What multi-touch attribution never actually measured
Set coverage aside. Even working perfectly, a path model answers a different question from the one you are asking. The gap is older than the privacy problem.
Three platforms, three answers, one budget to defend. Your GA4 export already holds the answer, and the first finding is free and never leaves your browser. Get your first finding free
By Joris van Huët, Founder & CEOUpdated 3 min read
Run the numbers for your store: the free customer journey credit calculator.
Set the coverage problem aside entirely and assume a path model sees every touch. It still answers a different question from the one a budget owner asks. That gap predates the privacy problem by a decade and is the more interesting of the two.
Two different questions
| The question a path model answers | The question you are asking |
|---|---|
| Given that this order happened, how should credit be divided among the touches on its path? | If I had not run this channel, would this order still have happened? |
The first is an accounting exercise. It takes conversions as given and allocates them, which means the total is fixed in advance and channels compete for shares of a number that does not depend on whether they existed.
The second is a counterfactual question. It asks whether the total itself would have been different, which is what a budget decision needs, because budget changes whether the channel runs.
Why the difference is not academic
Consider a channel that only ever appears on paths belonging to people who were going to buy anyway. Under any credit-allocation rule it receives credit, because it was present. Under a counterfactual question it receives close to nothing, because removing it would not have changed the outcome.
Branded search and retargeting are the two textbook cases, and they are textbook cases because they are common. The mechanism is set out in the correlation versus causation problem.
The part that surprises people
Changing the credit rule does not fix this. First-touch, last-touch, linear, time-decay and data-driven are all allocation rules, and every one of them divides a fixed total among observed touches. Moving from last-click to a data-driven rule changes who gets the credit; it does not change the question being answered.
That is why teams who upgrade their attribution model repeatedly and remain dissatisfied are not doing it wrong. They are solving the allocation problem better and the causal problem not at all. The comparison is in causal inference versus rule-based attribution.
What answering the real question requires
A comparison against something that did not happen. Constructed by randomisation in a holdout test, by an accident that behaved like one, or by modelling variation that already exists. All three are harder than allocation, and all three answer the question you actually have.
The weakest of them, the observational read, is still answering the right question, which is why it can be more useful than a sophisticated allocation rule despite being less precise.
Where this leaves the tooling
Path models remain useful for describing behaviour and for diagnosing where journeys break. They are the wrong instrument for allocating budget, and they were the wrong instrument before consent rates started falling.
The one-sentence version
Allocation asks who to thank. Causal inference asks what to fund. Only one of those is a budget question.
Related answers
Frequently asked questions
What is the difference between attribution and incrementality?
Attribution allocates credit for conversions that happened among the touches on their paths. Incrementality asks whether those conversions would have happened anyway. The first takes the total as fixed; the second asks whether the total itself would change.Does switching to data-driven attribution fix the problem?
No. First-touch, last-touch, linear, time-decay and data-driven are all credit allocation rules dividing a fixed total among observed touches. Changing the rule changes who receives credit; it does not change the question the model answers.Are path models useless then?
No, they are useful for describing behaviour and diagnosing where journeys break. They are simply the wrong instrument for deciding budget, and they were the wrong instrument for that before coverage started eroding.
Go deeper: Causal attribution, explained.
Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.
Have an audience of ecommerce operators? Earn 20% recurring commission by recommending the €99 read.
Keep reading
Terms in this article
- AttributionAttribution identifies user actions that contribute to a desired outcome and assigns value to each. It reveals which marketing touchpoints drive conversions.
- Attribution ModelAn Attribution Model defines how credit for conversions is assigned to marketing touchpoints. It dictates how marketing channels receive credit for sales.
- Causal InferenceCausal inference is the set of methods for estimating how much an outcome changes because of a cause, rather than alongside it. It works by comparing what happened with an estimate of what would have happened without the cause.
- CorrelationCorrelation is a statistical measure showing a relationship between variables; it does not imply causation.
- CounterfactualA counterfactual is what would have happened under conditions that did not occur, such as a customer's outcome had they not received the campaign they did receive. It is never observed, so every effect estimate has to stand in for it with a control group, comparable units or a forecast.
- Holdout TestA holdout test is an experiment that keeps a random part of an audience from seeing a campaign and compares its results with those of the part that could see it. The held-out part is the control group, so the gap in outcomes estimates what the campaign added.
- Multi-Touch AttributionMulti-Touch Attribution assigns credit to multiple marketing touchpoints across the customer journey. It provides a complete view of channel impact on conversions.
- RetargetingRetargeting is online advertising that targets users who have previously interacted with your website or content. Attribution analysis shows the causal role of retargeting in driving conversions and improving ad spend.