Put a number on your coordination friction.
Waiting, handoffs, and rework between teams consume time that never reaches a customer. This is the model I use to size it. Change any assumption and the arithmetic follows. Nothing is sent anywhere, and the calculation runs entirely in your browser.
Execution leakage
a year, or 9 to 16% of total salary expense.
| Layer | Salary | Low | High |
|---|---|---|---|
| Executive | — | — | — |
| Senior leadership | — | — | — |
| Team members | — | — | — |
What this is, and what it isn't.
The arithmetic is deliberately simple. Personnel cost is a share of operating expense, which is a share of revenue. That pot is split across three layers. Each layer loses some share of its week to waiting, handoffs, and coordination, and that time is costed at salary. The output is a range, because a point estimate would imply precision the inputs do not support.
Some inputs are benchmarked and some are my judgement, and it matters which. Operating expense as a share of ARR comes from a published survey of more than a thousand private B2B SaaS companies. The share of that spend going to people is my estimate. The time each layer loses to coordination, and the share you could recover, are also my estimates, drawn from what I have seen inside delivery organizations rather than from any survey. There is no published benchmark for coordination loss that I would be willing to cite.
The figures are a model, not a measurement from anyone's books. If you know your real personnel split, or you have measured how much of a week goes to coordination, put your own numbers in and my defaults stop mattering. The revenue-per-employee readout is there as a check: if it lands far above the industry median, your cost inputs are probably too low and the result is understated.
Two more things worth being straight about. Costing lost time at salary is the conservative reading, since capacity redeployed to product or customers is usually worth more than the salary attached to it. And no organization recovers all of it, because some coordination is simply how work gets done. That is why recovery is an input rather than a number I assert.
Where the numbers come from.
Operating expense as a share of ARR. Benchmarked.
Median total operating expense is 96% of ARR for bootstrapped companies and 101% for equity-backed ones, from the 2026 SaaS Capital Spending Benchmarks, their fifteenth annual survey, covering more than 1,000 private B2B SaaS companies. Their departmental medians are Sales 15%, Marketing 8%, Customer Support and Success 9%, R&D 22%, G&A 15%, hosting 5%, DevOps 4%. The default here is 96%, the lower of their two headline figures. SaaS Capital, 2026 Spending Benchmarks
Personnel as a share of cost. Derived, because nobody publishes it.
There is no clean published benchmark for this ratio, and there is a structural reason. Companies do not report personnel cost as a line item. It is spread across cost of revenue, sales and marketing, R&D, and G&A, so surveys report spending by department rather than splitting people from everything else. So I derived it instead.
Taking the SaaS Capital departmental medians above and applying a people-intensity to each: sales 85%, marketing 40% because programmes and ad spend are not people, support and success 90%, R&D 90%, G&A 60% because it carries rent and professional fees, hosting 0%, DevOps 80%, professional services 85%. That gives personnel at about 72% of total cost, or 77% if you exclude hosting and cost of revenue, and about 62% of revenue.
The intensities are my judgement, so treat the result as a considered estimate rather than a benchmark. Two things give me some confidence in it. Moving every intensity by 10 points only moves the answer to between 62 and 81% of total cost, so the conclusion is not delicate. And it lands independently on the same place as OPEXEngine, who found compensation alone exceeded 70% of revenue and had fallen roughly 10 points over five years, implying about 60 to 65% today against my derived 62%. OPEXEngine on the salary component of SaaS operating expense
Revenue per employee. Benchmarked, used only as a check.
Median revenue per employee is about $130,000 for private SaaS companies and about $283,000 for public ones, from the KeyBanc Capital Markets annual SaaS survey. It is not an input to the calculation. It is there so you can see when your cost assumptions imply an implausibly efficient company. Reported via OPEXEngine's 2025 sector brief
Rule of 40 does not set your total cost. Your margin does.
Total cost is 100% of revenue minus your operating margin, and margin is only half of a Rule of 40 or Rule of 60 score. Growth is the other half and does not enter this calculation at all. So there is no single Rule of 40 number to use. A company growing 40% at breakeven clears Rule of 40 at 100% of revenue. One growing 10% on a 30% margin clears the same bar at 70%. The benchmark median, 96%, is a roughly 4% margin, which clears Rule of 40 if you are growing 36% and Rule of 60 if you are growing 56%. Median growth is nearer 14 to 18%, which is why the median company is not at either bar.
Practically: use 100 minus your own operating margin and ignore the Rule score for this input. If you are profitable at a 25% margin, use 75%. If you are burning to grow, use 100% or more.
One caveat the model cannot resolve. It treats coordination loss as independent of how efficient a company is, and that cuts both ways. A company that reached Rule of 40 through operating discipline probably loses less than these rates suggest. A company that got there by cutting headcount may lose more, because the same coordination lands on fewer people. The percentage of salary expense is identical in every scenario, near 9 to 15%. Only the dollar base moves, which is why the ratio is the more portable number.
Coordination loss and recovery. My judgement, unbenchmarked.
The 10 to 20% for executives, 20 to 30% for senior leaders, and 5 to 10% for team members are my estimates from delivery work, expressed as a share of a 40-hour week. So is the 50% recovery default. I know of no credible published benchmark for either, and I would rather say that than cite something thin. These are the inputs to change first if you have measured your own.
Tell me what's stuck.
If the number surprised you, the interesting question is where the friction actually sits. Send me the initiative that isn't moving. Within a week I'll tell you whether I can help, what an engagement looks like, and what it costs.