A merit matrix is a table used in a pay review to recommend a salary-increase percentage from two inputs: a calibrated performance rating, and the employee’s position in their pay range. This page is the how-to for People leaders at a 200–2,000-person scale-up: what the grid is, how a merit increase differs from a promotion or a market adjustment, and a worked example in pounds sterling. It is not a guide to UK public-sector pay-review bodies.
Use it with the essay on whether performance reviews should be linked to pay, the performance management framework — which treats pay as one outcome of a calibrated grade — and the calibration meeting guide. The Excel workbook is free with your work email; the same grid and arithmetic also appear on this page. Disclosure: Ash Taherian is Product Sales Executive at Revolut People, which sells performance management software. Treat product mentions as coming from an interested party.
What is a merit matrix?
A merit matrix is a table that recommends a salary-increase percentage from two inputs: a performance rating, and the employee’s position in their pay range. Position in range is usually a compa-ratio — current base pay divided by the range midpoint — or range penetration, which is (pay minus the range minimum) divided by (the range maximum minus the minimum). The number in each cell is a recommended merit increase: a lasting rise in base pay given for performance in the current role, typically from a budget set aside for that purpose. It is a guideline, not an automatic payment, and it is not a promotion or a market adjustment.
Merit increase vs a raise, a promotion and a market adjustment
“Raise” is the umbrella word. A merit increase is one kind of raise: it pays for how someone performed in the role they already hold. A promotion pays for a different, higher role, and should come from a separate budget — this page is not a promotion guide; see when promotion decisions feel political if that is the problem you are solving. A market adjustment, or a cost-of-living rise, corrects the range itself or inflation; it is not a judgement on the individual’s last rating.
Mixing these into one pot is a common failure mode. The merit budget is then asked to fund promotions, to catch up people who sit below the market, and to keep pace with prices, as well as to differentiate performance. When that happens, high performers who are already well placed in range often wait, and managers start trading favours in the review. Keep the pots separate, and say so when you brief managers.
What you need before you use a matrix
Three things have to be true, or a tight matrix will reward the most generous managers rather than the strongest performance.
Calibrated ratings. Ratings must have been compared across managers before you apply the grid. The calibration meeting guide is the step immediately before the matrix: check for generous graders, thin evidence and outliers, and only then lock the grades. The performance calibration guide covers the cycle around that meeting.
A scale you can defend. Use the same labels the matrix columns will use, with written definitions. The playbook’s five-level scale — Poor, Basic, Intermediate, Advanced and Exceptional — is set out in employee performance rating scales. The worked example below uses those five levels. Many organisations label Advanced as “Exceeds expectations” and Intermediate as “Meets expectations”; the mechanism is the same.
A written pay range per role. You need a minimum and a maximum at least, and a midpoint if you have one, so position in range is not a guess. If you only have a minimum and a maximum, you can still use range penetration. If you have no range at all, do not run a tight matrix until you do.
If those conditions are not in place, say so plainly and loosen the formula, or wait a cycle. The essay on whether performance reviews should be linked to pay sets out three models — a formula, informed judgement, and a looser link. Before you attach pay tightly to a grade, run the performance health check and the review bias checker.
How a merit matrix works
Build a grid. The columns are the five rating levels. The rows are bands of position in range — for example below, around and above the midpoint. Each cell holds a recommended percentage, not an automatic payment. The pay range itself hangs off a published job level; how those levels are defined is the job architecture and levelling framework.
This page uses compa-ratio: current base pay divided by the range midpoint. A ratio of 1.00 is the midpoint; 0.84 is 16% below it; 1.12 is 12% above it. Range penetration is the alternative: (current base pay minus the range minimum) divided by (the range maximum minus the minimum). Both are valid. Pick one definition and use it consistently in the cycle.
Higher performance and a lower position in range should generally mean a higher recommended percentage, so the grid moves people towards the range rather than pushing already-high pay further out.
Treat every cell as a guideline. If you change a recommendation, record a one-line written reason in an exception log. Unrecorded adjustments are how a fair-looking grid becomes an unfair outcome.
Worked example in GBP (illustrative only)
Illustrative only. The percentages, salaries and people below are a planning illustration. They are not a market rate, not a 2026 or 2027 average, and not a Revolut People default. Do not copy them into your cycle as if they were a benchmark. The company, the band and the employees are fictional.
Fictional setting: Harbour Ledger, a 600-person UK scale-up. Senior Analyst band: minimum £40,000, midpoint £50,000, maximum £60,000. Position in range is the compa-ratio (current base pay ÷ £50,000).
