UK Salary Guide 2026: how to read a benchmark without being misled by it.
We do not publish a table of national salary averages, because a national average for a specialist UK role is a number that describes nobody. This is the more useful thing: how to build and read a benchmark you can defend.
A usable UK salary benchmark compares role content rather than job titles, against a named set of employers who genuinely compete for the same person, with total reward decomposed and London weighting reported separately from base pay.
Published national averages fail on all four counts, which is why bands built from them so often fail to clear the market.
Why we do not publish a salary table
It would get more traffic. It would also be, for any individual reader, wrong — and wrong in a way that is impossible to detect from the outside.
A published figure for “Head of Finance, UK” blends a role running a two-person team in Hull with one running a forty-person function in the City. It averages employers who responded to a survey with those who did not. It folds London weighting into a national number. It reports a midpoint without a distribution, which hides the only thing that matters: how wide the range is and where the edges sit.
Someone then sets a band from it, cannot hire, and concludes there is a skills shortage.
The five questions a usable benchmark answers
- Which employers actually compete for this person? Not your industry peers — the employers who would genuinely be the alternative offer. These are often in different sectors entirely.
- What is the distribution, not the midpoint? A median of £72,000 means something quite different in a range of £68–76k than in a range of £55–95k. The second tells you the job is defined inconsistently across the market, which is itself useful.
- What does total reward look like, component by component? Base, bonus target and actual, pension contribution, equity or LTIP, car or allowance, and the benefits that genuinely move UK candidates.
- What is the regional differential for this specific function? Not a blanket regional percentage. Pay does not scale uniformly across UK functions, and in scarce specialisms the differential has almost disappeared.
- What is London weighting worth here, separately stated? Always as its own figure, never folded into base.
How to decompose total reward properly
| Base salary | The contractual figure. The only component most surveys report, and rarely more than two-thirds of the actual package at senior level. |
|---|---|
| Bonus — target | The stated opportunity. Useful, but frequently aspirational. |
| Bonus — actual | What was genuinely paid in recent cycles. Where establishable, this is the number candidates actually price. |
| Deferral and clawback | In UK financial services, a material share of senior bonus is deferred over several years. A headline figure that ignores this is not comparable across sectors. |
| Pension | Employer contribution rate, and whether it exceeds the UK auto-enrolment minimum. A ten-point difference here outweighs several thousand pounds of base. |
| Equity, options and LTIP | Share options, growth shares, carried interest or long-term incentive plans. Decisive in scale-up and private equity settings and absent from most surveys. |
| Car or allowance | Still material in UK sales, field and senior operational roles. |
| Private medical and protection | Cover level and whether family is included. |
| Holiday entitlement | Days above the statutory minimum, and whether purchase is available. |
| Working pattern | Mandated office days. Now one of the strongest single predictors of whether a UK candidate engages at all — and not a benefit that any survey prices. |
Common benchmarking mistakes in UK organisations
Benchmarking the title instead of the job. Two roles with the same title can differ by an order of magnitude in scope. Benchmark against comparable content: team size, budget, reporting line, decision rights.
Choosing comparators by sector rather than by competition. A retailer recruiting a pricing analyst competes with gaming, insurance and utilities, not with other retailers. The comparator set should be who you lose candidates to.
Applying one regional percentage across every function. The differential varies sharply by role. In scarce technology and engineering specialisms national remote hiring has compressed it almost to nothing, while in operational roles it has held.
Using a midpoint without the distribution. The width of the range tells you how consistently the market defines the job, which frequently matters more than the midpoint itself.
Benchmarking once and treating it as permanent. A band from three years ago in a moving market is not evidence, and defending it is how vacancies become long-term vacancies.
Using a benchmark in the conversation with finance
The argument that works is not “the market pays more”. It is: here are eight named employers who compete with us for this person, here is where each of them sits, here is our current position in that distribution, and here is the observed consequence — this many approaches, this many declines, this share citing compensation.
That is a cost-of-vacancy argument supported by evidence rather than a pay-rise request supported by a survey. It is a considerably easier conversation.
What we deliver
An Audentia benchmark, itemised.
- A named comparator set, agreed with you before the research begins
- Role-content matching rather than job-title matching
- Total reward decomposed into every component above
- Distribution and range, not only the midpoint
- London weighting isolated and quantified as its own figure
- Per-function regional differentials across the UK locations you care about
- Your current position mapped inside the distribution
- A written recommendation on where the band should sit, and why
Salary benchmarking — questions
Why not just buy a published salary survey?
Because of what a survey can and cannot know. It reports the employers who responded, matches on job title rather than job content, and usually blends London weighting into a national figure. For a broad sense of direction it is fine. For setting a band you will have to defend, it is not.
What is London weighting and why isolate it?
It is the premium attached to working in and around London, reflecting cost of living and labour competition. When it is folded into a national average, the resulting figure overstates regional pay and understates London — so it misleads you in both directions at once.
How many comparators do we need?
Six to ten named employers who genuinely compete for the same person is usually enough to produce a stable picture. Thirty loosely relevant ones produce a wider range and less confidence, which feels more rigorous and is not.
Should we benchmark to the median or above it?
It depends on what you are buying. Median keeps you in the conversation; upper quartile buys speed and choice in a competitive market. The mistake is choosing a position without knowing where the distribution actually sits — which is the point of benchmarking in the first place.
How often should bands be re-benchmarked?
Annually for stable functions; every six months for anything where the market is genuinely moving. A band that has not been examined in three years is not a band, it is a historical artefact.
Need a band you can actually defend?
Tell us the role and who you think you compete with. We will tell you whether you are right about the comparator set before we quote anything.