A merit increase matrix turns an annual pay budget into a repeatable decision rule. It combines employee performance with position in the salary range and assigns a recommended raise percentage to each combination. The grid gives managers a starting point while compensation teams keep control of cost and range movement.
The hard part is building a matrix that fits the workforce you actually have. A grid can look reasonable and still miss the budget once real employees are mapped into its cells. The Conference Board reported a median US salary increase budget of 3.5% for 2026 and projected the same level for 2027, while its projected median 2027 merit budget was 3.1%, according to its 2026-2027 salary budget report. Those figures are useful context, but they aren’t a ready-made matrix.
A sound process starts with your own salary ranges and eligible payroll, backed by clear performance and pay rules. External surveys should inform those choices rather than replace them. WorldatWork’s 2026-2027 Salary Budget Survey covers 24 countries and supports cuts by factors such as industry and geography. The goal is to build a merit increase matrix that reflects your workforce and holds up when the numbers are tested.
TL;DR
A merit increase matrix usually uses performance on one axis and position in range on the other. Employees with stronger performance and lower pay relative to the range midpoint generally receive higher recommended increases. Employees with weaker performance or pay already high in the range generally receive less.
Before the grid is published, model every cell against real employee counts and salaries. The weighted cost must fit the approved merit budget. Keep promotion increases, market corrections, minimum-wage changes, and other special adjustments outside the core matrix unless your policy explicitly funds them from the same pool.
· Confirm salary ranges and eligible employee data.
· Calibrate performance ratings before pay recommendations begin.
· Set range-position bands and draft the cell values.
· Cost the grid against payroll and adjust the percentages.
· Test pay outcomes and exceptions before release. A salary increase matrix should guide decisions without hiding weak inputs.
What a Merit Increase Matrix Does
A merit increase matrix recommends a base-pay increase by crossing 2 factors: performance and position in the salary range. A person rated highly and paid below midpoint may land in a higher cell than a person with the same rating who is already near the top of the range.
That structure gives the pay cycle a common rule. Managers still need judgment for unusual cases, but the starting point is no longer an open-ended percentage. SHRM’s HR guidance describes the same basic idea in its salary increase matrix resource, where annual merit increases are tied to a structured matrix rather than handled as isolated manager requests. For compensation planning, the value is consistency across teams and a clearer path from budget to individual recommendation.
A merit increase matrix also separates 2 questions that managers often mix together. Performance answers how strongly the employee contributed during the review period. Range position answers how the employee’s current salary sits against the organization’s pay structure. Combining them helps avoid a flat raise that gives the same percentage to employees in very different pay situations.
A matrix is a policy tool. If job levels are wrong or salary ranges are stale, the grid can repeat those problems. Design work starts before anyone types percentages into a table.
Why a Matrix Belongs in Compensation Planning
A matrix helps translate a finite pool into differentiated increases. That matters because annual pay movement is shaped by more than one signal. The US Bureau of Labor Statistics reported that wages and salaries for civilian workers rose 3.2% over the 12 months ending June 2026, while private-industry wages and salaries rose 3.1%, according to the latest Employment Cost Index release. Those figures describe broad wage movement. They don’t tell an employer what each employee should receive.
Employee-level wage growth can also move differently from employer budget surveys. The Atlanta Fed’s Wage Growth Tracker showed 4.1% median wage growth in August 2026, with 3.6% for workers who stayed in the same job and 5.0% for job switchers. That gap can matter when HR is trying to retain people in roles where outside offers are moving faster than the normal annual cycle.
Inflation adds another reference point but shouldn’t be treated as a merit formula. The Bureau of Labor Statistics reported that the US Consumer Price Index rose 3.4% over the 12 months ending August 2026 in its August 2026 CPI release. A cost-of-living measure and a merit increase serve different purposes. Mixing them inside one percentage can make it hard to tell how much of a raise reflects performance and how much reflects other pay pressure.
