Guide · Pay reviews

Your first performance and pay cycle

Keep the first cycle light. Make one clear pay decision per person and explain it.

Short answer: A startup's first performance and pay cycle should be lighter than it looks in a big company's handbook. Make one clear pay decision per person, with one reason, compared across teams before it is final. Tell people beforehand how decisions will be made. The habits from the first cycle last for years, so a simple process done well beats a full one done badly.

The borrowed template

Picture a 60-person company in Tucson that makes route-planning software for heating and cooling contractors. It closed its Series A fourteen months ago and has doubled since. It has never run a performance or pay review. Raises so far have happened when someone asked, when someone threatened to leave or when a new hire's offer made an existing salary look low.

The chief operating officer decides the company will run its first proper cycle in January. She finds a template from a much larger company where a friend works. It has a five-point rating scale, self-reviews, peer feedback from three colleagues each, a calibration committee, a merit matrix linking ratings to raises, promotion nominations and an equity refresh round, all on one timeline.

The company has seven people managers. Five of them have never managed anyone before this year. None has written a performance review. The founders have set aside about 4% of payroll for raises, a number picked because it sounded normal.

Six weeks later the cycle is half done and already late. Managers have spent evenings writing reviews. Two teams have rated almost everyone "exceeds expectations" and one has rated almost everyone "meets". The matrix turns those ratings into raises that differ by team more than by performance. Nobody has yet told employees how any of this will work.

This is a composite of a common situation at companies this size, not a client. The details and figures are illustrative.

Why the first cycle is hard

The first cycle sets two good goals against each other. The first is rigor. A company that has paid people through individual deals wants a process that looks fair and that it can defend: Ratings, criteria, a record of why each decision was made. The second is capacity. A 60-person company has new managers, no compensation specialist and a product to ship. Every hour spent on the process is an hour taken from the work it is meant to reward.

Large companies resolve that tension with infrastructure the startup does not have. A compensation software provider that tracks merit cycles describes the full process as running nine to twelve weeks from ratings to employee communication [1]. That is a long time for a startup, run by people doing it for the first time.

The managers are the second difficulty. A global workplace study put manager engagement at 22% in 2025, nine points below its 2022 level [3]. Managers are being asked to do more with less support and first-time managers at a growing startup are usually the least supported of all. A process that depends on detailed written reviews from every manager depends on the people with the least time and the least practice.

The third difficulty is that first cycles leave a mark. Ratings and raises from the first cycle become the starting point for the next one. A field study of performance ratings at a large financial services firm found that managers anchored on the previous year's ratings, which kept gaps in place even after the company stopped showing managers their employees' self-ratings [4]. A first-cycle rating given in a hurry by an untrained manager can follow someone for years.

Two terms will help. A rating is the label a manager gives a person's performance over the period. A merit increase is a raise to base pay tied to that performance, as distinct from a raise for a promotion or a change in the market.

Five traps in a startup's first pay cycle

"Copy the process from a bigger company"

What it looks like: The company adopts a full performance and pay process from a larger employer or a software template: Rating scales, self-reviews, peer feedback, a calibration committee, a merit matrix and several separate decisions on one timeline.

Why smart teams fall into it: It looks professional and complete. Founders who worked at larger companies remember the process and assume the parts all need to be there. A full template also feels safer than designing something smaller, because nobody can say a step was missed.

What it costs: Each part of a big company's process exists to solve a problem of scale: Hundreds of managers, thousands of ratings, a compensation team to check them. At 60 people most of those problems do not exist yet, while the cost of each step does. Managers spend weeks on documents that change very few decisions. The cycle runs late and the pay decisions, the part people care about most, are made in a rush at the end. Employees learn that the process is heavy and the outcome is unclear, which is the opposite of what the company set out to show. A first cycle should include only the parts that change a decision.

"Let the first ratings set the raises"

What it looks like: Managers rate their people and a matrix turns each rating into a percentage. The pay decision is automatic once the rating is in.

Why smart teams fall into it: It feels objective. A formula removes the founder's personal judgment, which is exactly what the company wanted to move away from after years of individual deals.

What it costs: The formula is only as good as the ratings and first-cycle ratings are the least reliable a company will ever have. Managers have no shared sense of what "exceeds" means. Some rate generously to protect their team, others strictly to look rigorous. A formula applied to those ratings turns differences between managers into differences in pay and does it with an air of precision. In one compensation software provider's data for the first half of 2026, the median raise was 4.7% for people rated above expectations, 3.3% for those meeting them and 2.0% for those below [1]. At a company where one team's managers rate high and another's rate low, that spread lands on teams, not on performance. In a first cycle, the rating should inform the decision, not make it. Our guide on merit matrices and manager judgment looks at where each approach breaks.

