Attribution vs Incrementality
Why attribution and incrementality answer different questions about marketing performance.
The core idea
Attribution is a rule system for assigning conversion credit to observed marketing touchpoints. The model is not the customer journey itself; it is a lens applied to the data that survived collection, identity matching, and event processing.
Why different systems disagree
Two systems can observe different touchpoints, use different lookback windows, resolve identities differently, or apply different credit rules. Even when both ingest the same purchase, they can legitimately assign that purchase to different channels.
Common approaches
Single-touch models put all credit on one interaction. Multi-touch models distribute credit. Modeled approaches may use statistical or machine-learning methods. Incrementality and media mix modeling ask different questions and should not be treated as mere variants of click attribution.
Choose a model for a decision
Use a model because it supports a specific question: acquisition source, last measurable influence, channel contribution, budget allocation, or experiment interpretation. Changing models without changing the decision can create the illusion of precision while simply moving credit around.
Practical rule
Keep event definitions and windows consistent when comparing performance over time. If the model changes, annotate the change. Otherwise a reporting improvement can look like a marketing improvement when nothing about customer behavior actually changed.
For teams that have outgrown native reporting, the next practical step is to compare independent tracking platforms against the measurement gaps you have actually documented—not against an abstract feature checklist.