Contractors, Freelancers, and the Margin You Thought You Bought
A £450/day contractor and a £340/day employee are not £110 apart — they are two different cost shapes, and agencies routinely misprice both. When the contractor is genuinely cheaper, when they quietly eat the margin, and why contractor cost so often never reaches the project report at all.
Comparing a contractor's day rate to an employee's day cost is the most natural calculation in agency resourcing, and it is done wrong in both directions. Wrong when the contractor "looks expensive": a £450/day contractor against a £55k developer's £340/day fully-loaded cost is not a £110 premium, because the employee costs £340 on every working day of the year — bench, ramp-up and quiet Augusts included — while the contractor costs precisely nothing the day the project stops. Wrong when the contractor "fills in for a while": a long-running contractor doing steady, predictable work an employee could do is paying the flexibility premium every single day for flexibility nobody is using — routinely £25–35k a year of quiet margin leak per seat.
Underneath the pricing question sits a measurement problem that is usually worse than either mistake: contractor cost frequently never reaches the project margin report at all. The hours are in Jira; the invoices are in accounts payable; and at agencies that cost worklogs at a blended employee rate, the contractor's actual, higher cost silently never lands on the project that consumed it.
This piece gives the two cost shapes, a rule for when each one wins, a worked example of the long-stay contractor, and the plumbing fix for the invisible-cost problem.
There is a question every agency answers weekly without ever quite writing down the maths: do we put an employee on this, or bring in a contractor? The inputs feel simple — a day rate on one side, a salary on the other — and the comparison gets made in seconds, usually in whoever's head is nearest the staffing sheet.
I made that comparison casually for years, and both of the ways I made it were wrong. The contractor is not simply "more expensive per day", and the employee is not simply "cheaper if we have one free". They are different cost shapes — one variable, one fixed — and treating them as two prices on the same axis is how agencies end up paying premium rates for work with no premium in it, while simultaneously believing contractors are too dear for the work that actually justifies them.
Two cost shapes, not two prices
Start with the honest per-day numbers. A mid-level employee on £55,000 fully loads to roughly £65k a year — employer's NI, pension, equipment, licences — which across ~190 deliverable days is about £340 a day. A mid-level contractor in the same stack quotes, say, £450 a day, with the NI, pension, kit, holidays and gaps between gigs all folded into the rate. That is the comparison as usually made: £110 a day apart, contractor loses.
Now add the dimension the day-rate comparison hides: which days you pay for.
- The employee costs £340 on every one of ~220 working days a year, regardless of what those days contain — delivery, bench between projects, ramp-up on a new account, the quiet fortnight in August. The cost is fixed; the value delivered against it varies.
- The contractor costs £450 only on days you booked. The day the project pauses, the cost stops the same afternoon. No bench, no ramp you're funding, no quiet August on your payroll.
So the £110 gap is not a premium for the same thing — it is the price of transferring utilisation risk. And whether that transfer is worth £110 a day depends entirely on how much of that risk you actually have. An agency with a deep, steady pipeline is buying insurance against a fire that rarely starts. An agency with spiky, uncertain demand is buying insurance it will collect on within the quarter.
When each shape wins
The clean rule: pay the fixed shape for steady work, the variable shape for spiky work. In practice:
| Situation | Better buy | Why |
|---|---|---|
| Steady, forecastable workload, 9+ months visible | Employee | You'll use the fixed days; the premium buys nothing |
| A spike: one project, 6–12 weeks, then nothing | Contractor | The premium is cheaper than the bench that follows |
| Niche skill needed occasionally | Contractor | You cannot keep a specialism warm on the bench |
| Skill you sell repeatedly and want to compound | Employee | Contractors take the learning with them at the gate |
| Covering a hiring gap you're actively closing | Contractor, briefly | Fine as a bridge; ruinous as a lifestyle |
The expensive failure hides in the last row. The contractor who arrived "for six weeks while we hire" and is still there fourteen months later is the agency's most reliable quiet leak. Run the numbers on that seat: 220 days at £450 is **£99k**, against ~£65k fully-loaded for the employee who would by now exist if the hiring had happened. That is £25–35k a year, per seat, paid as a flexibility premium on work that stopped being flexible a year ago — invisible in any weekly view, because each individual week the contractor was genuinely needed.
The discipline that fixes it is boring and effective: every contractor engagement gets a review date at booking, and at each review the question is not "are they useful?" (they always are) but "is this work still spiky?" The day the honest answer is no, the seat is mispriced.
The measurement problem underneath
All of the above assumes the contractor's cost is at least landing on the project. At a surprising number of agencies, it isn't — and this is the part that corrupts the margin numbers rather than just the resourcing decisions.
The failure has a standard anatomy. The contractor logs hours in Jira like everyone else, so the hours are in the delivery data. But their cost arrives as an invoice, into accounts payable, into a "subcontractors" line in the management accounts — a different system, a different cadence, a different owner. Then one of two things happens. Either the margin report costs worklogs at a blended employee rate, quietly pricing £450/day hours at £340-day maths; or, worse, the contractor isn't in the cost ledger at all and their hours are costed at zero. In both cases the project consuming the contractor reports a margin it does not have — and precisely the projects leaning hardest on expensive external help are the ones most flattered, one more variant of the quiet losses that surface at year-end with no owner.
The fix is a rule, not a tool: anyone who logs hours has a real cost rate in the ledger — employees at fully-loaded cost, contractors at their actual day rate converted to hours, updated when the rate changes. In Saldo contractors are simply people with their own cost rate: their Jira worklogs are costed at their number the moment they land, so the project that used £450/day hands reports £450/day cost, continuously, next to overhead and everything else. One ledger, no blended fictions.
One more asymmetry worth pricing: who carries the overrun
A final wrinkle for fixed-price work. When a fixed-price project overruns with employees on it, the overrun costs you their fully-loaded rate — painful, but the cost of the extra fortnight is the salary you were paying anyway. When it overruns with contractors on it, every extra day is new money out of the door at the full external rate. The estimate variance — and estimates run systematically optimistic — is carried by the agency at the contractor's price.
The implication for pricing is straightforward and widely skipped: contractor-staffed fixed-price work needs either a fatter contingency or a contract shape that shares the variance. Quoting contractor-delivered work at the same margin assumptions as employee-delivered work prices the risk at zero, and the risk is not zero.
The shorter version: contractors and employees are not a cheap option and a dear option — they are a variable cost and a fixed one, and each is the wrong shape for the other's work. Pay the premium where the spike is real, review the seats where it isn't, and above all make sure the contractor's actual rate reaches the project report — because the margin you thought you bought was computed on somebody's blended average. If you'd like to see what your projects look like with every pair of hands costed at its true rate, the 15-minute demo runs on your real Jira data.
Continue inside Saldo