Deciding whether to make, buy or partner for a service
Score a support function on cost, control, capability and switching risk, then decide. The research basis is thinner than the brief assumed, and this piece says so.
Most support functions are outsourced for the wrong reason or kept in-house out of habit. The usual trigger is a supplier quote that looks cheaper than the team's payroll. The decision then runs on that single number until a contract renewal or a service failure forces a second look. The pattern is not confined to companies. According to the European Commission, 55% of procurement procedures in the EU use lowest price as their only award criterion. Buyers pick the number that is easiest to compare, and the cost of that choice tends to appear later.
This article sets out a framework for support services such as finance operations, payroll, HR administration, IT service desks and facilities. It asks four questions of each service: what it costs in full, how much control you must keep, how much capability you already have, and how hard it would be to leave. The brief behind this piece assumed a body of business-school research on outsourcing. Our source pack did not contain one. Harvard Business School's website describes its programmes and case-based teaching, but none of the sources we checked tests make-or-buy decisions. The tests below are therefore editorial judgement, informed by how public buyers are asked to think about procurement. Treat them as a structure for your own evidence, not as a finding.
Chapter one: Cost: price the whole arrangement, not the invoice
Start with the all-in annual cost of each option. For the in-house version, that means salaries and on-costs, but also the hours your managers spend supervising the work, the software licences, and the space it occupies. For the outside version, add the time needed to write and monitor the contract, the charges for change requests, and any fees for ending the arrangement early. The test is simple. If the supplier's price is lower only when you leave out the management time, the saving is not yet real.
What usually goes wrong is a comparison of year one against a steady-state internal cost. The switching charges and the extra supervision arrive in year three, when nobody is comparing. Ask finance to build a five-year view for both options and to show the assumptions in the same table. Score cost from 1 to 5, where 5 means the outside option is clearly cheaper once every line is counted and 1 means it is clearly dearer. Anything in the middle should not be decided on cost alone.
The Commission's reading of public buying is that buyers are probably not paying enough attention to quality, sustainability and innovation. The same error appears in private buying. A bid that wins on price can lose on the service you actually receive, and a cost comparison that ignores service quality is only half a comparison. Ask each bidder to describe the service levels it will meet and what it will charge if it misses them. The answer tells you more than the headline rate.
Chapter two: Control: decide what you must be able to direct
Control is the ability to change how a service is delivered, on your timetable, without asking a third party's permission. It is not the same as ownership. A supplier can run a payroll run on your behalf while you still set the pay calendar, approve exceptions and decide who has access to the data. The useful question is narrower: which three service levels would hurt you most if they slipped for a week? Write them down, then check whether your contract and your reporting let you direct the supplier on each one.
Score control from 1 to 5, where 5 means you must hold direct control over delivery. The Commission identifies principal public buyers in energy, transport, waste management, social protection, health and education. In each of those sectors, a failed service reaches people outside the buying body. Private firms rarely work at that scale, but the logic carries over. If a failure would reach your customers or your regulator within days, give control a high score, and be cautious about any option that leaves you unable to act on that failure yourself.
Chapter three: Capability: can you run what you buy?
Buying a service does not remove the need to manage it. The European Commission's recommendation on the professionalisation of public procurement, adopted in October 2017, says buyers need skills, knowledge and integrity. It points to training, career management, e-procurement tools and templates as the means of building them. The Commission backs the recommendation with a library of 90 examples of good practice and tools. The point for a private leader is that managing a supplier is a trade, and the trade has to be present in the building.
The test is practical. Could someone on your team write the specification, judge the supplier's monthly report and spot a missed service level without asking the supplier to explain it? If yes, score capability high, and the buy option becomes safer. If no, the partner option may be the honest answer. A partner can supply skills you lack while your own people learn to direct the work. Score from 1 to 5, where 5 means the skill already sits in your team. A low score is not a reason to avoid the decision. It is a reason to name the gap and price the cost of closing it.
Chapter four: Switching risk: price the exit before you sign
Switching risk is how hard, slow and expensive it is to leave. The Commission's 2025 evaluation of the 2014 EU public procurement directives found that the directives had only partly achieved their objectives. Among the problems it named were increased complexity, lack of buyer flexibility and weaker tender competition. Those features also make a contract hard to leave. A buyer tied to one provider has little room to change course when the market moves, and a thin market leaves little to switch to.
Ask three questions. How long would it take to move this service elsewhere? What would the move cost, including extracting your data and running two services in parallel? Do at least two credible providers exist? Score switching risk from 1 to 5, where 5 means you would be stuck. The National Bureau of Economic Research describes Working Paper 34755, Learning to Quit?, by Esther Bailey, Daniel Fehder, Eric Floyd, Yael Hochberg and Daniel J. Lee, a randomised field experiment with innovation-driven entrepreneurs. Its description says entrepreneurship training may help founders recognise when a venture no longer warrants more time and resources. The lesson for a buyer is a habit rather than a result: write the conditions for ending the arrangement before the contract is signed, while everyone is still friendly.
To turn the four scores into a decision, use our judgement rather than a formula. Make the service in-house where control and capability both score 4 or 5 and switching risk is low enough that you can afford to be wrong. Buy where control scores 2 or lower, the all-in cost gap is clear, and at least two providers exist. Partner where control is high but capability is thin, or where switching risk is high enough that a supplier must share the commitment with you. A set of middling scores is a sign to define the arrangement in more detail, not a signal to partner by default. Partnering does not avoid the decision. It only moves it into the contract.
Answer yes or no before you choose
If you answer yes to most of the questions in a row, the framework points the same way. If the answers split, write down the disagreement before you decide.
- Would a failure of this service reach customers or a regulator within a week?
- Have we written down the three service levels that matter most?
- Could someone on our team write the specification and review the supplier's reports?
- Does the all-in cost, including management time and exit fees, still favour the supplier o
- Could we move this service to another provider within a notice period we can live with?
- Are there at least two credible providers in this market?
- Have we written down the conditions under which we would end the arrangement?
Outgrace checklist
Where it breaks
The framework fails in three places. The first is when the cheapest option is also the only option. A market with one credible bidder cannot be scored on cost, and the Commission's evaluation named weaker tender competition among the problems it found. The second is when the service carries something that makes your product or your reputation different. Public rules allow buyers to weigh environmental, social and innovation aspects alongside price, and the Commission plans to update its guidance on those criteria. A private firm has the same reason to keep close a service that shapes its quality. The third is when the rules change. On 9 September 2026 the Commission adopted a proposal for a Public Procurement Act, which aims at simpler, more flexible and more digital procurement. It is a proposal, and it does not yet change the rules. If your buyer is a public body, check which regime applies. EU rules apply above a monetary threshold and national rules apply to lower-value tenders. That question belongs to your procurement lawyer, not to this scorecard.
Improving public procurement can yield big savings: even a 1% efficiency gain could save €20 billion per year.
On Monday, pick one support service that takes the most management attention. Score it on the four tests and work through the checklist. Get the all-in cost from finance in writing, not from the summary of a bid. Ask the person who would run the service to write the control and exit conditions in plain language, in no more than a page. If the result points to partnering, put the division of direction into the contract before you sign. If the buyer is a public body, or the contract touches regulated data, put the tender and legal questions to your lawyers before you approach any supplier.