"Make every decision at once"

What it looks like: The first cycle handles merit raises, market adjustments, promotions, title changes, equity refreshes and sometimes bonuses, each with its own rules, budget and approvals, all in the same few weeks.

Why smart teams fall into it: The company has not looked at pay systematically before, so there is a backlog. It seems efficient to clear it in one pass.

What it costs: Each decision is reasonable alone. Together they create a set of overlapping raises that nobody can explain. One person gets a small merit raise, a market adjustment and a new title. They cannot tell which part was for what. Another gets one larger raise and wonders why their colleague got three. Managers struggle to keep the rules apart and budgets leak from one pool into another. The simpler approach is one decision per person: Does their pay move, by how much and for which main reason. Promotions can still happen and market gaps can still be fixed. They just land as one change with one explanation. Our note on merit, promotion and market adjustments covers why mixing the three hides what each one is for.

"Let each manager run it their way"

What it looks like: The founders give managers a budget and a deadline. Each manager decides how to assess their team and what to propose. Proposals are approved team by team.

Why smart teams fall into it: It is the opposite reaction to the borrowed template. Managers know their people, the company is small and trust is part of the culture. Checking every manager's numbers can feel like second-guessing.

What it costs: Without one look across teams, nobody sees that two people doing similar work at similar levels are getting very different raises because they report to different managers. Generous managers' teams do better than strict managers' teams, whatever the performance. The people who notice first are the ones on the stricter teams. It also leaves the company with a weak record. Under US federal guidance, a merit system counts as a defense to an equal pay claim only if its criteria are set in advance, communicated to employees and applied consistently [6]. A process that varies by manager is hard to describe as one system. One short session where the founders and managers look at all the proposals side by side catches most of this.

"Tell people when it's done"

What it looks like: The cycle happens out of sight. Employees learn about it when their manager tells them their raise, or when it shows up on a payslip.

Why smart teams fall into it: The process is new and leaders are not sure it will work. Announcing it in advance feels like making promises before the details are settled.

What it costs: The first cycle is when employees learn how pay works at the company. If they learn it only from the outcome, they will build their own theory of how decisions were made and many of those theories will be worse than the truth. Research on reactions to pay raise decisions found that the fairness of the process shapes trust in the supervisor and commitment to the organization, separately from how people feel about the amount [5]. A short note before the cycle, saying what will be decided, on what basis and when, costs almost nothing. In the EU, the criteria used for pay progression must be easily accessible to workers anyway [7]. Our guide on explaining a small raise to a strong performer covers the conversation once the numbers are set.

What a good first cycle looks like

A good first cycle does not try to be complete. It tries to produce one fair, explainable pay decision for each person and to leave the company with habits it can build on. At 50 to 150 people, the first cycles that hold up share these qualities.

The salaries going in were sound enough to build on. If pay was set one deal at a time and some people are far out of line, a merit cycle will not fix it. Raises are a percentage of what people already earn, so they carry the old gaps forward. Our guide on cleaning up founder-set salaries covers what to fix first.

Every part of the process changes a decision. Self-reviews, peer feedback and detailed rating scales are kept only if they would change what someone is paid or how they are developed. Most first cycles need less than they think.

Each person gets one decision with one main reason. Whether that reason is performance, a promotion or a gap to the market, the person can say afterward why their pay moved.

The proposals are compared across teams before they are final. The founders and managers look at all proposed changes together, by level and by team. The question is not whether each number is right, but whether similar cases were treated alike.

The result is checked by group. Before decisions go out, someone looks at the raises by gender and by team. A first cycle that gives one group consistently less is a pattern to understand now, not in three years.

Managers know the reasons before the conversations. A short briefing for each manager is worth more than a long review form.

Employees knew the rules in advance. What would be decided, on what basis and when. Nothing in it was untrue. Nothing was promised that the company might not keep.

The company wrote down what it learned. Which parts helped, which did not and what to change next time. The second cycle is where a startup's process starts to become its own.

The calls only you can make

The weight of performance in pay. Some startups want a strong link from the first cycle. Others want to fix market gaps first and add differentiation later. Both are defensible. The choice tells your people what you reward.

