Marketing attribution assigns credit for an outcome to interactions in a measured journey. A model determines how that credit is allocated.
It helps organise reporting, but it is not automatically proof of causation. A channel receiving credit does not establish how many sales would have disappeared without that channel.
Why are attribution models needed?
People can encounter a business through several routes before acting. A search, an article, an email and a later direct visit may all appear in the observed history.
A reporting system needs a method for allocating credit. Different methods can answer different questions and produce different channel results.
Last touch versus wider models
A last-touch approach concentrates credit on a qualifying interaction near the outcome. Other approaches consider multiple observed interactions.
Neither label provides a complete account without the platform's rules. Windows, eligibility and reporting scope affect which interactions are considered.
What is an attribution window?
It defines a period within which interactions may be eligible for credit under the system's rules. A short window can produce a different view from a longer one.
Choose and interpret the period in relation to the buying cycle. A complex service may involve a longer process than a familiar repeat purchase.
Why do platforms disagree?
Platforms observe different signals and use different methods. Several systems can claim credit for the same underlying order.
Do not simply add their attributed revenue together. Reconcile claims with actual business totals and explain where definitions overlap.
Attribution versus incrementality
Attribution allocates observed credit. Incrementality concerns the additional outcome caused by activity compared with what would otherwise have happened.
Answering the second question generally needs a suitable evaluation design. Routine attribution reports do not independently establish the counterfactual.
What should a business report?
State the model, window and value definition. Include outcome quality and known measurement limitations.
Use the result to support decisions rather than to settle every budget debate with a single dashboard. Context from experiments, customer conversations and operational records can add evidence.
How should changes be handled?
Record changes to models, event definitions and tracking. A new reporting configuration can redistribute credit without changing total sales.
When comparing periods, distinguish business movement from measurement movement. That makes conclusions easier to assess and act on.
Further reading
For reference, see the supporting documentation. Explore related explanations of Return on ad spend, Google Analytics 4, Customer journey.
A practical example
Illustrative example: an order follows a social advert and a later branded search. Both platforms report a contribution under their own rules. The business compares those claims with the actual order value and avoids counting the same sale twice in its overall revenue.
Why it matters
Attribution provides a reporting framework for complex journeys. Understanding its assumptions prevents allocated credit from being mistaken for demonstrated incremental value.