| Position in range (compa-ratio) | Poor | Basic | Intermediate | Advanced | Exceptional |
|---|---|---|---|---|---|
| Below midpoint (under 0.95) | 0% | 2.0% | 3.5% | 5.0% | 7.0% |
| Around midpoint (0.95 to 1.05) | 0% | 1.0% | 3.0% | 4.0% | 5.5% |
| Above midpoint (over 1.05) | 0% | 0% | 2.5% | 4.0% | 5.0% |
| Employee (fictional) | Current base | Compa-ratio | Calibrated rating | Recommended % | Increase | New base |
|---|---|---|---|---|---|---|
| Amira Khan | £42,000 | 0.84 | Advanced | 5.0% | £2,100 | £44,100 |
| Ben Cole | £56,000 | 1.12 | Advanced | 4.0% | £2,240 | £58,240 |
| Chloe Reid | £56,000 | 1.12 | Intermediate | 2.5% | £1,400 | £57,400 |
Amira and Ben have the same calibrated rating, Advanced. Amira sits well below the midpoint (compa-ratio 0.84), so the matrix recommends 5.0%, or £2,100, taking her to £44,100. Ben already sits above the midpoint (compa-ratio 1.12), so the same rating recommends 4.0%, or £2,240, taking him to £58,240. Amira’s percentage is higher, which is the point of the grid: it closes more of her gap to the range. Ben still receives more pounds, because his starting salary is higher. The matrix recommends a fair percentage; it does not try to equalise the cash amount.
Chloe holds the same position in range as Ben, with an Intermediate rating, so her recommended increase is 2.5%, or £1,400. The difference between Ben and Chloe is the rating; the difference between Amira and Ben is position in range.
Download the merit matrix template
The Excel workbook has a blank matrix, a compa-ratio and range-penetration calculator, a budget-check sheet, an exception log, and the fictional Harbour Ledger worked example from this page. It is free with a work email.
Building the matrix against a budget
Cell values are not copied from another company’s grid, and they are not a published “typical” budget. Set them so that, applied to your actual population and calibrated ratings, the recommendations fit the merit pot. Keep promotions and market corrections out of this pot. The check below uses a fictional team of eight Senior Analysts in the same Harbour Ledger band. Salaries, ratings and percentages are illustrative only.
| Employee (fictional) | Current base | Rating | Position in range | Recommended % | Increase |
|---|---|---|---|---|---|
| Amira Khan | £42,000 | Advanced | Below midpoint | 5.0% | £2,100 |
| Ben Cole | £56,000 | Advanced | Above midpoint | 4.0% | £2,240 |
| Chloe Reid | £56,000 | Intermediate | Above midpoint | 2.5% | £1,400 |
| Dylan Patel | £48,000 | Intermediate | Around midpoint | 3.0% | £1,440 |
| Elena Walsh | £50,000 | Intermediate | Around midpoint | 3.0% | £1,500 |
| Farid Osman | £44,000 | Basic | Below midpoint | 2.0% | £880 |
| Grace Adeyemi | £52,000 | Exceptional | Around midpoint | 5.5% | £2,860 |
| Hassan Rahman | £40,000 | Poor | Below midpoint | 0% | £0 |
| Total | £388,000 | 3.20% | £12,420 |
Total base pay is £388,000. The recommended increases sum to £12,420, or 3.20% of that payroll. If the merit pot for this band is 3.0% of current base pay, or £11,640, the first pass overruns by £780. Adjust the cells and re-run; do not quietly shave individuals at the end without a recorded reason.
One illustrative re-run: reduce Exceptional / around midpoint from 5.5% to 4.5% (Grace’s increase falls from £2,860 to £2,340, saving £520) and Advanced / above midpoint from 4.0% to 3.5% (Ben’s increase falls from £2,240 to £1,960, saving £280). The new total is £11,620, just under the pot. Record those two cell changes as you would any other exception. If a later pass still does not fit, change the grid again or reopen the pot — do not hide the gap in one-off cuts only the People team can see.
Do not use a published “average merit increase” as a substitute for this check. Those figures, when they are published at all, are often a total salary-increase budget, or a basic pay-award intention, and they are not a cell in your matrix.
Two named surveys, both opened on 13 November 2026, show the distinction. CIPD’s Labour Market Outlook, Summer 2026 (published 16 August 2026) finds that, among employers looking to increase, decrease or freeze pay in the next 12 months, the median expected basic pay increase remains at 3% overall, and has stayed at 3% in the private, public and voluntary sectors for the last six quarters (apart from the voluntary sector in summer 2025). That is a basic pay-award intention, not a merit-only budget; the report notes that a basic award is only one component of pay growth. The table is based on employers expecting and able to estimate a pay award (total n=824, of whom 507 were in the private sector).
WorldatWork’s 2026–2027 Salary Budget Survey — a 16 July 2026 Workspan Daily article and the public topline PDF, survey closed May 2026 — defines total salary-increase budgets as the sum of merit, general or cost-of-living, and other increase budgets. For the United Kingdom, the mean total budget was 3.8% actual in 2026 and 3.7% projected for 2027 (medians 3.6% and 3.5%). The public topline does not publish a separate merit-only figure, so do not treat 3.7% or 3.8% as “the average merit increase”.
Where the matrix sits in a pay review
The sequence is the same whether your effective date is in the spring or later in the year. Calibrate first. Apply the matrix. Run the pay review — typically line manager, then department, then reward or People. Log every adjustment to the recommendation. Check outcomes by rating, gender and level before letters go out. Hold the pay conversation separately from the development review. The existing essay’s step plan for one pay cycle is the place for the week-by-week plan; this page does not copy it.
