"The data supports it" gets said a lot more often than it's actually earned. Evidence-based decision-making isn't just pointing at a number — it's being able to explain why that number is the right one to look at, what it doesn't tell you, and what would change your mind.
What it actually requires
A genuinely evidence-based decision rests on three things: a question specific enough to be answered, data collected in a way that could actually answer it, and honesty about the parts the data leaves uncertain. Skip any one of those and you get a decision that looks evidence-based without being one.
Where it breaks down
In practice, the most common failure isn't bad data — it's picking the metric that already agrees with the preferred conclusion. Good research means being willing to look for the number that might say no.
How I apply it
In my work, this shows up as making sure a recommendation traces cleanly back to the analysis that produced it, and flagging explicitly where the evidence is thinner than the confidence of the conclusion.