Attribution in Marketing: Definition & Models

What does attribution mean?

Attribution—or, in German, “Zuschreibung” or “Zuordnung”—answers the question of which touchpoints with the brand are credited with a share of the conversion. This attribution is determined by an attribution model, which distributes credit according to fixed rules:

In the “last-click” model, the final touchpoint before the conversion receives all the credit.

In the “first-click” model, it’s the first touchpoint.

In the linear model, all touchpoints receive equal shares of the credit.

Data-driven models calculate the shares based on observed data and user behavior.

Depending on the model, the channels that are deemed successful—and thus where the budget is allocated—change.

Where does the attribution model originate?

The principle stems from traditional advertising effectiveness research, which has always sought to determine the contribution of different advertising touchpoints to a purchase decision. In online marketing, it became technically possible to track visitor paths and measure attribution starting in the 2000s.

For a long time, the last-click model was preferred because it was the easiest to implement technically. With the emergence of large analytics and advertising platforms, the data-driven model then came to the forefront.

Why is attribution important?

Attribution is a tool for budget management. If you count only the last click, you neglect all the previous efforts that built trust: expert articles, social media presence, and recommendations.

As a result, you end up overestimating the channels immediately preceding the conversion. Attribution seeks to distribute success more transparently across the relevant touchpoints, leading to better investment decisions.

Advantages

Attribution provides a more realistic picture of channel effectiveness than any individual analysis.
It gives measures with long-term effects the support they would otherwise lack in budget discussions.

Challenges

Due to consent requirements, browser restrictions, and device switching, the data set has become fragmented. Touchpoints outside of digital channels do not appear in the data at all.

In general, there is no single objectively “correct” model. It is important to make a deliberate, documented decision in favor of a specific model; otherwise, each department will compare its numbers based on different criteria.

Example

A purchasing manager has been reading a vendor’s white papers for months. She then clicks on a LinkedIn ad, searches for the company name on Google, clicks on the search result, and submits an inquiry.

In the last-click model, the web search receives all the credit. As a result, the company decides to cut the budget for industry articles and LinkedIn—a major mistake that, over time, leads to fewer web searches.