Information Flow Bottleneck Identification¶

Why It Matters¶

Organizational performance is constrained by information flow. Critical information exists somewhere in the organization but fails to reach decision-makers. Decisions are made with incomplete information because relevant knowledge is trapped in silos. Problems are diagnosed incorrectly because the people investigating lack access to diagnostic data. The default assumption is that information flows freely and missing information reflects lack of effort. Reality: information flow is constrained by structure—formal reporting lines, communication channels, attention capacity, and trust networks. Bottlenecks aren't malicious; they're structural. Making them visible allows redesigning around them.

The Key Move¶

For any decision or problem, ask: "What information would be needed to handle this optimally? Where does that information currently exist? Through what path would it need to travel to reach the decision-maker?" Trace the path step by step, identifying points where: information must pass through individuals with limited bandwidth, cross organizational boundaries requiring translation, compete for attention with other priorities, or require trust that doesn't exist. These are your bottlenecks. Test: can information actually make the journey, or does it die at a bottleneck?

Classic Application¶

The Bay of Pigs invasion failure (1961). Field intelligence indicating the invasion would fail existed in the CIA and State Department. This information needed to reach President Kennedy's decision-making team. But it bottlenecked at multiple points: lower-level analysts hesitant to contradict leadership assumptions, intermediary briefers summarizing away dissenting views to avoid appearing uncertain, senior advisors filtering information to avoid appearing weak on communism. Each bottleneck removed information that would have prevented the decision. The problem wasn't lack of information; it was bottlenecks that prevented its flow to the decision point.

Surprising Application¶

On-call incident response in software operations. Critical performance degradation occurs. The on-call engineer needs to diagnose it. Essential information exists across multiple monitoring systems, tribal knowledge from past incidents, and contextual understanding held by the engineer who wrote the code (now asleep). Information bottlenecks: monitoring systems not integrated, requiring manual correlation; documentation from past incidents not searchable or linked to current symptoms; tacit knowledge not externalized; escalation paths unclear. The on-call engineer thrashes while people who could solve it in minutes lack context or notification. The fix isn't "better engineers"—it's eliminating bottlenecks through unified observability, runbooks that capture tacit knowledge, and escalation systems that route problems to knowledge-holders.

Failure Modes¶

Creating information flood by removing all filters—some information reduction is necessary; not all information is equally valuable. Over-designing information systems that are too complex to use—the bottleneck becomes system complexity itself. Assuming technology alone solves information flow—often bottlenecks are social (trust, status, motivation). Ignoring that some information bottlenecks are protective—preventing information overload or maintaining confidentiality. Treating all information as equally urgent—prioritization is necessary filtering.

Go Deeper¶

Galbraith, "Organization Design: An Information Processing View," Interfaces 4(3), 1974; Burt, "Structural Holes and Good Ideas," American Journal of Sociology 110(2), 2004.