This is where a structured salary increase matrix helps. It gives performance and range position a defined role while separate policies handle promotions and market adjustments. Finance can see what the approved pool pays for, and managers can see the boundaries.
Gather the Inputs Before You Build the Grid
Build the matrix after the inputs are clean. You need current base salary, the assigned range, job or grade, performance rating, eligibility, effective date, and known pay actions. You also need a clear definition of which employees are funded by the approved pool.
External wage data can help test whether salary structures are still credible. Regional movement can differ even when the national average looks stable. BLS publishes Employment Cost Index data by area, and the June 2026 table shows different 12-month wage changes across major metropolitan areas. That doesn’t mean every location needs a different grid, but it does mean a national midpoint can become weak if the jobs compete in sharply different labor markets.
Confirm the Salary Structure First
Check each salary range before calculating employee position. If market pay has moved while the structure hasn’t, many employees may appear underpaid against an old midpoint. That can push too many people into the most generous columns.
Look for grades with unusual concentrations below minimum or above maximum. Those patterns can point to old pay decisions or weak market references. You need to know the issue exists before you use the range in a merit increase matrix.
Define Eligibility and Payroll Scope
Set the rules for who receives a recommendation. Common questions include how to treat new hires, employees on leave, workers with recent promotions, people with active performance plans, and employees whose salaries already exceed range maximum. Decide whether part-year service changes the award or only changes eligibility.
Calculate eligible payroll using the same population finance will use. A 3% pool on eligible payroll differs from 3% of total payroll when a material group is excluded. Document the denominator before modeling the cells.
Separate Other Pay Actions
Keep merit separate from other salary actions. Promotions reflect changed job scope, while market or legal adjustments answer different pay issues. If those actions are pushed through the same pool, the final distribution becomes hard to interpret.
Your compensation planning file should carry separate fields for each action. The final salary can reflect several changes, while the merit increase matrix still shows the portion produced by the merit rule.
Set the Performance Axis
The performance axis should use the rating scale employees actually receive, but fewer levels often make the grid easier to govern. A 5-point performance system can work, yet many organizations collapse similar ratings into 4 matrix rows if the distinction between adjacent levels isn’t meaningful enough to justify a different pay outcome.
Performance quality matters more than the number of rows. Research has repeatedly shown that evaluation systems can carry bias when criteria are vague or managers apply standards differently. An analysis from Harvard Business Review on bias in performance evaluations argues for clearer criteria and more consistent evaluation practices. If ratings feed directly into salary decisions, weak rating discipline becomes a pay problem.
Calibration can reduce obvious differences between managers, but it needs rules of its own. Harvard Business Review has also examined how calibration meetings can introduce bias, especially when discussion changes ratings without a clear standard or documented evidence. For a merit increase matrix, the practical lesson is to finish rating calibration before the grid is applied and keep a record of why material rating changes were made.
A workable performance axis might include Exceptional, Exceeds Expectations, Meets Expectations, and Below Expectations. The exact labels matter less than the definitions behind them. Each level should have written criteria and enough examples for managers to distinguish one from another. If almost everyone receives the same rating, the matrix loses most of its ability to differentiate pay.
Build the Compa-Ratio Axis
Compa-ratio shows how base salary compares with the range midpoint. Divide current base salary by midpoint. A salary of $72,000 against an $80,000 midpoint produces 0.90. A salary equal to midpoint produces 1.00, while pay above midpoint produces a value above 1.00.
The measure is useful because the same performance result can call for different salary movement depending on current pay position. Someone performing strongly at 0.84 may have more room and a stronger pay-position case than someone with the same rating at 1.12. The matrix can recognize that difference without asking every manager to invent a percentage.
Your bands should fit the shape of your salary ranges and workforce. A simple design might use below 0.90, 0.90 to 0.99, 1.00 to 1.09, and 1.10 or above. Wider ranges may need a different split. If the organization uses broad salary bands, the midpoint can cover employees at very different career stages, so one fixed set of cutoffs may be too crude.
