What does it mean that information bottlenecks?

When you cannot get the information you need to proceed the information flow is the constraint.

Example: You are a product manager ready to finalize next quarter's roadmap. The data you need lives in three places: customer churn reasons in the support team's Zendesk instance, usage analytics in a data warehouse only the engineering team queries, and competitive positioning notes in a sales deck that was last updated six months ago. You email three people. One replies in four days with a CSV you cannot parse without context. One replies in nine days with a dashboard link that requires access you do not have. One never replies. Three weeks pass. Your roadmap is late — not because you lacked the judgment to prioritize features, not because you lacked the time to build the document, but because the information you needed to make defensible decisions was scattered across systems, owned by people with other priorities, and delivered in formats that required translation before use. The bottleneck was never your planning skill. It was the information pipeline feeding your planning process.

Try this: For the next five working days, keep an information request log. Every time you need a piece of information to proceed with a task — a number, a document, a clarification, a dataset, an answer — write down: (1) what you needed, (2) where you looked first, (3) how long it took to obtain, and (4) in what format it arrived versus what format you needed it in. At the end of five days, calculate your average information retrieval time and classify each request into one of the four types: scarcity (could not find it), latency (found it but waited too long), format mismatch (got it but had to transform it), or overload (had too much and could not extract the signal). Identify which type dominated. That is your primary information bottleneck pattern.