Over the past decade, organizations have spent unprecedented sums on data. New warehouses. New platforms. New dashboards. New analytics teams. New chief data officers. Almost any executive you talk to will tell you their organization has been on some version of a "data-driven transformation" for at least the last several years.
And yet, if you ask the same executives a different question — has the way our organization actually makes decisions changed in any meaningful way? — the answer is almost always a long pause, followed by some version of "not really."
The gap between these two answers is the most important thing happening in enterprise data and AI right now. And almost nobody is talking about it directly.
The investment was real. The change wasn't.
It is not that the data investments didn't produce anything. They produced dashboards. They produced reports. They produced an enormous amount of data infrastructure, much of which works well. But the assumption underneath those investments — that better information would automatically produce better decisions — turned out to be wrong.
Better information does not automatically produce better decisions. Better information, in the hands of people and organizations that have not changed how they think, mostly produces more confident bad decisions and more sophisticated rationalizations of choices that were going to be made anyway.
Better information, in the hands of people and organizations that have not changed how they think, mostly produces more confident bad decisions.
This is uncomfortable to say, especially for those of us who have spent careers helping organizations build data capability. But it is what the evidence keeps showing. Survey after survey of enterprise data programs finds the same pattern: high investment, broad deployment, modest impact on the actual decisions that matter.
The bottleneck was never information. It was judgment.
The simplest way to put it is this: most organizations don't have a data problem. They have a judgment problem.
Judgment is what happens between information and decision. It is the cognitive work of asking what the data actually means, what is missing from it, whether it answers the question being asked, whether the question being asked is the right one, what assumptions are baked into the framing, what the costs of being wrong look like in each direction, and a dozen other things that no dashboard can do for you.
For decades, we have been investing in everything around judgment — the data, the tools, the platforms, the people who maintain them — and almost nothing in judgment itself. We treated it as something individuals just had, rather than as a capability that organizations have to deliberately build.
What this actually looks like in practice
Consider a typical scenario. A senior leader asks for "the numbers" on a strategic question. An analyst pulls together a dashboard. The dashboard gets shared in a meeting. The meeting moves through the dashboard in ten minutes and arrives at a decision that the senior leader, in retrospect, was already inclined to make.
The dashboard played a role in this story. But not the role its sponsors imagined. It did not actually shape the decision. It validated a pre-existing inclination, gave it the rhetorical authority of evidence, and let everyone go home feeling that the choice had been made on the basis of data.
This pattern is not anyone's fault, exactly. It is the natural result of building information systems without also building the cognitive habits, social norms, and decision practices that would allow information to actually count.
So what do you do?
You start by reframing what "data strategy" is for. Most data strategies are essentially infrastructure strategies in disguise — they describe what platforms will exist, what data will be collected, what governance will be applied. They almost never describe how decisions will be different as a result.
A real data strategy starts from the decisions. It asks: which decisions matter most to this organization, and what would it take to make each of them measurably better? The answer is almost never just "better data." It includes who is in the room, how the question is framed, how disagreement is surfaced, how outcomes are tracked, how learning gets fed back into the next decision.
That is decision architecture work. And it is the work that has to come before, alongside, and after the technology work — not optionally, but as the precondition for the technology work producing any return at all.
If you would like to read more on this, the related essays below extend the argument in two directions: into the human capabilities side, and into the organizational systems side. The first book, Turning Data Into Wisdom, works the cognitive habit side; the forthcoming Beyond the Tech works the organizational systems side.