I. The apparent condition

A decision is brought to leadership with supporting evidence: a performance report from one system, a cost figure from another, a customer signal from a third. Nothing in the presentation is missing outright. Each figure is accurate within its own system. The decision proceeds on the assumption that, taken together, this evidence describes the situation.

II. The structural problem

Financial systems, customer systems, transactional platforms, support environments, and analytics tools are each built to answer a specific, narrower question well — not to answer the broader commercial question leadership is actually facing. A financial system correctly answers what was spent and what was earned. A customer system correctly answers who engaged and how. A support system correctly answers what went wrong and how often. None of them is built to answer the composite question a consequential decision actually depends on, and none of them is incomplete in a way that would show up as an error inside its own environment.

The evidence, in other words, is not absent. It is distributed — held correctly, in isolation, across systems that were never designed to be read together. Assembling it into a single decision-relevant picture is not something any one of these systems does automatically; it is work that has to happen separately, and it does not always happen at all.

III. The distinction

Visible evidence is any information that exists somewhere in the organization’s digital systems and can, in principle, be retrieved. Decision-complete evidence is the specific combination of information, drawn from wherever it resides, that is sufficient to support the judgment actually being made. An organization can have an abundance of the first and a scarcity of the second. This is not a paradox — it is the predictable result of systems built around functional boundaries rather than around the shape of the decisions leadership eventually has to make.

IV. The evidence

A consistent pattern appears across organizations with genuinely fragmented systems: each system is optimized for its own primary purpose, and that specialization is often a deliberate and reasonable design choice rather than a failure of planning. The cost of this specialization is that information relevant to a cross-functional decision naturally accumulates in several places rather than one, and no single system is responsible for noticing when a decision requires more than what it alone contains.

A second pattern compounds the first: when evidence is distributed this way, what gets used in a decision tends to be whatever is easiest to retrieve, not whatever is most relevant to the decision. The system that is fastest to query, most familiar to the team preparing the recommendation, or most recently updated tends to dominate the picture presented to leadership — not because it holds the most decision-relevant evidence, but because it is the most accessible. Evidence that would materially change the picture, sitting one system away, may never enter the conversation at all.

A third pattern follows from the first two: organizations operating this way tend to respond to problems after they surface rather than anticipating them, because the fragmented evidence base makes it structurally difficult to see a developing situation until it has already crossed into one system in a way that triggers attention.

V. The executive consequence

A decision can be made in good faith, using accurate evidence, prepared by capable people, and still rest on a structurally partial account of the situation — not because anyone withheld information, but because no one system, and no one person assembling the case, was positioned to see all of it at once. This is distinct from a data quality problem. The data is not wrong. The decision is under-informed because the organization’s digital structure was never built around the shape of this particular judgment.

VI. The institutional implication

The correction is not integration for its own sake, nor a call to consolidate systems — a project of that scale is a separate question with its own costs and tradeoffs. The correction that belongs to this territory is narrower: before a consequential decision is finalized, leadership should be able to ask which systems hold evidence relevant to it, and confirm that the case actually in front of them draws from those systems rather than merely the ones that were easiest to reach. A decision built from accessible evidence is not the same thing as a decision built from complete evidence, and an organization that cannot tell the difference will consistently mistake the first for the second.