Most contract staffing decisions are made on the wrong number

When a team needs capacity and someone proposes contract, the analysis that follows is almost always the same. Take the contractor's bill rate, annualise it, and set it against the fully loaded cost of an employee. The contractor looks expensive. Somebody notes they come with no benefits and no severance, the numbers get closer, and the decision goes to whoever argues more forcefully.

Both sides of that comparison are real, and the comparison itself is close to meaningless. It prices two different instruments as though they were the same instrument bought on different terms. A contractor is a way of buying capacity with an end date attached, and the end date is the entire product.

Priced that way, the question changes shape completely. It stops being “which is cheaper per hour” and becomes “does this work genuinely end, and what happens when the person who did it leaves?” Those are answerable questions. They just aren't the ones on the spreadsheet.

Some work has a real end date. A migration completes. A certification is achieved. A release ships, a backlog clears, a pilot concludes with a decision to build or stop. When that's genuinely true, contract is the correct instrument, and hiring a permanent employee for it creates a problem you'll have to solve later with a reorganisation or a redundancy.

Other work only looks bounded. The platform team stood up for a migration that now runs the platform. The analyst brought in for a reporting push who's become the person who knows how the reporting works. The end date existed on a plan and then quietly stopped existing, and nobody revisited the instrument because the arrangement was working.

The test is simple to apply and slightly uncomfortable to answer honestly. Write down the date and the condition. What's true on the day this work stops? If nobody can state it, the work has an aspiration, not an end date, and it should be staffed permanently. Most expensive contract arrangements began as a genuine end date that nobody went back and checked.

Every contract engagement has an exit, and the exit has a cost that never appears in the rate comparison. When the person leaves, some proportion of what they learned leaves with them. For genuinely bounded work that proportion is small. The migration is done, the knowledge was about a system that no longer exists in that form, and the loss is close to zero.

For work that turned out to be permanent, the proportion is enormous. The contractor who spent 18 months becoming the only person who understands why a critical integration behaves the way it does is carrying institutional knowledge. The day their engagement ends, the organisation buys it back at a much higher price, usually by converting them at a premium, or by paying someone else to rediscover it slowly.

That's why the instrument matters more than the rate. A correctly-scoped contract engagement ends cleanly and costs exactly what it said it would. An incorrectly-scoped one accrues a liability quietly for a year and then presents the bill all at once. The bill is rarely recognised as the consequence of the original decision, because by then it looks like a retention problem.

The most common reason organisations reach for contract has nothing to do with the shape of the work. Headcount is frozen, the work isn't, and contract labour is funded from a budget line that isn't frozen. Everybody in the conversation knows this, and it's usually the pragmatic thing to do. The alternative is that necessary work doesn't happen because of an accounting boundary.

The trouble is what happens next. The arrangement works, so it continues. A year later the organisation has a meaningful share of its capability sitting outside its headcount, invisible in workforce plans, excluded from succession thinking, and costing more per unit of output than the employees doing similar work. When the freeze lifts, converting that population is a large, sudden and politically awkward exercise, and some of the best people will have moved on by then.

None of this argues against doing it. It argues for doing it consciously, with the end designed at the start: which of these engagements convert when the freeze lifts, on what timeline, and what's the plan if it doesn't lift. Written down at the beginning, that's a paragraph. Discovered at the end, it's a reorganisation.

Contract-to-hire is sold as risk reduction: try before you commit, on both sides. Where the work is genuinely uncertain and the arrangement is genuinely two-way, that's fair and it works. It's worth being clear-eyed about the selection effect, though, because it's strong and it runs in one direction.

The candidates most willing to accept a contract-to-hire arrangement are the ones with the least bargaining power: between roles, newer to the market, or without a competing permanent offer. The candidates least willing are the ones currently employed and being courted elsewhere, who won't resign a permanent position for a conditional one. If the role is one where the strongest people are typically already employed, contract-to-hire quietly filters them out before the process begins.

That's a trade, and for roles where the market is deep and the risk is real, it's often the right one. For scarce or senior roles it usually isn't, and organisations that apply it as a blanket policy tend to conclude the market for those roles is weak, when what they've actually done is decline to compete for most of it.

The sequence that works puts the rate last. First, establish whether the work has a real end date, stated as a date and a condition. Then decide what happens to the knowledge when the engagement ends, and whether that loss is acceptable. Then choose the instrument. Only then talk about the number, because by that point you know what you're buying and can tell whether it's worth the price.

Run in the usual order, the number drives the decision and the decision drives a year of consequences nobody attributed to it. Run this way, the rate conversation gets easier as well as better. A well-specified bounded engagement is a straightforward thing to price, and a supplier who understands the end date can staff against it properly rather than hedging.

Contract staffing is a good instrument, badly used more often than not, mostly because the decision is framed as a cost question when it's a scheduling question. The organisations that get value from it aren't the ones that negotiate hardest on rate. They're the ones that know, on the day they sign, what's true on the day it ends.