A salary increase matrix should distinguish compa-ratio from tenure. Someone can be low in range because they were recently hired or because pay has lagged. Use the ratio as a signal, then review the reason behind extreme values before approving large changes.
Performance-linked pay can also affect wage distribution. An NBER study on performance pay and wage inequality found that performance-pay jobs were more closely tied to worker characteristics and contributed to wage dispersion in the period studied. That is one reason to inspect how the matrix distributes dollars across groups and pay levels.
Before setting final bands, run a frequency table for the proposed columns. If 60% of employees land below 0.90, review the ranges or job matches before assuming the whole group needs aggressive increases.
Convert the Merit Budget Into Matrix Values
Start with the approved merit budget and work backward. If eligible payroll is $20 million and the approved rate is 3.2%, the pool is $640,000. That is the amount the matrix can allocate unless finance has approved a separate reserve.
Next, map every eligible employee into a performance row and range-position column. For each cell, calculate both headcount and current eligible salary. Salary weight matters because a cell with 10 highly paid employees can cost more than a cell with 25 lower-paid employees even when its percentage is smaller.
The budget test is a weighted calculation:
Merit cost for a cell = total eligible salary in the cell x proposed increase percentage.
Total projected merit cost = sum of the merit cost across all cells.
Projected merit rate = total projected merit cost divided by total eligible payroll.
That final rate must land inside the approved merit budget. If it doesn’t, revise the cell values and test again. This is where the grid becomes an actual compensation model.
A common way to draft the first version is to choose a reference cell, often Meets Expectations near midpoint, and set it near the central planning rate. Higher performance rows get larger values. Columns below midpoint get more room, while columns higher in the range get less. The size of each step should be deliberate. A 0.25-point difference between ratings may be too small to feel meaningful, while a 3-point jump can consume the pool quickly.
Don’t force the grid to look symmetrical. The percentages should reflect policy and budget. The pool is a payroll constraint, so test the matrix against payroll dollars rather than simple employee averages.
The process can be done in a spreadsheet with SUMIFS or lookup formulas. A pivot table can help review the distribution. The tool matters less than traceability. A reviewer should be able to see the employee input, assigned cell, recommended percentage, dollar increase, new salary, and new range position without rebuilding the logic from memory.
Worked Merit Increase Matrix Example
The example below uses a 4 x 4 grid. The percentages are illustrative. They aren’t market standards, and they shouldn’t be copied without cost testing. The purpose is to show how performance and pay position can work together inside one merit increase matrix.
| Performance Rating | Below 0.90 | 0.90-0.99 | 1.00-1.09 | 1.10+ |
| Exceptional | 6.0% | 5.2% | 4.4% | 3.2% |
| Exceeds Expectations | 4.8% | 4.2% | 3.6% | 2.6% |
| Meets Expectations | 3.6% | 3.2% | 2.8% | 1.8% |
| Below Expectations | 0.0% | 0.0% | 0.0% | 0.0% |
Read the table in both directions. Moving down a column reduces the recommendation as performance falls. Moving across a row reduces the recommendation as salary moves higher relative to midpoint. An Exceptional employee at 0.86 receives a larger recommendation than an Exceptional employee at 1.14 because the second employee is already paid further into the range.
Now consider an employee earning $68,000 with a range midpoint of $80,000. The compa-ratio is 0.85. If the employee receives an Exceeds Expectations rating, the example grid recommends 4.8%. The dollar increase would be $3,264, producing a new salary of $71,264. The new ratio would be 0.891, assuming the midpoint remains $80,000 on the effective date.
At workforce level, distribution controls cost. If many employees are below midpoint and ratings sit in the top 2 rows, this grid may run above a 3.2% pool. If most employees are near midpoint and ratings cluster around Meets Expectations, it may come in below.
This is why the salary increase matrix needs a population model beside it. The visible grid is the policy. The hidden engine is the weighted payroll calculation that tells you whether the policy can actually be funded.
Test the Matrix Against Payroll Before Approval
Cost testing should happen before managers see the grid. Once a percentage table has been shared, later reductions can look like the organization changed the rules after managers began making decisions. Build several scenarios first and keep the final version locked to the approved assumptions.
