A full-time customer support advisor on the 2026 National Living Wage costs a UK employer roughly £35,000 to £41,000 a year fully loaded, once you add employer NI, pension, holiday and sickness cover, recruitment, ramp, tooling and a share of management overhead. Salary is only about 60% of that. Most in-house cost models stop at salary. That is why most in-house cost models are wrong.
The short answer: the break-even point, stated up front
The break-even sits at roughly 1.5 full-time equivalents (FTE) of ticket demand. Below that, hire. Between 1.5 and 6 FTE, especially if you need any evening or weekend cover, outsourcing usually wins. Above about 15 FTE, some flavour of hybrid tends to beat either on its own. Everything from here is the working.
What changed for UK employers in April 2026 (and what changes in January 2027)
Eighteen months of statutory change has landed on UK employers, most of it flowing from the Employment Rights Act 2025. Support functions caught it worst. If you run a team that skews entry-level, part-time and high-turnover (which most support teams do), nearly every clause was written with your people in mind.
| Change | Effective Date |
|---|---|
|
NLW (21+) up to £12.71/hr from £12.21 (4.1% rise) 18–20 rate up 8.5% to £10.85 16–17 / apprentice up 6% to £8.00 | 1 April 2026 |
| Statutory Sick Pay payable from day one (no waiting period, no lower earnings threshold) | 6 April 2026 |
| Paternity and unpaid parental leave now day-one rights | 6 April 2026 |
| Collective redundancy protective award doubled from 90 to 180 days’ pay | 6 April 2026 |
| Employer NI at 15% (from 13.8%); threshold cut to £5,000 (from £9,100), frozen to April 2028 | Since April 2025 |
| Employment Allowance doubled to £10,500; £100,000 eligibility cap removed | Since April 2025 |
| Unfair dismissal qualifying service reduced from 2 years to 6 months; compensatory cap abolished |
Planned: 1 January 2027 (Subject to confirmation) |
Sources: National Minimum Wage (Amendment) Regulations 2026 (legislation.gov.uk); Employment Rights Act
2025 (legislation.gov.uk); Rates and thresholds for employers 2026 to 2027 (gov.uk); Employment Allowance
guidance (gov.uk).
Two of these matter more than the rest for support teams.
Day-one SSP quietly removes a cost buffer that shift-based operations have leaned on for decades. Under the old regime you paid nothing for the first three working days of any absence. Now you pay from hour one, with no lower earnings threshold. On a team where minor illness turnover runs high, that adds up faster than most rota planners have modelled.
The January 2027 unfair dismissal change is bigger than it sounds. Today, a new hire can be exited with limited claim risk for the first two years. From January 2027 that safe window shrinks to six months, and the compensatory award cap goes altogether. For a role with the attrition profile support carries, that changes the risk on every offer letter. Worth stressing the caveat: the ERA is being staged in through secondary regulations, and the exact commencement date is worth checking before you commit to a hiring plan built around it.
One thing pulls in the other direction. The Employment Allowance was doubled to £10,500 and the £100,000 eligibility cap dropped. For a smaller employer this can wipe out most of the total employer NI bill. Ignore it and your build-vs-buy model overstates the case for outsourcing; your FD will notice within about ten seconds.
The salary is about 60% of it: building the fully loaded cost
Start with the raw numbers. A support advisor on the 2026 NLW (£12.71 an hour), full-time at 37.5 hours a week across 52 weeks, earns £24,785 gross.
That is what putting one person on payroll costs before they touch a ticket. If you qualify for the full Employment Allowance and have not already absorbed it against other staff, you can knock up to £2,968 off the NI line for the first advisor you add. Whether you actually get that offset depends on your total NI bill and your existing headcount. It is not a discount you can bank on for every hire.
