Make or Buy, Decided on the Wrong Number
Outsourcing decisions taken on a full absorbed rate are systematically biased toward buying. The arithmetic that corrects it.
One piece of work, 200 hours, four ways of answering whether to outsource
Includes overhead that will not disappear.
Wage and on-costs only; the pool is already paid for.
Looks cheaper than the full rate.
Supplier, plus the overhead that stays, plus managing them.
Against the full rate the supplier wins by £1,184; against the honest comparison it loses by £3,850. The difference is the overhead that does not leave when the work does. One employer's own comparison, not a benchmark.
A make-or-buy decision taken by comparing a supplier's quote to an internal full absorbed rate is biased toward buying, because the internal figure contains overhead that will still be there after the work leaves. The supplier's price contains only the supplier's costs.
The practical lesson in “Make or Buy, Decided on the Wrong Number” is that a convenient number should begin a review rather than end one. A team assessing explore the available controls for monitask pricing can add timely time and project context, with transparent settings, proportionate access and a documented correction process around every material use.
Comparing them is comparing a number that includes a share of your building to one that does not.
As an outside benchmark for “Make or Buy, Decided on the Wrong Number”, see McKinsey organisation insights. Keep the source record, calculation rule, approval and resulting decision visible as separate steps.
What actually changes
The honest comparison is between the costs that would disappear if the work went, and the costs that would appear if it did.
Everything else is common to both options and belongs in neither side of the comparison.
What disappears
The hours themselves — wage and on-costs — if the people go or are redeployed to other chargeable work. Consumables. Any equipment specific to that work, eventually.
That is usually all. The supervisor, the premises, the systems and the central functions do not go.
What appears
The supplier's invoice. The cost of managing the supplier, which is real and routinely omitted. Transition cost. Any quality or rework cost that comes back to you.
The second of these is typically ten to twenty per cent of the contract value in the first year.
The redeployment question
The marginal saving is only real if the freed hours are used for other chargeable work or the people leave. If neither happens, the saving is zero and the supplier's invoice is a pure addition.
This is the single most common error in outsourcing arithmetic.
The step-cost reality
Overhead does fall eventually, in steps: a supervisor when enough work has gone, a floor when enough people have gone.
Modelling it as proportional makes small outsourcing decisions look better than they are; modelling it as fixed forever makes large ones look worse. Steps are the honest shape.
The capacity argument the other way
Where the operation is at capacity and the alternative to buying is turning work away, the comparison changes completely: the relevant internal cost is the margin on the work that would be displaced.
That is a much higher number and it frequently favours buying.
The reverse decision
Bringing work in-house is biased the other way by the same arithmetic: the internal cost looks like the full rate when the marginal cost is much lower.
Both decisions should use the same framework and usually do not.
What to compare, stated plainly
Costs that would stop, against costs that would start. Nothing else belongs in the comparison, in either direction.
The things people forget
Management of the supplier. The hours spent specifying and tendering. The rework that comes back. The knowledge that leaves and does not return. The price in year three.
Five items, all real, none in the spreadsheet.
The non-financial half
Capability, control, confidentiality, flexibility and risk. The arithmetic settles whether the decision is affordable; it does not settle whether it is wise.
Where the arithmetic is close, the non-financial considerations should decide, and saying so explicitly prevents a marginal number carrying a strategic decision.
The decision that is revisited
Outsourcing arrangements are usually assessed once, at the start, and then renewed. The arithmetic that justified them is rarely re-run three years later with actual figures.
Where it is, the common finding is that the retained cost was larger than modelled and the supplier's price has risen faster than internal pay. Both are predictable and neither is in the original case.
The year-three price
Suppliers price to win and reprice to renew. A comparison built on the first-year price is comparing a promotional rate to a permanent cost.
Where the contract has an index or a review mechanism, modelling the third-year price alongside the first is a ten-minute addition that occasionally reverses the decision.
What to keep if the work goes
Enough capability to specify, to judge quality, and to bring it back. That retained capability is a cost of outsourcing and belongs in the "costs that start" column.
Omitted, it is discovered later as an inability to manage the supplier, which is the most common way these arrangements disappoint.
The comparison done both ways
Worth running the reverse case as a check: if this work were already outsourced, would we bring it in? If the answer to both questions is no, one of the two calculations is wrong.
It is a quick consistency test and it catches most of the errors in this note, because the two arithmetics should be mirror images.
What to publish with the decision
The two columns — stops and starts — with each line named and sourced, and a statement of what is assumed about redeployment.
It is one page, and it is the page that gets looked at again in three years when somebody asks how this was decided.