Marketing Attribution: Understand the Limits of Credit

Define the conversion and reporting window

Agree on what counts as a conversion, which events are included, how duplicates are handled and how long after a touchpoint credit may be assigned. Document the rules before comparing channels.

Keep campaign data consistent

Use a shared naming convention for source, medium, campaign and creative. Audit redirects, cross-domain journeys, consent effects and missing tags that can make the same campaign appear under different labels.

Choose a model that matches the question

Last-click is easy to interpret but can over-credit the final touchpoint. Multi-touch models distribute credit using assumptions that should be made explicit. Neither should be treated as a causal estimate by default.

Compare attribution with other evidence

Review conversion paths alongside customer research, channel costs, sales feedback and time-lag patterns. For larger decisions, consider holdout tests, geo experiments or other credible incrementality methods.

Report uncertainty honestly

Show the model and time window, note data gaps and avoid adding attributed revenue across platforms when the same conversion may be counted more than once.

Common mistakes to avoid

• Choosing tools before clarifying the customer or business problem.
• Optimising a surface metric without checking downstream quality.
• Launching a process without an owner, baseline or review date.