Data Informs, Humans Decide
Data tells you what happened. It does not tell you what to do next.
The phrase "data-driven" has done more damage to product management than any other buzzword. It implies that data makes decisions. It does not. Humans make decisions. Data provides evidence that those decisions should consider.
The distinction matters because "data-driven" creates a dangerous illusion of objectivity. When a team says "the data says we should do X," they are hiding a chain of human judgment calls: which data to collect, how to interpret it, what timeframe to analyze, which segments to focus on, and what threshold counts as significant. Each of those choices is subjective. The data does not "say" anything. The analyst says something, using data as evidence.
The healthier framing is "data-informed." Data is one input alongside customer conversations, market context, technical feasibility, strategic alignment, and product intuition. A PM who ignores data is reckless. A PM who hides behind data is avoiding accountability. The best PMs hold data and judgment in tension, using each to challenge the other.
There is also the problem of measurability bias. If you only build what you can measure, you will never build anything truly new. The most important product decisions often involve uncertainty that no amount of data can resolve. Should we enter a new market? Should we pivot our positioning? Should we bet on a technology that does not have traction yet? These are judgment calls that data can inform but cannot make.
Use data to identify problems, size opportunities, and evaluate outcomes. Use human judgment to decide what to do about what the data shows. And be honest about which one is driving the decision.
“A PM who ignores data is reckless. A PM who hides behind data is avoiding accountability.”
When this goes wrong
Using "the data says" to avoid making a judgment call. Only building features that can be A/B tested. Ignoring qualitative feedback because it is not "statistically significant." Confusing correlation with causation and building a roadmap around it.
In practice
- ✓Say "data-informed" instead of "data-driven" and mean it
- ✓For every data-backed decision, name the judgment calls the data did not make
- ✓Balance quantitative data with qualitative customer feedback
- ✓Be willing to make decisions with incomplete data when the cost of waiting is high