The base scenario should use the final calibrated ratings and current salaries. Then test at least 2 stress cases. One can shift more employees into higher performance ratings. Another can test what happens if the lowest range-position column contains more payroll than expected because of late market adjustments or new hires. These aren’t predictions. They’re checks on how sensitive the model is to plausible data changes.
Use weighted averages rather than averaging the cell percentages. If 1 employee gets 6% and 100 employees get 2%, the average of the 2 cell values is 4%, but the workforce cost is nowhere near that figure. Payroll dollars and employee placement determine the real result.
Research on pay and productivity also shows why it’s risky to treat performance pay as a simple uniform rule. An NBER study on within-firm pay inequality and productivity found that pay-productivity relationships varied across earnings levels and were connected with performance-pay practices in the firms studied. That doesn’t tell an employer what matrix percentages to use, but it supports a broader point: performance-linked pay can shape the distribution of earnings inside the firm, so the distribution should be measured after the model is applied.
A practical budget file should show eligible payroll by cell, employee count, proposed percentage, projected dollars, overall rate, and variance from the approved merit budget. Keep a separate override field so finance can see the difference between matrix cost and final submitted cost.
Set an escalation threshold before the cycle starts. For example, an override above a defined number of percentage points may require compensation approval, while smaller moves may sit within manager discretion. The exact threshold is a policy choice. The important part is deciding it before exceptions begin to accumulate.
Handle Promotions, New Hires, Range Maximums, and Market Adjustments
Special cases can break a clean matrix when the rules are vague. Set the treatment for each case before manager worksheets are released so similar situations don’t get different pay actions.
Promotions
A promotion changes the job, so track the related increase separately from merit. If a promotion falls near the review date, define which salary enters the matrix and whether the employee remains merit eligible. The order of calculations can change the final salary.
Document the sequence in the policy. Keep merit and promotion amounts in separate fields so the cost stays visible and the promotion isn’t mistaken for an unusually high matrix recommendation.
New Hires and Partial-Year Employees
New hires create a different question because performance evidence may be limited. Set a clear hire-date cutoff or proration rule and apply it the same way across comparable employees.
Put proration in the calculation rather than leaving it to manager judgment. Apply the factor after the matrix assigns the full-year percentage and keep both values visible for review.
Employees Near or Above Range Maximum
An employee near the maximum may have little room for base-pay growth. Decide whether salary is capped at maximum or whether an approved one-time payment is used for any excess. The answer affects cost and communication.
Don’t force the matrix to solve an outdated maximum. If many strong performers sit at the top of the range, review the structure. A one-time payment can limit fixed payroll growth, but it doesn’t repair a weak range.
Market and Equity Adjustments
Market corrections should remain separate. If an employee is far below the intended range position because the job was priced incorrectly, a very large merit percentage can blur performance and market pay. Handle the correction in its own field, then apply merit according to written policy.
Use the same approach for internal equity corrections. The annual cycle may be a practical time to process them, but the funding source should be explicit so legacy pay issues don’t consume the approved pool.
Check Pay Equity and Legal Risk
A matrix makes pay decisions more structured, but structure alone doesn’t remove risk. Review both the inputs and the outputs. The US Equal Employment Opportunity Commission explains that federal law covers compensation discrimination across salary and other forms of pay, and its compensation discrimination guidance notes that job content rather than job title is central when evaluating substantially equal work under the Equal Pay Act.
The Equal Pay Act also recognizes a merit system as one possible basis for wage differences, but the system still has to operate lawfully. The Department of Labor publishes the statutory language in the Equal Pay Act of 1963, including the rule against sex-based wage discrimination for equal work subject to stated exceptions. Employers should have legal counsel review their own obligations because federal requirements can sit alongside state or local rules.
Test whether rating distributions differ materially across protected groups and comparable job populations. A difference isn’t automatically evidence of discrimination, but it tells you where to ask more questions. Then compare recommended increases, override rates, final increases, and post-cycle range positions.
