Correlation vs Causation
Correlation indicates that two variables move together (e.g., ad spend and sales). Causation indicates that one variable produces change in another. Attribution systems are correlation-based ('It happened after the click'). SpendSignal is causality-oriented ('It wouldn't have happened without the spend').
The Short Version
Just because the rooster crows before sunrise doesn't mean the rooster caused the sun to rise.
Visual Explanation
Attribution vs Incrementality
One measures correlation. The other measures causation. See the difference.
The Retargeting Fallacy
Retargeting ads are highly correlated with sales because they target people who are already shopping.
Attribution confuses this correlation with causation, giving credit to the ad for a sale that was likely going to happen anyway.
How it works
Attribution observes: Event A (Ad) -> Event B (Sale)
Incrementality asks: If No Event A, would Event B still occur?
The difference is Causation
Common Misconceptions
Assuming 100% causality (The biggest error in marketing)
Optimizing for correlation (leads to buying low-value, high-intent traffic)
Ignoring the counterfactual
Frequently Asked Questions
QWhy does it matter?
If you pay for correlation, you waste money. If you pay for causation, you grow the business. It is the difference between expense and investment.
QCan correlation ever verify causation?
It can suggest it, but it cannot prove it. You need experimental design or counterfactual modeling (like SpendSignal) to prove causation.