What does it mean that human bottlenecks in team systems?

In collaborative work specific people are often the constraint.

Example: A four-person development team ships features through a standard pipeline: build, code review, QA, deploy. Three developers finish their work in a day or two, but every pull request must be reviewed by the senior engineer — the only person the team trusts to catch architectural mistakes. She reviews carefully, averaging ninety minutes per PR. With four developers producing roughly two PRs each per day, eight requests land in her queue daily. She can review five. By Friday, fourteen PRs are waiting. Three developers sit idle for stretches of the afternoon, blocked on review. The team's throughput is not four developers times their velocity. It is one senior engineer times her review capacity. She is indispensable, respected, and the single constraint throttling the entire system.

Try this: Identify one collaborative workflow you participate in — at work, in a side project, or in a household. Map every step that requires a specific person''s involvement before the next step can proceed. For each person-dependent step, count how many items are currently waiting for that person. If you find a step where more than three items are queued behind one individual, you have found a human bottleneck. Write down: ''In [system], [person/role] is the constraint because [N] items are waiting on them and no one else can do [step].'' If you are that person, write it about yourself.