Most bottlenecks are structural, not personal.

When delivery slows, the loudest narrative is often “we need more people” or “this vendor is slow.” Sometimes that is true. More often the constraint is a queue nobody named: a single approver, an environment that is always busy, a handoff where work sits because ownership is fuzzy, or a dependency team that is not staffed to match your timeline. The work is visible; the wait time is treated as normal weather.

Calling that a motivation problem or a talent problem feels satisfying. It is usually a system design problem wearing a human face.

Map the actual flow

Before you optimize, sketch how work really moves: not the ideal RACI from the deck, but the path an item takes from intent to production.

Look for:

• Queues: where items sit longest, and whether that time is tracked or invisible.
• Approval stacks: single points where everything funnels.
• Context loss: handoffs where the next person has to rediscover what the last person already knew.
• External dependencies: contracts, security reviews, other teams with their own full plates.

If you cannot point to where work waits, you are probably measuring activity (meetings held, tickets closed) instead of flow (time from commit to value, or from request to resolution). Flow metrics are uncomfortable because they make the bottleneck undeniable.

Fix sequencing first

Throwing people at a bad sequence rarely fixes it. If step B always waits on step A, two teams working in parallel on B do not help until A is predictable.

Better sequencing and clearer ownership often unlock more value than adding headcount or buying another tool:

• Make the constraint visible: WIP limits, a shared board, or a simple aging report on “stuck” items.
• Protect the constraint: if one role or system is the limit, everything else should queue before it in an orderly way, not flood it randomly.
• Reduce batch size: smaller chunks of work clear queues faster and surface problems earlier.

Sometimes the fix is political: a policy that requires three signatures for a low-risk change is the bottleneck. No amount of “working smarter” compensates until someone with authority agrees to change the rule or delegate.

When it really is personal

People matter. But chronic underperformance in one role is worth separating from chronic overload in a role design that no one could sustain. If you fix the structure and the same pain persists, you have better signal for a coaching or staffing conversation.

Start by assuming the system is honest. It usually is, and it is usually hiding in plain sight.