Ask an owner what it would cost to replace them and the usual answer is one salary, a general manager's, guessed at. Ask a buyer's analyst and the answer is often two or three salaries, because the analyst has noticed that the owner also runs sales, prices every job and approves every hire. The gap between those answers is a familiar shock in the sale of an owner-run business, and it lands at a bad moment: in the buyer's model, after the owner has already settled on a view of what the business is worth.
Role Costing is the driver that takes the surprise out. It names every seat the owner occupies, deducts a benchmarked, fully loaded market cost for each, and produces a pro forma profit and loss showing the earnings after those costs, so the number the owner quotes is the number a buyer's model will reach.
Where Role Costing sits
Two drivers take normalized SDE to a pro forma figure a buyer's model will confirm. This essay covers the first.
Normalized SDE
- Role CostingThis essayNames every seat the owner fills and deducts a benchmarked, fully loaded market cost for each
- Procedure CaptureTurns each owner seat into a job a market-rate hire could do, with the owner's remaining decisions on a short list
Pro forma EBITDA
The owner's seats, from a log
The first piece is a functional organization chart. It has a seat for every function the business runs, at a minimum sales, delivery, finance, people and administration, and the owner's name in each seat where the owner's own time log shows a meaningful share of that seat's work.
The log is a modest tool, chosen on purpose. For four weeks the owner records their time in half-hour blocks, each tagged by function, and goes over it each week with whoever is running the crossing. The log decides which seats carry the owner's name, because the owner's sense of their own job is a poor guide to where the time goes. Owners remember the strategic work. They forget the hours of estimating, chasing invoices and answering staff questions that fill the days in between.
Where the owner's time goes
Illustrative. Four weeks of the owner's time, logged in half-hour blocks and tagged by function. The log, not the owner's recollection, decides which seats carry the owner's name: any function taking a meaningful share of the owner's time.
Share of the owner's logged time
In this illustration a function becomes an owner seat once it takes more than a fifth of the owner's logged time, the dashed line. Sales and estimating are over it, so both are costed as seats. Operations, at 15 percent, sits under it, which is worth knowing: it is the seat most likely to cross the line as the business grows.
Show the numbers
| Function | Share of the owner's logged time |
|---|---|
| Sales | 32% |
| Estimating | 22% |
| Operations | 15% |
| Finance | 13% |
| People | 9% |
| Admin | 9% |
| Illustrative: a fifth | 20% |
The chart is drawn by seat. The seats are defined first, from the functions, and the names go in afterwards, with any seat nobody fills left visibly empty. A chart that starts from the people and draws boxes around them describes the current staff, and it hides the very thing the exercise is meant to show, which is the seats only the owner fills.
Once the chart exists, a buyer can see how many seats the owner holds and no longer has to guess. What the owner does is known, and that is what takes off the discount for certainty.
The seats that get forgotten
Some seats are obvious, like running the business or leading sales. The ones that cause trouble in a buyer's model are the jobs the owner does without thinking of them as jobs. Pricing every quote is one. Chasing overdue accounts is another, and so are hiring, firing and settling disputes between staff, negotiating with the two or three suppliers who matter, and being the person who knows how the systems work and fixes them when they stop. A buyer will have to pay someone for each of these, and each is easy to leave off a replacement plan drawn from the owner's job title instead of the owner's week.
Other forgotten seats belong to someone else. A spouse who keeps the books unpaid, a parent who answers the phones on Fridays, a relative who does the payroll as a favour: the work is real, it costs the business nothing, and after the sale nobody will do it for free. The log should cover anyone whose unpaid work the business depends on, and each of those seats is costed at market like the owner's.
A market cost for each seat
Each seat with the owner's name in it gets a benchmarked replacement cost, from a compensation survey or a recruiter's quote for the region and company size, loaded for payroll taxes and benefits. The total comes off normalized SDE to give pro forma EBITDA.
Each seat is priced from at least two sources, a published survey and a recruiter's quote, and the schedule records the figure used and where it came from. A second source catches a survey that does not fit the role, and the recorded source means someone other than the owner can defend the figure.
The figure is then loaded to full cost, since a hire costs the business more than the salary on the advertisement.
A salary is not the cost of a hire
Illustrative. The deduction for an owner seat is what a hire will cost the business, not the salary on the advertisement: the base pay, the employer's payroll taxes, the benefits, and the bonus the market expects for the role.
In the illustration, a seat with a $95,000 base salary costs the business about $125,000 a year once payroll taxes, benefits and the bonus the market expects are added. Deducting $95,000 would understate the seat by nearly a quarter, and a buyer's model will not make that mistake. A deduction sized from evidence also comes out smaller than one sized from a buyer's suspicion, which is how the work takes off the discount for magnitude.
A pro forma the business can bear
The pro forma profit and loss has to stay positive after every replacement cost, with a staffing plan that sets the seat, timing and budget for each hire or promotion and shows the cash to fund the hires before they pay for themselves.
It is built month by month from the normalized statements, normalized SDE less the loaded replacement schedule, so the seasonal pattern in the cash the hires need can be seen. The hires are put in order by how much of the owner's time each frees per dollar of cost, each with a start date and a budget, and the seat the owner means to keep is chosen and stated in advance.
Ordering by time freed per dollar can upset the owner's plans. The instinct is often to hire the most senior role first, a general manager to take over everything. The log often shows that a less senior seat, an estimator or an office manager, frees more of the owner's week for less money, and freeing that time first is what lets the owner hire for the senior seats and hand them over properly.
What has to be in place first
The driver needs normalized SDE from the previous span, because the pro forma is built on it and a reconstructed figure cannot be costed. It needs the owner to log time honestly for four weeks. The chart is only as true as the log, and an owner who records the role they ought to have defeats the exercise. And it needs a compensation survey or a recruiter, budgeted before the costing starts, because a guessed salary is the thing the driver is there to remove.
Knowing that it works
The owner's logged time by seat, reported weekly while the log runs, shows which seats cross the line and get costed. The share of owner seats with a sourced benchmark should reach all of them. And once the pro forma exists, its EBITDA margin after replacement costs, reported monthly, is the number a buyer will model.
The pro forma holds if the owner can present it to the outside accountant and defend every salary and its source with the lead out of the room. The owner should also be able to say in one sentence which seat they will keep and why, with that seat's market cost already in the pro forma. The second answer goes further than it looks, because the seat the owner keeps decides which buyer the business is being prepared for: one who wants the owner to stay, or one who wants it to run without them.
What it does to value
Role Costing is the one driver on this bridge that lowers value. It takes the cost of the owner's seats off the honest earnings, and the framework calls that a correction, because a buyer's model would have made the same deduction anyway. The business itself is unchanged. The owner has learned the figure the buyer was always going to use.
So it ranks last on value and is still done early. Knowing the number lets the owner decide whether the crossing is worth making before spending money on it, and a pro forma that matches the buyer's model takes away one of the largest reasons a buyer has to reopen the price after diligence.
Why it slips, and what holds it
The owner drifts back into seats that were costed and then filled, stepping in to help a struggling hire and staying. Or a manager leaves and the owner absorbs the seat instead of refilling it.
Redrawing the chart every year and logging the owner's time for one week each quarter means a drift back into a seat shows in the log within a quarter, before it shows in the earnings. Refreshing the replacement benchmarks with the annual review means a vacated seat can be refilled at a known cost, so it is not absorbed by default.