The bottleneck is usually the managing partner. We say that with the empathy of partners who have been that bottleneck, because the pattern is so common that it reads as a developmental stage rather than a flaw. This is a process leak, and the cost compounds quietly. Delayed proposals, delayed vendor decisions, delayed hires. None of it shows up as a line item. All of it shows up in margin about a year later.
The process bucket is where this leak lives. It is not a cost leak, because no one is overpaying for anything, and it is not a revenue leak, because the rate card is intact. The drag comes from the speed at which the firm can decide, and the speed at which the firm can decide is capped by the calendar of one person.
How the Bottleneck Develops
The firm grew. The approval structure did not. That is the entire mechanism, and it is worth sitting with, because it explains why the sharpest operators end up here.
When a firm is small, routing every decision through the founder is correct. The founder has the context, the relationships, and the most to lose. A single set of eyes on every vendor payment, every new hire, and every client proposal is not a control problem at that stage. It is good judgment applied where judgment is scarce.
Then the firm doubles, and doubles again. The vendor payments multiply. The hires multiply. The proposals multiply. The approval structure, though, stays exactly where it was: one person, one inbox, one signature. Every expense over $500 still routes through the same desk it routed through when the firm was a quarter of its current size.
This is not a discipline problem. The managing partner has not gotten worse at deciding. The firm has simply outgrown the infrastructure that was built to decide for it. The missing piece is delegation infrastructure, and almost no firm builds it on purpose. It has to be installed deliberately, usually after the absence of it starts to hurt.
What It Actually Costs
The cost is real, and it is mostly invisible because it lives in time rather than dollars.
Start with proposals. A proposal that waits two to five days for a signature is two to five days of billing that has not started. At a firm billing in the hundreds of thousands per engagement, a few days of delayed start, repeated across a year of proposals, is a meaningful number on its own.
Then vendor decisions. When a vendor contract needs a sign-off that does not come, the firm runs on the expiring agreement at the old, non-negotiated rate. The renewal that would have saved money waits in a queue. The savings are not lost forever, but they are deferred quarter after quarter, which has the same effect on this year's margin.
Then hires. A hire delayed is revenue delayed, because the work that hire would have produced does not get produced. In a professional service firm, where the people are the product, a slow hiring decision is one of the most expensive forms of waiting there is.
The published taxonomy does not name approval bottlenecks separately, because they cut across all three buckets. In the diagnostic work we do, the recoverable range for this pattern falls in the $40K to $200K window for a firm somewhere between $5M and $50M in revenue. That figure is BaxterLabs-observed rather than pillar-published, and the difference is worth stating plainly. It is what the mechanism tends to cost when we find it, not a number on the public profit leak page.
The Diagnostic Question
Here is the question we ask, and it is uncomfortable on purpose.
How many decisions in a normal week require your sign-off, and how many of those would the firm make correctly without you in the loop?
The first number is usually easy to estimate and larger than expected. The second number is the one that lands. At most firms in this size range, the honest answer is that a large share of those decisions would be made correctly by someone else, on their own, with no drop in quality. The approvals are not adding judgment. They are adding latency.
When the second number is high, the finding is not that the managing partner is doing something wrong. The finding is that the firm has capable people waiting on a queue that exists only out of habit.
The Structural Cure
The cure is structural, which is the good news, because structural problems have structural fixes that do not depend on anyone working harder.
Set approval thresholds by dollar amount. Decisions below a clear line, say anything under $5,000, do not require partner approval at all. They get made by the person closest to the work, who usually has the better context anyway. The line can move as trust builds, but the principle is fixed: the partner approves fewer things at a higher level, not more things at every level.
Then make delegation feel safe, because the reason most partners hold the line is not control for its own sake. It is the fear of not knowing. Replace approval-by-approval visibility with a weekly spend summary. The partner still sees everything. They just see it as a digest after the fact rather than as a gate before it. Visibility and control are not the same thing, and a firm that treats them as the same pays the bottleneck tax indefinitely.
The goal is not to stop approving things. The goal is to approve fewer things, at a higher level, with the infrastructure underneath to make the smaller decisions safely. That infrastructure is the piece the firm never built while it was busy growing. Installing it is what turns a one-person decision calendar back into a firm that can move at its own size.
Three years of that latency compounds quietly inside the operating line. By the time it surfaces in margin, the cause looks like a hundred small delays and the cure looks like a habit no one remembers starting.