A rule-based or data-driven method for assigning credit for important actions to recorded marketing touchpoints.
An attribution model determines how credit for an important action is assigned to recorded marketing touchpoints. It may use a defined rule or a data-driven algorithm. The model changes how a report distributes credit; it does not change which transaction or inquiry actually happened.
A journey may include several interactions before the action. Different models and reporting scopes can emphasize different parts of that recorded journey. Check the model, eligible channels and relevant windows before comparing a channel’s attributed results across reports.
Attribution is limited by the available data and the model’s assumptions. It should not be presented as proof that a channel independently caused the action or that removing the channel would produce an exactly proportional loss. Unobserved interactions and measurement gaps remain important.
Takeaway: Use attribution to inform questions about the customer journey, not as a perfect allocation of responsibility. Explain the model when presenting channel performance.
Example
A customer first finds a business through search, later follows an email and eventually submits an inquiry. A model may credit those recorded touchpoints differently. The inquiry is still one business event, even if fractional credit appears across several channels.
What to check
Keep model settings and event definitions documented. Investigate differences in scope before accusing a report of being wrong. Combine attributed data with business evidence and appropriately designed evaluation rather than relying on a single model to decide every budget change.
For more detail, see Get started with attribution.