"We are a data-driven company." It is a phrase you hear in almost every corporate presentation, said with pride, like a badge of modernity. But behind it hides a confusion that costs dearly: most companies that call themselves "data-driven" do not quite understand what that means, nor whether it is even what they should want to be. There is a subtle but profound distinction between being data-driven — letting the data decide — and being data-informed — letting the data advise a decision that remains human. Confusing the two leads to errors in both directions: sometimes you obey the data blindly when you should not, other times you ignore it when it was exactly what you needed.
The difference is not one of vocabulary. It is a choice about the role data plays in decisions — and that choice changes everything, from the company's culture to the quality of the decisions it makes. Understanding when data should decide and when it should merely advise is one of the most important and least discussed strategic skills in modern management.
This article does not argue that one approach is always better than the other. It argues something more useful: that each type of decision calls for its own approach, and that maturity lies in knowing which to use at each moment, instead of applying a slogan to everything.
What being data-driven means
Being genuinely data-driven means that, in a given decision, the data has the final say. You define a metric, measure it, and the result determines the action without human judgment overriding it. It is the philosophy behind A/B testing taken to the limit: you show two versions, measure which converts more, and pick the winner, full stop — even if someone's intuition said otherwise. In these situations, letting the data decide removes bias, politics and ego from the equation, and that is precisely what makes it so powerful.
This approach shines when decisions are frequent, measurable and reversible, and when the goal is clear and single. Optimizing a button's color, an email's subject, the order of some results — small, testable decisions where human opinion adds nothing the numbers do not say better. Here, being data-driven is not coldness; it is intellectual honesty, accepting that the data knows more than our intuition.
What being data-informed means
Being data-informed is different and, for many decisions, wiser. Data comes in as an important voice — perhaps the most important — but not as the only nor the final one. It informs, contextualizes, alerts, but the decision also integrates what data does not capture: the strategic context, the market knowledge, the company's values, the long-term vision, the experience of those who have seen similar situations. The human decider listens to the data with attention and respect, and then decides with the whole picture in front of them.
This approach is the right one when decisions are big, rare, hard to measure or laden with consequences that data alone does not cover. Entering a new market, launching a radically new product, deciding the strategy of the coming years — decisions where historical data says a lot about the past but little about a future that does not yet exist. Here, blindly obeying the data would be driving while looking only in the rear-view mirror.
The dangers of confusing the two
- Being data-driven where you should be data-informed: obeying a metric in strategic decisions leads to optimizing the number at the expense of what it does not measure — like maximizing short-term profit while destroying the long-term customer relationship.
- Being data-informed where you should be data-driven: letting intuition override the data in small, testable decisions opens the door to bias and ego where the numbers would decide better.
- Using data as an alibi: pretending a decision already made on instinct was "data-driven", choosing only the numbers that confirm it.
- Paralyzing while waiting for perfect data: refusing to decide without total certainty, when many decisions have to be made with incomplete information.
The mistake of treating data as absolute truth
There is a particular danger in a misunderstood data-driven culture: treating data as if it were an objective and complete truth, when it is always a partial representation of reality. Data measures what is easy to measure and ignores what is not. A number can say a decision "worked" according to a metric, while it destroys value in ways that metric does not capture — long-term satisfaction, reputation, team morale. Trusting data blindly is trusting blindly what we chose to measure, forgetting all that was left out.
That is why human judgment remains irreplaceable in the decisions that matter. Not because intuition is better than data — often it is not — but because a good decider sees what the data does not show, questions whether the right metric is being measured, and integrates factors no number expresses. Data is a powerful instrument in the service of judgment, not a substitute for it.
A concrete case
A company prided itself on being radically data-driven: every decision had to be justified by a metric, and the team tirelessly optimized the numbers. For a while, it worked well — the short-term indicators rose. But one decision in particular revealed the fragility of the approach taken to the extreme. The data showed that a certain aggressive sales practice increased short-term conversion, and so it was adopted and reinforced, because "the data said it worked". What the conversion metric did not capture was that this practice annoyed customers and undermined trust in the brand — an effect that would only show up months later, in churn and reputation, far from the number they were optimizing. When the damage became visible, it was late. The company learned, the hard way, that being data-driven in a decision that required being data-informed — that required weighing what the short-term data did not show — had cost dearly. They reworked the culture: they kept data-driven discipline in small, testable decisions, but started treating strategic decisions as data-informed, listening to the data but deciding with the whole context. Decisions improved precisely because they stopped applying a single approach to everything.
Maturity lies in knowing how to choose
True data maturity lies not in always being data-driven, nor in never being. It lies in recognizing, for each decision, which approach it calls for. Frequent, measurable, low-risk decisions call for data-driven discipline, to get bias out of the way. Big, rare decisions laden with consequences the numbers do not capture call for data-informed judgment, so as not to drive looking only in the rear-view mirror. Knowing how to classify decisions and apply the right approach to each is what distinguishes a truly sophisticated organization from one that merely repeats the fashionable slogan.
Seen this way, the question "are we data-driven?" turns out to be ill-posed. The right question is: "do we use data appropriately to each type of decision?" — letting it decide where that makes sense, and letting it advise human judgment where that is what the decision requires. That is the attitude that makes data create value instead of becoming a straitjacket.
In practice
Next time you hear "the data says we should do X", pause and ask: is this a decision where the data should decide, or one where it should merely advise? Does the metric we are following capture everything that matters, or are there consequences it does not see? Asking these questions is not devaluing the data — it is using it with the wisdom to know when to obey and when to think. Does your company apply a data slogan to every decision, or does it know, decision by decision, when data should rule and when it should merely inform?
