Salary Divided by Hours Is Not a Rate. It Is a Wish.

Almost every company that costs a job, sets a charge-out rate or budgets a project is using a number built the same way: annual salary divided by a nominal number of working hours. It is simple, everyone understands it, and it is nowhere near the true cost of employing that person — because it is wrong in both the numerator and the denominator — always in the same direction, which is why nobody catches it. The rate is too low, so work looks more profitable than it is, so nobody investigates.

The numerator has to be the fully loaded cost of employment: wage plus every allowance, plus the accruing liabilities the employee is earning as they work, plus the recurring costs of having them employed at all. The denominator has to be the productive hours genuinely available after leave, public holidays, training and the normal non-productive share of a working year. Neither correction is difficult and neither is controversial once it is written down. What makes them worth doing is the size of the combined error, which is large enough to change pricing decisions.

What belongs in each half of that fraction, the two things that should not be in there at all, and how to build the number in an afternoon.

What belongs in the numerator

Work down this list for one role in your own business and the gap from the salary figure will be obvious.

Wage and all allowances — basic, housing, transport, mobile, site, shift. The total cash cost, not the basic.

The accruing liabilities. End-of-service gratuity and annual leave are earned every day the person works, which means they are a cost of the period in which the work happened rather than of the period in which they are paid out. Leaving them out is the single largest omission in most rate calculations, and it is dealt with in treating both as liabilities that accrue rather than as annual journals.

Insurance, medical and any other mandatory cover.

Visa, permit and document costs, amortised over the period they cover rather than expensed in the month they land. These are consistently left out because they sit in an administrative account rather than in payroll, which is one more reason the expiry register should carry a cost field.

Recruitment cost, amortised over expected tenure. If a role turns over every eighteen months, that is a real per-hour cost of that role, and it is the number that makes retention arguments concrete.

Recurring employment costs: accommodation or allowance where provided, transport where provided, tooling, protective equipment, uniforms, the licences and certifications the person needs to do the job, and their renewal.

Training and certification required to keep them able to work.

Airfare or repatriation provision where the contract or policy provides for it.

The pattern is that everything in payroll gets included and everything outside payroll gets forgotten. A useful test: go through the general ledger looking for accounts that only exist because you employ people, and check each one against your rate build-up.

What belongs in the denominator, and what does not

This half is smaller and produces a surprisingly large part of the error.

The starting point is the contractual hours in a year. From that, subtract annual leave, public holidays, expected sick leave at a rate taken from your own history rather than from a guess, and training and mandatory non-productive time.

Then the judgement. For a billing business, subtract the realistic non-billable share — the administration, the internal meetings, the proposal writing that is part of the job and is not a client's hour. That figure should come from measurement rather than from optimism, and the machinery for measuring it is the subject of utilisation and capacity planning.

Two things that should not be in the denominator. Overtime hours, if overtime is separately paid — including them credits the base rate with hours that carry their own cost. And hours nobody has ever worked: a denominator built on a theoretical year that the business has never achieved produces a rate that only recovers in a year that does not happen.

The two arguments this settles

Charge-out rates. A professional services business pricing from an unloaded cost is running a margin lower than it believes by a consistent amount. Because the error is uniform, it does not show up as one bad project — it shows up as a business that is busy and not making money, which gets diagnosed as a pricing problem in the market rather than as an arithmetic problem at home. The neighbouring effects are worked through in client profitability.

Make, buy or subcontract. An in-house cost that omits half of employment cost will lose every honest comparison against a subcontractor, and companies do genuinely subcontract work they should keep, and keep work they should subcontract, on the strength of a number that was never right. This is also where the recruitment and turnover component earns its place: a role with high turnover is more expensive per productive hour than its salary suggests, and that is exactly the role most likely to be compared against an external supplier.

For a contractor, the same rate feeds every job cost, so an understated labour rate makes every project's margin look slightly better than it is — the error that never prompts an investigation because it never produces bad news.

Building it in an afternoon

1. Pick three roles, not all of them: one that bills, one that does not, and one with high turnover.
2. Build the numerator from the list above, taking the figures from the ledger rather than from memory. Where a figure does not exist — recruitment cost per hire, say — estimate it explicitly and mark it as an estimate.
3. Build the denominator from the calendar and from your own absence history.
4. Compare the result to the rate you are currently using. Write the ratio down. That number is the conversation.
5. Decide what to do with it: rates, or pricing, or nothing — but as a decision rather than as an oversight.
6. Then agree who owns the rate and how often it is rebuilt. Once a year, on a date, by a named role.

The data this rests on is the employee master, the wage structure and the service dates — the same three that everything else in this area depends on, and the reason the payroll module is the one everybody defers and regrets. Where the rate is currently a number somebody inherited and nobody can derive, establishing what it should be is a short exercise with an immediate effect on pricing, and it is one of the first things worth settling in any implementation where job costing is the reason for the project.