Data-Driven Decision Making

Data-driven decision making means turning the question into a metric and an acceptance criterion before choosing, then measuring the outcome of that choice afterwards.

The full loop has five links: a clear question, a metric that answers it, a threshold agreed in advance, the decision, and a second measurement. Dropping the third link is the most common failure, because a threshold set after seeing the data turns any number into support for the preferred answer. A second hard limit is that observational data does not establish causation; a correlation between a behaviour and retention can come from a third factor, and only random assignment removes that ambiguity. Being data driven also does not mean waiting for certainty, it means writing down explicitly what the choice was based on.