Signal Extraction in an Age of Information Obesity

What is information obesity and why does it degrade decision quality?

Information obesity is the organizational condition where the volume of available data exceeds the cognitive capacity to process it meaningfully, producing decision paralysis, analysis addiction, and a paradoxical decrease in actionable insight as data volume increases.

I built a reporting dashboard that tracked 14 operational metrics. Leadership consumed the dashboard daily. Decisions improved for approximately 6 weeks. Then a request came to add 8 more metrics. Then 5 more. Within 4 months, the dashboard displayed 27 metrics. No one could identify which 3 mattered most for any given decision. The dashboard had evolved from a decision support tool into a data consumption ritual, something to be viewed rather than used.

This pattern is consistent. Organizations that lack clarity about what they need to know compensate by measuring everything, hoping that comprehensiveness will substitute for comprehension. It does not. Herbert Simon’s observation that “a wealth of information creates a poverty of attention” is not a warning about the future. It is a description of the present.

How does signal extraction differ from data analysis?

Signal extraction differs from data analysis in that analysis asks “what does this data show?” while extraction asks “what decision does this data enable?” focusing the analytical effort on the minimum viable information needed to act rather than the maximum available information that can be displayed.

When I redesigned the reporting system, I replaced the 27-metric dashboard with 3 decision-specific views:

• Capacity view: 4 metrics (room utilization, instructor availability, enrollment pipeline, budget burn rate) that answered a single question: “Can we accept more programs this quarter?”
• Quality view: 3 metrics (completion rate, student satisfaction, instructor evaluation scores) that answered: “Are the programs we are running meeting their objectives?”
• Risk view: 3 metrics (enrollment decline rate, cancellation count, error frequency) that answered: “What is likely to go wrong in the next 30 days?”

Each view served a specific decision. The remaining 17 metrics were not deleted. They were moved to a diagnostic layer accessible only when a decision required deeper investigation. The distinction is architectural: the primary layer enables decisions. The diagnostic layer enables investigation. Mixing them produces a dashboard that does neither well.

Why do organizations resist information diets?

Organizations resist information diets because reducing the volume of reported metrics feels like losing visibility, and in political environments where information is power, no stakeholder willingly surrenders access to a metric that might prove useful in a future negotiation.

When I proposed reducing the dashboard from 27 metrics to 10, the resistance was immediate and revealing. Each stakeholder argued that their specific metrics were essential. The arguments were not about decision quality. They were about institutional positioning. The marketing team needed their metrics to justify budget requests. The operations team needed their metrics to demonstrate workload. The executive team needed all metrics to maintain the appearance of comprehensive oversight.

This is information obesity as organizational behavior: the accumulation of metrics not for decision-making but for political purposes. The dashboard becomes a territory map where each department’s metrics represent their claim to organizational resources and attention. Reducing metrics is experienced not as simplification but as reduction in status.

What principles guide effective signal extraction?

Effective signal extraction follows three principles: anchor every metric to a specific decision, impose a maximum metric count per decision context (I use 3-5), and regularly audit whether each metric has influenced an actual decision in the last 90 days.

The 90-day audit is the most powerful tool I have found. For each metric, I ask: “Has this metric changed a decision in the last quarter?” If no, it is a vanity metric, a number we track because we can, not because it informs action. I have performed this audit on three different reporting systems. In each case, 40-60% of metrics failed the test. They were consuming dashboard space, cognitive attention, and maintenance effort without influencing a single decision.

The discipline of signal extraction is ultimately a discipline of renunciation: deciding what not to measure, what not to display, and what not to consider. In an age that celebrates the accumulation of information as an inherent good, the capacity to deliberately limit one’s intake, to fast from data that does not serve a decision, is both the rarest operational skill and the most necessary. The signal is not found by adding more noise. It is found by removing everything that is not the signal.