9 December 20257 min read
An agency wins a project, buys delivery from a partner at a lower rate, and books the difference as margin. Six months later the project is profitable on the spreadsheet and everyone involved is exhausted, because the spreadsheet counted the rate difference and not the three people who spent half their week making the arrangement work.
The mistake is treating the markup as profit rather than as a budget that has to cover a specific list of costs. Get the list right and the number follows.
What the margin has to pay for
Before arguing about percentages, write down what the difference between your sell rate and your buy rate is actually funding.
- Management: the project manager or delivery lead who directs the partner, runs the ceremonies and handles the client. This is the largest line and the most consistently forgotten.
- Review: the time your own senior people spend reading work before it reaches the client. Not optional, and not free.
- Rework: the share of work that comes back. On a mature partnership this is small; in the first months it is not.
- Sales and account cost: you won this, and you will spend time keeping it.
- Onboarding: the ramp before a new partner person is productive, which you pay for even when the client does not.
- Risk: fixed-price overrun, a person leaving mid-project, a currency movement on a cross-border rate, payment terms where you pay the partner before the client pays you.
- Actual profit, which is what is left, and should be a real number rather than a rounding error.
Anyone who has run this list once stops being surprised that a small markup produces no profit. The rate difference was never the margin. It was the gross figure from which seven things are deducted.
Why thin margins fail in a specific way
A thin margin does not simply reduce profit. It changes behaviour, and always in the same direction.
The first thing cut is review, because review is invisible to the client until it is missing. Then management attention thins, because the delivery lead is put on three engagements instead of one. Then the partner notices they are being managed loosely and fills the gap with their own judgement, which may be fine or may not, and either way is no longer your standard. Then quality varies, the client complains, and the rework that follows consumes what was left of the margin.
A margin too thin to fund review is a margin that will be spent on rework instead, at a worse time and in front of the client.
The same logic applies at the other end. Squeezing a partner's rate below what lets them keep good people is a false economy with a delay built in. You will not feel it this quarter. You will feel it when their strongest engineer leaves and the replacement takes two months to become useful, on your project, at your cost.
Cost-plus, rate card, or outcome
Three pricing structures show up, and they suit different arrangements.
Cost-plus, where you disclose the partner's rate and add an agreed management fee, is transparent and works with sophisticated clients who are effectively buying your delivery management. It also invites the client to wonder what they are paying you for, so it needs your management value to be genuinely visible.
A blended rate card, where you quote one number regardless of who does the work, is the common shape. It protects the partner's rate from view, lets you move people without repricing, and puts the burden on you to keep the blend honest as the mix of senior and junior time changes.
Outcome pricing, where you sell a defined deliverable, holds the largest upside and the largest risk. It only works when the scope is genuinely specified and the acceptance criteria are written, because every ambiguity is now yours to pay for. If you sell an outcome, buy an outcome: match a fixed-price sale with a fixed-price statement of work from the partner rather than buying hourly capacity against a fixed commitment.
The terms that quietly move the margin
Two commercial details affect profitability as much as the headline rate.
Payment timing. If you pay the partner on thirty days and the client pays you on sixty, you are financing the engagement, and at scale that is a real cost with a real interest rate. Either align the terms or price the gap.
Currency and rate stability. On cross-border arrangements, agree which currency the rate is denominated in and how long it is fixed for. A rate that is renegotiated mid-engagement destroys a fixed-price quote, and a rate fixed forever will eventually be one the partner cannot deliver on. An annual review with a stated notice period is the usual settlement.
Volume commitments are worth understanding in both directions. A guaranteed monthly minimum should buy you a better rate, because it lets the partner staff against certainty. Do not commit to a minimum you cannot fill, though: paying for unused capacity is precisely the bench cost you subcontracted to avoid.
Set the number, then check it against reality
The practical method is not to pick a percentage from an industry rule of thumb. It is to price the seven cost lines for the specific engagement, add the profit you actually want, and see what sell rate that implies. Then check it against the market: if the implied rate is not sellable, the problem is either that the work is not worth what it costs to deliver well, or that you are buying the wrong shape of capacity. Both are worth knowing before the contract, not after the second invoice.