Skip to content

Create a merit matrix

A merit matrix decides what increase to recommend for someone, based on how they performed and where they already sit in their salary band. It gives every manager in a cycle the same starting number, so a proposal begins from a rule rather than from a blank field.

Required role: Administrator.

Create the rating scale you want the matrix to use. The rows of the grid come from that scale’s levels, so the grid stays empty until you choose one.

Grades and their salary bands also need to exist, because the columns of the matrix describe where someone sits inside their band.

  1. Go to Compensation → Merit Matrix.

  2. Select Create Matrix.

  3. Enter a Name.

    Example: 2026 Standard Merit Matrix.

  4. Enter a Description if the matrix needs context for whoever inherits it.

  5. Choose the Rating Scale.

  6. Fill in each cell of the grid, setting the recommended percentage and its Min and Max.

  7. Set Default matrix if new cycles should pick this one automatically.

  8. Set Active.

  9. Select Save Matrix.

OptionDescription
NameWhat the matrix is called. It appears when an administrator picks a matrix while creating a cycle.
DescriptionOptional context for whoever configures the next cycle.
Rating ScaleThe scale whose levels become the rows of the grid.
Default matrixMarks this as the matrix offered first when a cycle is created.
ActiveWhether the matrix can be chosen for a new cycle.

Each cell sits at the intersection of a rating and a position in the salary band. The band positions are fixed: Below Min, Min–25%, 25–50%, 50–75%, 75%–Max, and Above Max.

Every cell carries three numbers. The recommended percentage is what the manager sees pre-filled; Min and Max are the range they can move within without the proposal counting as an exception.

Example: at HC Corp UK Ltd, an engineer rated at the top of the scale who sits in Min–25% of their band is low-paid for strong performance, so that cell usually carries a higher recommendation than the same rating at 75%–Max.

The point of the second dimension is that a rating alone does not say whether someone is underpaid. Two people with identical ratings can deserve different increases because one is already near the top of the band.

The grid refuses to save numbers that contradict each other: a minimum above the recommended figure, a maximum below it, or a minimum above the maximum. It reports Some cells have validation errors — fix them before saving rather than saving a grid that cannot be applied.

A matrix already used by a cycle keeps working. Deleting it removes it from the list for future cycles, and any cycle referencing it continues from the copy taken when that cycle was created.

That copy is what makes a finished cycle readable a year later: the recommendations it produced still make sense, because the grid behind them did not move.

The matrix becomes selectable when you create a compensation cycle. Once a cycle launches, each employee’s rating and position in range are read against the grid, and the resulting percentage arrives in front of their manager as a starting recommendation rather than a decision.