The costs almost everyone forgets: recruitment, ramp, cover, attrition, management
Then layer the costs most spreadsheets skip.
| Line | Typical annual cost (per FTE) |
|---|---|
| Recruitment (agency fees, ads, hiring manager time) | £1,500 – £3,000 |
| Ramp to competence (6–10 weeks at ~50% output) | £2,000 – £3,500 |
| Sickness cover (higher exposure under day-one SSP) | £800 – £1,500 |
| Holiday cover for 28 statutory days | £1,500 – £2,500 |
| Attrition replacement (30–50% turnover, annualised) | £1,500 – £3,000 |
| Tooling seats (helpdesk, phone, QA, workforce management) | £600 – £1,200 |
| Share of team lead salary (1:8 management ratio) | £4,000 – £6,000 |
Combine everything and one fully loaded UK support advisor lands at £35,000 to £41,000 a year. You can push it lower with the right set-up: small employer, Employment Allowance eating most of the NI, a thin tooling stack, no dedicated team lead yet. That gets you down to about £30,000. Push the other way with London weighting, richer tooling and a dedicated QA function, and £45,000
arrives without much effort.
Either way, the £24,785 salary line is around 60% of the real figure. If your build-vs-buy model stops at salary, you have understated the in-house side by a fifth to a third.
There is one further cost that is hard to price in a spreadsheet: the wrong-hire premium. From January 2027, a poor hire caught in month five costs materially more to exit than the same mistake today. That will not show up in your cost model. It will show up in your FD’s face when you explain it.
What outsourced support actually costs in the UK: the three pricing models explained
UK outsourced support prices out three ways in practice.
Dedicated agent (UK onshore)
A named advisor working your queue, often dressed as part of your team. Typical rates run £35 to £45 an hour, or £42,000 to £55,000 a year equivalent once you include the vendor’s margin, management, cover and QA. You are paying for capacity rather than tickets, so under-utilisation is your problem, not theirs.
Shared per-ticket
A pool of agents handling multiple clients, priced per resolved contact. Tier-1 typically runs £3.50 to £5.00 per ticket, sometimes lower at volume. You pay for what you use. On the other hand, you are one of many voices in the agent’s day, so knowledge depth on your product is thinner and consistency of tone is harder to hold.
Worked example A: a 12-person SaaS business adding tier-1 support
Twelve-person SaaS, product-led, roughly 4,500 contacts a year. Call it one FTE of demand, all business hours. Technical product. Founders still touching escalations personally.
- In-house: one advisor at £38,000 fully loaded, cost per contact around £8.50 in year one (higher during ramp).
- Outsourced dedicated (UK onshore): about £42,000 equivalent, cost per contact £9.30.
- Outsourced shared per-ticket: 4,500 × £3.50 = £15,750/yr. Cheaper by half on paper. In practice it only works when your tier-1 is genuinely generic, which for a technical SaaS it usually is not.
Handoff friction to your engineers eats the saving and then some. In-house wins. The premium buys product proximity, and product proximity is what your customers are paying for. Revisit at 2 FTE of demand.
AI plus human hybrid
AI resolves the deflectable tier (password resets, order status, common FAQs) and hands the rest to human agents. Typical: a platform fee of £500 to £2,000 a month plus £1 to £3 per human-handled contact. Cheapest per contact at volume once it is bedded in. The setup is heavier than the other two options, and the AI needs training on your actual content, which most vendors quietly assume is your job.
A fourth option worth naming: nearshore and offshore dedicated. Rates drop meaningfully; £15 to
£25 an hour is realistic, and quality has improved sharply in the last five years. A lot of UK SMEs still keep tier-1 onshore for brand-voice reasons, though nearshore works well for tier-2 and back-office work.
The comparison that matters most for the decision is not just the hourly or per-ticket rate. Three other dimensions:
Time to live. In-house takes six to ten weeks from offer to competent. Dedicated outsourced: two to three. Shared per-ticket: one to two. AI hybrid: three to six, most of which is content training.
Contract flex. In-house carries statutory notice and, from January 2027, unfair dismissal protection kicking in at six months. Dedicated outsourced contracts typically run 30 to 90 days notice. Shared per-ticket and AI hybrid are usually month-to-month.
Out-of-hours loading. In-house evenings, weekends and bank holidays cost 20% to 33% in shift premiums, plus the rota overhead of building a compliant pattern for two or three people. Outsourced dedicated typically adds 15% to 25%. Shared per-ticket and AI hybrid often include out-of-hours in the rate, or add 10% to 15%.
Rates are indicative UK market as of Q3 2026. Pull two or three real quotes against your actual ticket profile before committing. Written quotes will surprise you in both directions.
