shaped model when you believe both the initial introduction and the final push are most critical.

When not to use it

Avoid relying on a single, simplistic model for all analysis. "Last Touch" attribution, while easy to calculate, is a major footgun. It ignores all the preceding interactions that built awareness and trust, often leading teams to incorrectly cut budgets for top-of-funnel marketing. Attribution is also less meaningful for single-interaction conversions; in that case, the single touchpoint gets 100% credit regardless of the model.

One canonical example

A retail company wants to understand which parts of its site are critical to conversion. They use a "Participation" model, which gives 100% credit to every unique page a user visited before making a purchase. The report shows that while the final purchase happened on the checkout page, the "product comparison" and "customer reviews" pages were visited by 80% of converting users. This highlights their importance in the journey, even though they weren't the "last touch."