Whether to use ratings at all. A short scale gives managers a shared language. No scale avoids labeling people in a first cycle where the labels are unreliable. If you rate, decide whether people will see their rating.

How big the budget is. Recent US surveys put typical merit budgets a little above 3% [1] [2]. A startup can spend more or less depending on its cash, how far behind the market it is and how much it wants to keep people. The number should come from those choices, not from what sounds normal.

Who has the final say. The founders, a small group of leaders or the managers within limits. Whichever you choose, say it in advance.

How often to run it. Once a year is common. Some startups growing quickly run a lighter cycle twice a year. More often means fewer people waiting a long time for a change and more time spent on the process.

The data and the law

  • Raise sizes by rating. A compensation software provider's data for the first half of 2026 shows a median raise of 3.4% for eligible employees. By rating, the medians were 4.7% for above expectations, 3.3% for meets and 2.0% for below. The same provider describes the full cycle as taking nine to twelve weeks [1]. These are the provider's own figures.
  • Merit budgets. A survey of 756 US employers in spring 2026 found an average merit increase of 3.1% and that most employers base raises on performance together with where a person's pay sits against the market or peers [2].
  • Managers. A global workplace study covering 2025 found manager engagement at 22%, down from 31% three years earlier [3].
  • Ratings carry forward. A field study at a multinational financial services firm from 2015 to 2018 found race gaps in manager ratings, with women of color receiving the lowest final ratings. Hiding employees' self-ratings from managers barely changed the gaps, because managers anchored on the previous year's ratings [4].
  • Process and trust. A study of 217 employees' reactions to pay raise decisions found that the fairness of the process explained more of their trust in their supervisor and commitment to the organization than the amount did [5]. This study is older than the others here. We cite it because its finding is the basis of most later research on pay fairness.
  • United States, merit systems. Under federal guidance on equal pay, a merit system is a defense only if it was not adopted with discriminatory intent, is based on criteria set in advance, has been communicated to employees and is applied consistently. Systems that rely heavily on subjective judgment get close scrutiny [6].
  • European Union. The pay transparency directive requires the criteria used to set pay, pay levels and pay progression to be easily accessible to workers, objective and gender-neutral. Member states may exempt employers with fewer than 50 workers from the pay progression part (Article 6) [7].

This is general information, not legal advice. Rules and their national versions differ. Check the rules where you employ people.

Our view

We think a startup's first performance and pay cycle should be the lightest process that still produces fair, explainable decisions. That means one pay decision per person with one main reason, a single look across teams before anything is final, a check by group and a short note to everyone beforehand about how it will work.

The temptation is to make the first cycle look like a mature company's. We would resist it. The full machinery can come later, once managers have done a cycle, ratings mean something and the company knows which parts it needs. A first cycle people understand is worth more than a complete one nobody can explain.

Sources

  1. Pave, merit increase data for the first half of 2026 (26 June 2026), checked 2026-10-30
  2. HR Dive, Peanut butter pay raises are not mainstream (8 May 2026), on Mercer's QuickPulse US Compensation Planning Survey of 756 employers, checked 2026-10-29
  3. Gallup, State of the Global Workplace: 2026 Report (2025 data, 141,444 employed respondents), checked 2026-10-30
  4. Bohnet, Hauser and Kristal, Can gender and race dynamics in performance appraisals be disrupted? The case of social influence, Journal of Economic Behavior and Organization 235 (2025), summary by the Harvard Kennedy School, checked 2026-10-29
  5. Folger and Konovsky, Effects of procedural and distributive justice on reactions to pay raise decisions, Academy of Management Journal 32(1), 1989, pp. 115 to 130 (abstract), checked 2026-10-29
  6. US EEOC, Compliance Manual Section 10: Compensation Discrimination, 10-IV.F.1 (merit and incentive systems), checked 2026-10-29
  7. Directive (EU) 2023/970 on pay transparency, Article 6 (EUR-Lex), checked 2026-10-28

Merit matrix or manager judgment?

Merit matrices fail at the edges and manager judgment fails at scale. Five ways merit decisions break and what a sound process looks like.

Cleaning up founder-set salaries

Founder-set pay is inconsistent, not wrong one by one. Five traps in the cleanup and why the system has to come before individual fixes.

A small raise for a strong performer

When a strong performer gets a small raise, the number is already set. Five traps in that conversation and why to explain the system instead.

See Resources for every guide and the self-check.