The conversation usually starts the same way. A founder, somewhere between forty and four hundred staff, has a support queue that has become a problem. Response times are slipping, the team is three people who are all doing other jobs as well, and somebody has calculated that hiring four more would cost roughly what the last funding round allocated to engineering.
So they look at outsourcing, and they look at it as a cost decision.
It is not, at this stage. At twenty seats a British scale-up is not buying labour arbitrage — the saving on that headcount is real but not transformative. It is buying something the business cannot assemble internally at speed: a functioning service operation, with supervision, planning, quality and coverage already attached.
Understanding that changes what you should be evaluating, and it is why advisers who match smaller programmes to suitable providers ask what is broken before they ask what the budget is.
What you are actually buying
Mostly not people, which is the part founders find counter-intuitive.
Time
Doing it yourself means advertising, interviewing, inducting, drafting escalation rules, procuring tooling and locating somebody who has run a service desk before. Two to three months if nothing goes wrong and every candidate accepts. A provider holding trained staff already shortens that dramatically, and when the queue is visibly failing that compression is the thing being purchased.
Coverage
A British team of four cannot cover extended hours, weekends and holidays without either overtime or gaps. A provider can, because the arithmetic works at their scale rather than yours.
Management
You are buying team leaders, a planner, a quality function and an operations manager — roles no scale-up at this size has any business hiring individually.
Elasticity
The ability to add six seats in March and remove them in July without a redundancy process.
“Founders come to us with a cost spreadsheet and leave with a timeline,” said Ralf Ellspermann, founder and chief strategy officer at PITON-Global. “The question that actually matters to a scale-up is not what a seat costs. It is whether the queue is still broken in six weeks.”
How small is too small?
There is a size below which outsourcing is usually the wrong answer, and it is worth being honest about where it sits.
Below roughly eight to ten seats, the overhead of managing an external relationship — the governance, the knowledge transfer, the quality oversight — tends to exceed the benefit, unless the requirement is specifically for coverage the business cannot otherwise provide. UK analysis of why UK businesses look offshore for customer service points the same way: at that size the overhead rarely repays itself.
“Below ten seats you are buying a management relationship you have nobody to manage,” said John Maczynski, chief executive of PITON-Global. “Above thirty you are buying a function. In between, what you are really buying is a decision about which of the two you want.”
Above that, the arithmetic shifts, and it shifts decisively somewhere around twenty to thirty seats, where the internal alternative starts requiring its own management layer.
The other threshold is maturity rather than size. A business whose processes live entirely in the heads of three long-serving staff will struggle, because knowledge transfer has nothing to transfer. That is not a reason to avoid outsourcing. It is a reason to spend three weeks writing things down first, which is valuable work the business should have done anyway.
Ask about authority, not about rate
Not “what is your hourly rate?” but “what will your team be allowed to decide?”
A team that must seek permission before refunding, crediting, extending a trial or escalating a bug is not really a support team. It is a relay station, and customers work that out quickly — what they experience is a business that has to ask itself for approval.
This is the single most common failure in first-time outsourcing arrangements, and it is entirely self-inflicted. The work is transferred and the authority is withheld, usually because nobody got around to writing down what the team may settle.
“Settle what the team may decide before launch and the programme works,” said John Maczynski, chief executive of PITON-Global. “Leave it and somebody invents an answer at two in the morning during your first real incident, and you will not like what they chose.”
Choosing a provider at this size
Scale-ups are a difficult segment to buy for, because the large providers are not interested and the small ones vary enormously.
Three criteria separate the viable from the plausible.
Existing programmes that resemble yours. Not sector experience in the abstract, but live programmes now, at comparable size, with similar contact types. A provider for whom you would be the first client of your kind will learn at your expense.
A named team rather than a pool. At twenty seats you need the same people accumulating knowledge about your product, not a rotating allocation from a general floor.
Willingness to start small. A provider insisting on a minimum commitment well above your actual requirement is solving their problem, not yours.
The real shortlist is shorter than the market appears to offer: most firms claiming capability at this size are configured for something considerably larger, and will run your programme as a rounding error on a floor built for somebody else.
The first ninety days
Resist the temptation to go live at full volume because the queue is painful.
Give it a month of preparation before anything touches a customer. Then a month of live work at deliberately low volume with every interaction read. Then a month of stepping volume up, gated on quality rather than on the calendar.
It feels slow when the queue is already failing. It is substantially faster than the alternative, which is going live early, generating a quality problem in week three, and spending the following quarter repairing confidence rather than handling volume.
Success, at this stage
Not a cost line that has fallen, though it probably has.
It is that the founder stops thinking about the support queue. The function runs, somebody competent owns it, the metrics are visible, and the attention that was going into firefighting is going into the product instead.
That is what the decision was actually about, and it is worth saying out loud at the start — because a business that evaluates this purely on rate will select a cheap provider and discover, three months later, that it bought the wrong thing entirely.