Cost per ticket vs cost per FTE: which metric should drive the decision
In-house managers measure headcount. Outsourcers quote per ticket. The two numbers cannot be compared until you convert.
The conversion is simple. Take your fully loaded annual FTE cost and divide by contacts resolved per year. That gives you your true internal cost per resolved contact. Do that once and every outsourced quote suddenly makes sense.
A competent tier-1 UK advisor resolves somewhere between 4,000 and 7,000 contacts a year. The spread is mostly about ticket complexity and tooling quality, not agent talent. Against a fully loaded £38,000, that puts internal cost per resolved contact at £5.40 to £9.50 for a well-run team. If you are above £10, you have a productivity problem, not a pricing problem.
If you do not yet know your resolved-contacts-per-FTE, stop reading and go measure it. Nothing
else in this piece works without that number.
Worked example B: an ecommerce brand covering Q4 peak
Consumer brand, baseline 4,500 contacts a year, tripling to roughly 13,500 across November and December. This one is worth walking through in prose, because it is the shape that catches most ecommerce teams out.
Run it all in-house and you need three FTE to hit peak. That is £114,000 fully loaded. For ten months of the year, two of those three sit on their hands or get pushed onto low-value work. You have bought Q4 capacity twelve months a year.
Run it as a hybrid instead: one permanent in-house advisor for the baseline, plus shared per-ticket overflow for the extra 9,000 contacts in peak. That is £38,000 + (9,000 × £3.50) = £69,500 all in. Same coverage. Saving of roughly £44,500 a year. And you have bought the flexibility to expand or contract volume without a redundancy conversation in January.
The permanent hire matters, though. Pure per-ticket for the baseline usually degrades brand voice over time. Everything above baseline is the right kind of work to hand out.
Worked example C: a 60-person firm adding out-of-hours cover
Sixty-person B2B firm. Two in-house day advisors already on the books at £76,000 combined. A new enterprise contract now requires 24/7 P1 response, and the overnight volume works out at roughly 2,500 contacts a year.
- Cover the nights in-house: two more advisors at a 25% shift premium, roughly £95,000, plus rota headroom for holidays and sickness, plus the lone-worker supervision requirements you will need to work through with HR. All in, the combined day-and-night bill lands north of £176,000.
- Cover the nights outsourced: shared per-ticket at £4.00 works out at £10,000, or a dedicated night pod at £30,000 to £40,000. Combined bill: £86,000 to £116,000.
Days in-house, nights outsourced. It is not close.
When hiring in-house is genuinely the right call
There are situations where keeping everything in-house is the right answer. Not many, but they exist.
The first is product complexity, specifically the kind where months of accumulated context noticeably improve resolution quality and no amount of knowledge-base documentation closes the gap. If your escalation rate to an outsourced tier-1 would run above about 40%, you are not really outsourcing; you are paying twice.
The second is brand voice as core product. Some consumer brands live and die on tone. If your customers would notice within a message or two that they were no longer talking to your own people, in-house wins by default.
The rest are more prosaic. Stable, predictable, business-hours-only volume. Total demand that sits below 1.5 FTE with no realistic growth path above it. A regulatory regime (certain regulated finance, health or defence work) where outsourced data access is either impossible in practice or so expensive it stops making sense.
Everywhere else, the flex of an outsourced contract tends to pay for itself. That covers spiky volume, out-of-hours needs, multilingual coverage, uncertain future headcount and seasonal peaks. The January 2027 unfair dismissal change tilts things further, because a hire made under the old rules and reviewed under the new ones carries a different risk profile in the first year.
Your decision framework: seven questions and a calculator
Seven questions to answer honestly before you commit either way.
- Fully loaded cost per advisor, not the salary.
- Resolved contacts per FTE per year (measured, not guessed).
- Demand shape: flat, spiky, or seasonal?
- Any evenings, weekends or bank holidays required in coverage?
- Time-to-competence in your specific product, in weeks.
- Realistic annual attrition rate, from your last three years of data.
- From January 2027, is your cost of a wrong hire in month five higher than it is today?
If you can answer the first two accurately, you are already in the top quartile of teams making this decision. If you cannot, that is the work before the decision, not after.
