The observation
Jay Galbraith’s 1974 paper for Interfaces, “Organization Design: An Information Processing View,” offers a foundational account of why this happens, and it predates the modern dashboard by several decades — which is itself informative, since it suggests the underlying mechanism is not a technology problem but a structural one that technology has simply made more visible. Galbraith, drawing on Herbert Simon’s work on the cognitive limits of decision-makers, argued that organisational form is shaped by task uncertainty: as the amount of information that must be processed to make a good decision rises, an organisation has to adopt some strategy for coping with the load — self-contained tasks that reduce how much coordination is needed, vertical information systems that route detail to the right level, or lateral relations that let information move sideways rather than only up and down a hierarchy.
The relevant implication for performance data specifically is that collecting more information does not, on its own, solve the information-processing problem Galbraith describes — it can worsen it, if the organisation’s capacity to interpret, prioritise and act on that information has not grown at anything close to the same rate as its capacity to collect it. A dashboard with twenty metrics is not twenty times more useful than a dashboard with one; past a certain point, each additional metric competes with the others for the same finite pool of attention, and the marginal value of each new data point tends to fall as the total volume rises, even though the cost of producing and maintaining it does not.
Why it happens
The mechanism compounds because collecting data has become progressively cheaper while interpreting it has not. Modern systems make it straightforward to instrument almost any activity and generate a metric from it, which removes what used to be a natural constraint on how much data an organisation could accumulate. Interpretation has no equivalent cost reduction: understanding what a metric actually means, checking whether it still reflects reality, and deciding what — if anything — to do differently because of it, still requires the same scarce, human, judgement-intensive attention it always has. The result is a widening gap between data volume, which has grown roughly in line with how cheap collection has become, and insight, which has grown roughly in line with how much attention leadership teams actually have available — a much slower-growing resource.
Galbraith’s framework also helps explain why this gap is not evenly distributed. Organisations that have deliberately built ways to route information to the right level — his “vertical information systems” — or to let relevant people connect laterally around a specific issue without waiting for it to travel up and back down a hierarchy tend to cope with rising data volume better than organisations that have simply added more reports to an unchanged reporting structure. The problem, in other words, is rarely solved by adding a coordination mechanism after the data has already piled up; it is better addressed by designing, in advance, some deliberate strategy for how much information actually needs to reach a given decision-maker to make a good decision, and building the reporting structure around that answer rather than around what is technically possible to collect.
A further contributing factor, distinct from Galbraith’s original framing but consistent with it, is that more data creates an appearance of rigour that is not always matched by substance. A report with more metrics can look more thorough than one with fewer, even where the additional metrics add noise rather than signal — which creates an incentive, often unspoken, to keep adding rather than to prune, since removing a metric can look like removing scrutiny even when it is actually removing clutter.
What it looks like in practice
Consider an organisation where a leadership team now receives a substantially larger volume of performance data than it did several years earlier — more dashboards, more automated reports, more granular breakdowns by team, region and product line. The volume increase is treated, implicitly, as a sign of progress: more visibility should mean better-informed decisions.
In practice, meetings built around this expanded reporting increasingly spend their limited time simply establishing which numbers to trust and which to set aside, rather than acting on what any of them show. Individual team members develop their own informal shortcuts for which reports actually matter to their decisions and which have become background noise they no longer really examine — a workaround that functions reasonably well for each individual, while leaving the organisation’s formal reporting structure carrying a growing amount of data that nobody is meaningfully using, at real ongoing cost to produce and maintain.
This pattern is a natural, near-inevitable consequence of data collection getting cheaper without any equivalent investment in deciding, deliberately, what actually needs to be collected and at what level of detail for a given decision. It rarely announces itself as a crisis; it accumulates quietly, report by report, each individual addition reasonable on its own terms, until the organisation is producing considerably more data than any group of people has the attention to genuinely absorb.
The same pattern appears across many kinds of organisation; see Other Sectors for how it is read elsewhere.
Questions worth considering
- Of the performance data currently reaching this leadership team, how much of it changed a decision in the last quarter, as opposed to simply being reviewed and set aside?
- Has any metric currently in regular circulation ever been formally retired, or does the reporting structure only ever grow?
- Is information in this organisation routed to the level and the person who actually needs it for a specific decision, or does most of it travel the same broad distribution regardless of relevance?
This perspective is part of Performance.
Your team is facing similar challenges?
Subscribe to The Riddle Ledger
Perspectives are sent as they are published, with no promotional email, and you can unsubscribe at any time.