The performance process deserves special attention because subjective judgment can travel directly into pay. A 2026 NBER field experiment on hierarchy and performance pay found that oversight of supervisors reduced favoritism in bonus allocation in the public-sector setting studied. The setting isn’t the same as a corporate merit cycle, but the finding supports a practical control: important pay decisions should leave enough evidence for someone other than the direct manager to review them.
Keep the audit at the right level. A small department can swing sharply because of 1 or 2 employees, so counts matter. Compare like jobs where possible and review unexplained differences.
Pay-equity review also needs to look beyond the current year’s percentage. An employee can receive the same increase as peers and still remain underpaid because of earlier decisions. Post-cycle salary position should therefore sit beside increase percentage in the analysis. A fair-looking annual change doesn’t erase a weak starting salary.
Give Managers Guardrails Without Letting the Matrix Drift
Manager discretion has value when the data can’t capture the full context, but open-ended discretion can undo the point of a salary increase matrix. Define what managers may change, how far they may move, what evidence they must provide, and who approves the exception.
A simple approach is to give each manager the matrix recommendation plus a limited override range. Larger exceptions go to compensation or a review committee. The worksheet should require a reason code such as retention risk, recent role change, documented market issue, or correction of an input error. Free-text notes can add context, but reason codes make it easier to review patterns across the company.
Don’t let a manager move money between employees without showing the effect on the approved pool. If one recommendation rises, the file should show whether the department remains within allocation or needs approval.
Give managers a short explanation of the logic. They need to know what the rating means, how range position affects the recommendation, which actions sit outside merit, and when an override needs approval.
Communication with employees should stay separate from the calculation file. Managers can explain that annual increases reflect the organization’s pay process and individual circumstances, but they shouldn’t promise that a rating always equals one fixed percentage. Range position and budget can change the outcome for 2 employees with the same rating.
Review the Cycle and Fix the Next Matrix
At the end of the cycle, compare the approved model with final outcomes. Review overrides and dollar concentration. Also check whether any job families remain clustered below midpoint.
Start with budget accuracy. Compare the projected rate with the final rate after approved changes. If the difference is large, trace it to the underlying action rather than leaving it buried in one final percentage.
Then review differentiation. If Exceeds Expectations and Meets Expectations differ by only a few tenths of a point for most employees, the grid may be practically flat. If the gap is large, check whether the cost still matches the stated pay policy.
Review movement through the ranges as well. The matrix should show whether lower-paid employees with strong performance are moving toward midpoint at the pace you expected. It should also reveal whether employees high in range are repeatedly receiving base-pay increases that push them above maximum. Those patterns can guide the next structure review.
Audit manager behavior as well. Calculate override frequency and average override size by manager or business unit. Heavy use of exceptions may point to a local pay issue or weak rating calibration.
Use the findings in the next cycle. Compensation planning becomes easier when each annual matrix starts from evidence gathered in the prior year. Range bands and approval rules can then be changed for a reason rather than by habit.
Common Merit Matrix Mistakes to Avoid
The most common mistake is copying percentages from another employer. A grid built for a 4% pool and employees near midpoint can fail at a company with a 3% pool and many employees below midpoint. Percentages only make sense when tied to the employer’s own population.
Stale salary ranges are another problem. A low ratio looks precise, but it can mislead when the midpoint no longer reflects the intended market. Range quality has to be checked before the grid is trusted.
A third mistake is allowing rating inflation to become a budget problem. If managers learn that top ratings produce materially higher increases, pressure to move employees upward can grow. This is why performance calibration belongs before pay recommendations and why the rating distribution should be monitored over time.
Too much detail also hurts. A 7-row by 7-column grid can create thin cells and percentage differences that managers can’t explain. Use enough bands to support real policy differences and stop there.
Mixing merit with promotions or market corrections also clouds the record. Separate action types make the compensation planning file easier to audit and show why payroll moved.
