How to link performance appraisals to pay and bonuses

Managers finish the appraisal and hand in a score. Weeks later, someone else works out the bonuses in another spreadsheet, often without the ratings beside them, and that breaks trust.

• Scattered data. KPIs and appraisal results sit across Excel files and systems that don't talk to each other. It's the week before calibration, and HR is still pulling scores together.
• Inconsistent standards. Two managers rate the same performance differently. Nobody notices until calibration, if at all.
• Disconnected rewards. The bonus gets finalized from a spreadsheet that was never checked against the actual ratings.
• No answer to "why?" An employee asks why a colleague got a bigger bonus, and no one can explain it with objective evidence.

Fixing this means doing the steps in the right order, with everyone working from the same data.

Start by sorting pay into what moves with performance and what doesn't. For an Indonesian company it can look like this:

Two rules are worth knowing before you design anything. Under Government Regulation 36 of 2021 on wages, employers must draw up and apply a wage structure and scale that takes the company's ability to pay and its productivity into account, and tell each worker about it individually. And under Minister of Manpower Regulation 6 of 2016, workers with 12 months' continuous service or more receive one month's wage as THR (those with at least one month receive a proportional share), paid no later than seven days before the religious holiday. Check the current text of both with your legal team: this guide isn't legal advice.

Performance pay sits on top of a base that already follows a structure, so if the structure is missing, fix that first.

Because the rating isn't only about the person rated. A study of more than 4,000 managers, each rated for development by seven people (two bosses, two peers, two direct reports and themselves), found that each rater's own rating tendencies accounted for over half of the variation in ratings: 62% in one sample and 53% in the other. The managers' actual performance accounted for less than half as much.

So if you pay straight from raw ratings, part of each bonus depends on who the employee's manager happens to be. Calibration is where you take that out:

1. Put everyone's ratings side by side, by department, plant or level, before anything is decided.
2. Look for patterns before individuals. A manager whose whole team is rated top, or a plant that rates two bands below the others, is the first thing to discuss.
3. Ask for evidence. A rating should move in calibration only on KPI results or examples of behavior, however hard someone argues.
4. Freeze the ratings. Only calibrated ratings go into the pay calculation.

In DEUS Enhance, the calibration matrix shows everyone's evaluations together, so ratings that don't follow the same standard stand out.

Two simple tools do most of the work.

Bonus bands. Group the final ratings into a few bands and give each band a bonus range, for example as a multiple of monthly base pay or a percentage of a target bonus. Similar performance gets a similar reward, and the band is easy to explain. If the company or the unit had a weak year, shrink the pool, keeping the link between band and range.

A raise grid. A merit matrix sets each person's salary increase from two things: their performance rating and their position in the salary range for their grade, often measured as a compa-ratio (salary divided by the midpoint of the range). A strong performer paid below the midpoint gets a bigger percentage raise than an equally strong performer already paid above it. That moves pay toward the midpoint over time instead of further above it.

An example, in words rather than percentages, to show the shape only:

Decide the bands, the ranges and the grid before the ratings come in, and share the rules with managers. If you change them after seeing the names, people stop seeing the system as fair.

In Enhance you set up each bonus cycle and who takes part, group similar performance into similar rewards, and plan raises beside the performance behind them.

• Paying on a score nobody calibrated. Part of each bonus then depends on the rater (section 4).
• Too many decimals. A formula on a 3.27 rating looks more precise than the rating is, so bands are more honest.
• Changing the rules after the ratings. Agree the bands and the grid first.
• Mixing up rewards and allowances. THR and fixed allowances aren't levers for performance, and treating them as one confuses everyone.
• Measuring only results. If the appraisal ignores how people work (the competencies), the bonus can reward results reached in ways you don't want. See the guide on setting KPIs and competencies.
• No explanation for the employee. Each person should see their KPI results, their competency results, the calibrated rating and the band it put them in.