Correlation-Based Attribution
Correlation-based attribution assigns credit based on the sequential occurrence of events (Ad -> Sale). It assumes that because the marketing event happened *before* the sale, it *caused* the sale. This is the fundamental flaw of all touch-based tracking systems.
The Short Version
Superstitious analytics.
Visual Explanation
SpendSignal vs GA4
GA4 tracks events. SpendSignal measures incremental revenue. See the difference.
Prerequisites
Confusing Sequence with Consequence
A user decides to buy running shoes. They Google your brand name. They click your ad. They buy.
Correlation attribution gives the ad 100% credit. Reality says the user was already going to buy. The ad was just a toll-booth on their journey.
How it works
Log timestamps of ad interactions
Log timestamp of conversion
Draw a line between them and assign credit
Common Misconceptions
Thinking 'Data-Driven Attribution' solves this (It is still correlation-based)
Optimizing for 'Retargeting' because it has the highest correlation (and lowest causation)
Ignoring the counterfactual
Frequently Asked Questions
QIs correlation ever useful?
Yes, for understanding the 'Customer Journey' and touchpoint frequency, but NOT for calculating ROI.
QWhy is retargeting so correlated?
Because it targets people who already visited your site. High intent users are highly likely to click ads, creating a false signal of efficacy.