Correlation and causality while building digital products are often used interchangeably to derive conclusions as the human brain likes to find patterns even if an event is a result of an independent event. Subsequently, it can lead to a waste of effort on low-value features/enhancements.
Causation is when event A causes outcome B. On the other hand, correlation is simply a relationship. One event relates to the other but the event doesn’t necessarily cause the other event to happen.
Was the traffic increase because of the new design (causality)? Or was traffic simply up organically at the time when the new design was released (correlation)?
Both can co-exist, but correlation does not imply causation. False positives can lead to a decision in completely opposite directions.
How can we tackle this?
Firstly, while looking for patterns around us, we should detach from the observation & unless causation can be clearly identified, it should be assumed that we’re only seeing the correlation. Chamath Palihapitiya, while working at FB in the growth team, focused solely on deriving results basis metrics rather than gut-based decisions. Objectivity is the key.
Secondly, try running experiments once we find a correlation, we can test for causation by controlling the other variables and measuring the difference using Hypothesis testing or A/B/n testing.
More adept we become at measuring & identifying causation & correlation better we’ll get at prioritizing our efforts.