Don’t skip the output review. A matrix can stay within budget and still produce weak patterns. Review group outcomes and range movement before payroll is finalized.
Final Takeaway
A useful merit increase matrix converts pay policy into a repeatable rule and gives finance a model that can be costed before money is committed. The controls around the grid matter as much as the percentages inside it.
Build from clean ranges and calibrated ratings. Map the real workforce before finalizing percentages. Keep other pay actions separate and test exceptions before approval. The approved pool is then easier to explain.
Revise the model when the evidence calls for it. Heavy clusters below midpoint are one signal. Repeated overrides or strong performers stuck at range maximum should also shape the next compensation planning cycle.
FAQs
What Is a Merit Increase Matrix?
A merit increase matrix is a table used to recommend salary increases by combining an employee’s performance rating with the employee’s position in a salary range. The result is a percentage that can be applied to current base salary, subject to the employer’s budget and pay rules.
How Do You Calculate Compa-Ratio for a Merit Matrix?
Divide the employee’s current base salary by the midpoint of the assigned salary range. An employee earning $72,000 in a range with an $80,000 midpoint has a compa-ratio of 0.90. The value shows where current pay sits relative to midpoint.
What Should the Performance Rows Be in a Merit Matrix?
Use rating levels with clear definitions. Many employers use 4 or 5 levels. The matrix doesn’t need a separate row for ratings that managers can’t reliably distinguish.
How Many Compa-Ratio Bands Should a Matrix Use?
Use enough bands to create meaningful pay-position differences without creating tiny cells. A 4-band design is a common starting point, while broad ranges may need different cutoffs.
How Is a Merit Budget Different From a Salary Increase Budget?
A merit budget funds performance-related base-pay increases. A broader salary increase budget may also include promotions, market changes, equity corrections, or other salary actions. Employers should define which actions are included before calculating the pool.
Should Inflation Determine Merit Increase Percentages?
Inflation provides economic context, but it measures consumer prices rather than employee performance or range position. A cost-of-living adjustment should have its own stated policy.
Can Employees With the Same Performance Rating Receive Different Increases?
Yes. In a matrix that also uses range position, 2 employees with the same rating can receive different recommendations because one is paid lower in the salary range. The policy should explain that difference clearly.
What Happens When an Employee Is Above the Salary Range Maximum?
Follow a written policy for employees above range maximum. The treatment may cap base pay or use a one-time payment, subject to the pay structure and applicable law. Repeated cases may signal that the range needs review.
Should Promotion Increases Be Included in the Merit Matrix?
Usually, track promotion pay separately because it reflects changed job scope. Also define the calculation order when promotion and merit actions occur close together.
How Should New Hires Be Handled in the Annual Merit Cycle?
Set a hire-date cutoff or a written proration rule before the cycle begins. The rule should reflect how much performance evidence is available and should be applied consistently to comparable employees.
How Do You Know Whether a Matrix Fits the Budget?
Map every eligible employee to a matrix cell, apply the proposed percentage to the employee’s salary, and sum the projected dollar increases. Divide that total by eligible payroll. The resulting rate should fit the approved pool before the grid is released.
Should Managers Be Allowed to Override Matrix Recommendations?
Limited overrides can help address facts the grid doesn’t capture, but they should have boundaries. Define the permitted range, required reason, approval path, and budget treatment before managers receive their worksheets.
How Often Should a Merit Matrix Be Updated?
Review it for every annual pay cycle. Ranges and workforce distribution can change from year to year, so old percentages should be retested before reuse.
Can a Merit Matrix Fix Pay Equity Problems?
A matrix can make current decisions more consistent, but it can’t erase earlier pay differences. Review starting salaries and comparable job placement alongside ratings, increases, overrides, and post-cycle salaries.
What Is the Biggest Mistake When Building a Salary Increase Matrix?
The biggest mistake is choosing percentages before testing them against real employees and payroll. A grid can look reasonable and still exceed the budget or produce weak pay outcomes once the actual workforce distribution is applied.
