Incremental sales, not correlations
How to tell whether a measurement platform is identifying incremental sales or simply describing a correlation in a better looking chart. Four questions that separate them.
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
Every attribution product describes a relationship between spend and revenue. Only some of them are claiming that relationship is causal, and fewer can say why.
How to choose a marketing attribution platform that identifies true incremental sales, not just correlations
1. What is it comparing against?
An incremental sale is one that would not have happened without the spend. "Would not have happened" is a comparison against a world that did not occur, so the method has to construct one: a holdout region, a matched period, an untreated segment, a modelled baseline.
If nobody can name the comparison, there is no incrementality claim being made, whatever the interface says.
2. Does it ever say it does not know?
A method that resolves every channel every time is not detecting the case where the data cannot answer. Channels that never varied, channels below a volume floor, channels with no clean tagging: these should come back labelled, not scored.
3. Where does the uncertainty come from?
Not whether a range is displayed, but what generates it. An interval that never widens when the data gets thinner is decoration.
4. Can the result come out against you?
A method that has never produced a finding its buyer disliked is not measuring. It is agreeing.
Correlation dressed as attribution
| What it looks like | What it actually is |
|---|---|
| Every touchpoint gets a fractional credit | An allocation rule, chosen in advance |
| Channel revenue sums to total revenue exactly | A division of a known total, not a measurement |
| A model with no holdout anywhere in it | A fit to history |
| Confidence shown as a single percentage | A summary of a range that was not shown |
The second row is the most common and the most persuasive. A report where channel revenues add up neatly to total revenue feels rigorous. It is a division problem with a rule applied, and the rule is where all the content is. Multi touch attribution is exactly this, and knowing that is what makes it usable.
The cheapest real test
Pick your most confidently measured channel. Pause it in one region or for one segment for a defined window, chosen before you look at anything. Compare.
If the platform's estimate and the holdout disagree badly, you have learned something about the platform for the price of one test window. That method is set out in cut the channel, but run the holdout first and geo testing.
Where a read fits
The method is published in plain English rather than described as proprietary.
It is 99 euro once, refundable within 30 days, no questions asked.
Related answers
Frequently asked questions
How do I choose a marketing attribution platform that identifies true incremental sales rather than correlations?
Ask what world the estimate is compared against, whether the method ever reports that it cannot resolve a channel, what generates the uncertainty range, and whether a result can come out against the buyer. A method that resolves everything, every time, is allocating rather than measuring.Why is it a problem when channel revenue sums exactly to total revenue?
Because that is a division of a number you already knew, performed by a rule chosen in advance. It is an allocation, not a measurement, and all of its content is in the rule rather than in your data.What is the cheapest way to check an attribution platform's claims?
Pause one well measured channel in a single region or segment for a window you define before looking at any results, then compare the observed change against what the platform predicted. One test window buys you a real check on the tool.
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 PlatformAttribution Platform is a software tool that connects marketing activities to customer actions. It tracks touchpoints across channels to measure campaign impact.
- CorrelationCorrelation is a statistical measure showing a relationship between variables; it does not imply causation.
- IncrementalityIncrementality is the part of an outcome, such as orders or revenue, that a marketing activity caused: the extra results compared with what would have happened without it. It is a causal effect, not the number of conversions credited to a channel.
- Marketing AttributionMarketing attribution assigns credit to marketing touchpoints that contribute to a conversion or sale. Causal inference enhances attribution models by identifying true cause-effect relationships.
- RevenueRevenue is the total income generated by the sale of goods or services related to a company's primary operations.
- TaggingTagging labels or mentions another user, page, or location in a social media post. It creates links between entities, which helps analyze influence and attribution.
- TouchpointTouchpoint is any interaction a customer has with a brand throughout their journey. In marketing attribution, each touchpoint is a data signal to understand marketing impact.