On timing: many UK private-sector employers implement a pay award in the spring, often from April, but this is not universal and it is not a fact about scale-ups as a class. Incomes Data Research, in a September 2022 report for the Low Pay Commission on the influence of the National Living Wage on pay setting (PDF opened 13 November 2026), found that most pay reviews in its sample with a fixed date took place in April (76% of 50 respondents with a fixed date). The survey was targeted predominantly at low-paying sectors and drew 56 responses. Treat that as a dated observation from that sample, not as a rule for a 200–2,000-person scale-up, and work backwards from your own effective date. This page is not about UK public-sector pay-review bodies, and it is not a legal guide to pay-transparency duties — the reviews-and-pay essay covers that ground for EU staff.
Common mistakes
- Using uncalibrated ratings, so the matrix pays the most generous managers.
- Putting promotions and market fixes through the merit pot, so the pot is asked to do four jobs.
- Treating the cell as an automatic payment, with no exception log.
- Skipping the outcome check by rating, gender and level before letters go out. See also performance review bias.
- Paying for things the scorecard did not measure.
- Copying another company’s percentages, or a published “average increase”, as if they were a benchmark for your cells.
Where Revolut People fits
If you want a second look at the grid and the outcome check, bring a rating scale and a merit matrix with names removed to a 15-minute review with Ash, Product Sales Executive. It is a fit-check, not a product walkthrough that replaces your own arithmetic.
In Revolut People, Salary Reviews let organisations run structured compensation review cycles, with a configurable approval workflow and a central place for salary recommendations. Public docs (as checked 13 November 2026) describe them as useful for annual compensation review cycles, promotion-related salary changes, market alignment reviews, and department or company-wide adjustments. Default steps on that page are Line Manager Review, Head of Department Review, and Final Review (Compensation Manager); steps and reviewer types can be configured.
The same page states: “Salary Reviews are managed independently from Performance Reviews, but can be used alongside performance outcomes to inform compensation decisions.” Importing salary-review data (employee email, proposed new salary, currency, optional justification) collects recommendations only. It does not automatically update compensation records; final salary updates must be validated and applied separately.
Compensation bands define an acceptable pay range for a combination of attributes (the docs give examples such as specialisation, seniority, sublevel, location, and band type). Each band stores a lower amount, an upper amount and a currency, plus optional fields such as a target bonus percentage. Bands store a minimum and a maximum. There is no stored midpoint or range spread. During compensation reviews, the system can find the approved salary band that matches the employee and can also retrieve the next sublevel’s band.
Those pages do not describe a built-in merit matrix, a generated merit percentage, or a calculated compa-ratio or midpoint. A rating is not written into pay automatically.
FAQ
What is a merit matrix?
A merit matrix is a table used in a pay review to recommend a salary-increase percentage from two inputs: a performance rating and the employee’s position in their pay range (often a compa-ratio, which is current base pay divided by the range midpoint). It is a guideline for a merit increase — a lasting rise in base pay for performance in the current role — not an automatic payment, and not a promotion or a market adjustment.
How do you use a merit matrix for salary increases?
Calibrate ratings first, place each person in the cell that matches their rating and position in range, and treat the percentage in that cell as a recommendation. Check that the recommendations, applied to the real population, fit the merit budget. Record a reason for any change, check outcomes by group before letters go out, and hold the pay conversation separately from the development review. The percentages in any example grid are a planning model, not a published average for 2026.
What is a merit increase, and how is it different from a raise, a promotion or a market adjustment?
A raise is the umbrella word. A merit increase is a lasting rise in base pay given for performance in the current role, typically from a budget set aside for that purpose. A promotion pays for a different, higher role and should come from a separate budget. A market adjustment or cost-of-living rise corrects the range or inflation, not the individual’s last rating. Mixing these into one pot is a common failure mode.
What are merit increases for 2026 and 2027?
There is no single published “average merit increase” to copy into a matrix. CIPD’s Labour Market Outlook, Summer 2026 (opened 13 November 2026) reports a median expected basic pay increase of 3% for the next 12 months, including in the private sector — a basic pay-award intention, not a merit-only budget. WorldatWork’s 2026–2027 public topline (opened the same day) reports UK mean total salary-increase budgets of 3.8% actual in 2026 and 3.7% projected for 2027, defined as merit plus general or cost-of-living plus other increase budgets, not a merit figure on its own. The worked example on this page is an illustrative planning model, not a benchmark. The budget section above cites both sources in full.
What do you need in place before you use a merit matrix?
Three conditions: ratings that have been calibrated across managers; a rating scale narrow enough to defend, with written definitions; and a written pay range per role (a minimum and a maximum at least, and a midpoint if you have one) so position in range is not a guess. If those are not in place, a tight matrix will reward the most generous managers.
What happens in a pay review meeting once ratings are calibrated?
On the People-team side: apply the matrix to the calibrated ratings, treat each cell as a recommendation, log any exception, and take the proposals through approval — typically line manager, then department, then reward or People. Check outcomes by rating, gender and level before letters go out, and hold the pay conversation separately from the development review. This is not a guide to how an individual should ask for a rise.