Institutions often respond to uncertainty in a predictable way: they ask for more information. When a decision feels weakly supported, another report is commissioned; when a problem remains unclear, additional indicators are collected; when performance is difficult to interpret, a new dashboard is added. The underlying assumption is reasonable enough. If understanding depends on information, then insufficient understanding may appear to indicate insufficient information. Yet this diagnosis can fail in an important way. An institution can possess enormous quantities of accurate information and still struggle to understand what that information means.
The distinction matters because an information deficit and a sense-making deficit are not the same problem. An information deficit exists when relevant observations are missing. A sense-making deficit exists when observations are available but the institution cannot adequately interpret their significance, relate them to one another or determine what they imply for action. Both can produce uncertainty, but they require very different remedies. Collecting more information helps primarily with the first. Applied indiscriminately to the second, it can make the cognitive problem worse.
Imagine an institution trying to understand why a public programme is underperforming. It already possesses service statistics, regional comparisons, budget data, user complaints, staff reports, demographic information and several evaluations conducted at different moments. None of these sources is obviously irrelevant. Yet they point in different directions. Some suggest that implementation quality is the central issue, others that eligibility criteria are poorly designed, others that population needs have changed, and still others that apparent underperformance may partly reflect the way success is measured. The institution does not lack information. It lacks a sufficiently coherent interpretation of the information it already has.
Adding another dataset may increase the number of things the institution has to reconcile without resolving the disagreement. New evidence may introduce another time period, another methodology, another administrative boundary or another representation of the problem. The informational environment becomes richer while the interpretive burden becomes heavier. What appears from outside as an evidence shortage may therefore be a bottleneck in the institution’s capacity to transform evidence into understanding.
This is one reason modern institutions can become informationally powerful without becoming cognitively proportionate to that power. Digital systems make it possible to collect, store and retrieve quantities of data that would have been unimaginable to earlier administrations. Sensors, administrative records, surveys, digital interactions and automated analytics continually increase observational capacity. But the ability to observe more does not automatically generate the ability to understand more. Information becomes institutionally useful only when someone — or some institutional process — can determine how different observations relate, which distinctions matter, what context changes their significance and what interpretation deserves confidence.
The paradox is that information growth can sometimes conceal this limitation. When institutions cannot explain a problem, asking for more data feels active and rational. It creates visible work, produces new artefacts and postpones the uncomfortable conclusion that the difficulty may lie not in what is known but in how what is known is being interpreted. The organisation can continue expanding its informational apparatus without confronting weaknesses in its interpretive architecture.
More information can also increase contradiction. Different datasets may have been designed for different administrative purposes, cover different populations or encode categories that are only superficially comparable. A financial system may describe a programme through expenditure categories, a service database through cases processed, a survey through user experience and an evaluation through estimated causal effects. Each source can be accurate within its own frame while offering a different representation of the same institutional reality. Accumulating them does not automatically produce a unified picture.
Nor does technical centralisation solve the problem by itself. Bringing data into one platform may improve access, but access is not understanding. A dashboard can place indicators next to one another without explaining their relationship. A data warehouse can integrate formats while leaving conceptual differences untouched. An institution may therefore create increasingly sophisticated information infrastructure while the central cognitive question remains unresolved: what does all this evidence, taken together, actually tell us?
The challenge becomes greater when information arrives at different temporal scales. Some indicators update daily, others annually. Some describe immediate activity, others long-term outcomes. A sudden change in one measure may look alarming when viewed over a week and insignificant when viewed over a decade. More information can therefore increase not only informational volume but interpretive ambiguity, because the institution must decide which temporal frame makes the observation meaningful.
A similar problem appears when information is produced by specialised parts of an institution. Legal teams, economists, operational managers, frontline staff, technical experts and policy analysts may each possess high-quality knowledge while interpreting the same situation through different professional frameworks. The result is not necessarily error. Each perspective may reveal something real. Yet increasing the volume of specialist information does not resolve the question of how those perspectives should be related. The institution can become better informed within each domain while remaining uncertain about the whole.
This is why the phrase “data-driven” can sometimes obscure more than it clarifies. Data can constrain speculation, reveal patterns and challenge assumptions, but it does not eliminate interpretation. Decisions about what to collect, how to classify it, which comparisons matter and what counts as significant already involve conceptual choices. Once information exists, further judgement is required to connect observations to explanations and explanations to action. Treating data as if it carried its own meaning risks mistaking informational abundance for institutional understanding.
None of this reduces the importance of collecting good information. Institutions genuinely do suffer from information deficits, and many public failures occur because relevant evidence was unavailable, ignored or never collected. The point is not that more information is harmful. It is that the value of additional information depends on the problem the institution is actually facing. When the bottleneck is observational, more evidence can transform understanding. When the bottleneck is interpretive, additional evidence may simply expand the material that must somehow be made sense of.
The distinction has practical consequences. Before commissioning another dataset, report or analytical system, an institution can ask what uncertainty the additional information is expected to resolve. Is a crucial fact genuinely missing? Is the institution unable to discriminate between competing explanations because existing evidence is insufficient? Or does it already possess substantial evidence but lack agreement about how that evidence should be interpreted? These questions shift attention from the quantity of information toward the cognitive function the information is supposed to perform.
They also reveal why measurement systems can become self-expanding. New indicators often generate new questions, which generate demands for further indicators. This can be productive when each layer increases explanatory power. But it can also produce an accumulation cycle in which the institution becomes ever more precise about fragments of reality while its overall understanding remains weak. The bottleneck is not the absence of observations but the absence of a framework capable of relating them.
This becomes especially important as artificial intelligence increases the speed and scale of information processing. AI systems can classify records, identify correlations, summarise documents and detect patterns across datasets that humans could not inspect manually. These capabilities can dramatically expand the informational field available to institutions. Yet expanded processing does not eliminate the question of significance. A system may reveal a pattern without establishing why the pattern matters, whether it reflects a causal mechanism or how it should change institutional judgement. Greater processing capacity can therefore intensify the difference between seeing more and understanding more.
The deeper lesson is that institutional cognition requires transformations between information and understanding. Information provides observations about the world; understanding requires those observations to acquire relationships, context and significance. When that transformation is weak, additional information can accumulate faster than the institution’s capacity to interpret it. The result is a distinctive cognitive overload: not simply too much data for people to read, but too much unresolved meaning for the institution to organise coherently.
More information can therefore make an institution understand less when accumulation increases interpretive burden faster than interpretive capacity. The problem is not informational abundance itself, but the assumption that every uncertainty is an information deficit. Sometimes an institution does not need another dataset, another dashboard or another report. It needs to understand what the information it already possesses means, how its different parts relate and why they point towards one interpretation rather than another. Institutional intelligence begins partly with recognising that knowing more and understanding more are not the same achievement.