Performance Reviews for a Small Team, Without the Jargon

A performance review for a small team should take an hour per person, twice a year, and answer four questions: what went well, what didn't, what to focus on next, and what the person needs from you. Anything heavier than that, at 20 people, costs more than it gives.

Who this is for: founders and managers of teams of 10 to 50 people setting up reviews for the first time, or fixing a process that nobody likes.

Why yearly reviews fail at 20 people

Most small companies copy a big company's annual appraisal. It usually goes like this. In March, someone remembers reviews are due. Managers fill in long forms about a whole year they barely remember. Ratings cluster in the middle. Pay rises go out, people compare notes, and the next eleven months pass without another honest conversation.

That fails a small team for three reasons.

A year is too long. In a startup, someone's role can change twice in twelve months. Feedback about January is useless in December.

Long forms hide the point. Twenty competencies with five-point scales produce a lot of boxes and very little clarity. Nobody leaves knowing the one thing to do differently.

Surprises break trust. If someone first hears about a problem in their annual review, the manager has failed, not the employee. In a team of 20, people sit next to you. There is no excuse for a surprise.

The fix is not a better form. It is a lighter rhythm and a few good questions.

A light, twice-a-year review cycle

Here is a cycle that works for teams of 10 to 50:

Two things make this work.

First, the monthly one-to-ones. They are where feedback actually happens. If you keep them, the six-monthly review becomes a summary of things already said, which is exactly what it should be.

Second, the self-review comes first. When people write their own view before reading their manager's, the conversation starts with what they think, not with a defence.

Pick review months that avoid your busiest periods. For a company with a March financial year end, October and April work well. Don't hold reviews in the same week as month-end, a fundraise or a big launch.

The four questions

Every review, self-review and manager review alike, answers the same four questions. Keep them in plain language.

1. What went well in the last six months?

Ask for specifics: projects, outcomes, moments. "Shipped the new billing flow two weeks early" is useful. "Good team player" is not. Encourage people to include things nobody saw.

2. What didn't go well, and why?

This is not a list of failings. It is a look at what got in the way: unclear priorities, a skill gap, too many projects, a broken process. Some of the answers will be about you as the manager. That is the point.

3. What should you focus on in the next six months?

Agree two or three things, not ten. Mix one stretch goal with one or two that build on strengths. Write them so you will both know in six months whether they happened.

4. What do you need from me?

Training, more context, fewer meetings, a clearer brief, a different project. Write it down and treat it as your homework.

Ratings: keep them simple

You don't need ratings to have a good conversation. You do need them, or something like them, when it's time to decide pay rises, because otherwise the decision ends up resting on who argued hardest.

If you use ratings, use a short scale with plain descriptions:

Most people in a healthy team will sit in "strong in the role". That's good, not a problem to fix. Don't force a curve on 20 people. A forced curve made sense, arguably, in companies with thousands of people. In a small team it just guarantees that someone good gets a label they didn't earn.

Linking reviews to pay

Pay is where reviews go wrong most often. A few rules help.

Separate the conversations. Hold the review first. Share the pay decision a week or two later. When both happen in one meeting, people only hear the number.

Decide the budget first. Before reviews, decide the total amount you can spend on increases this year, as a percentage of the salary bill. Then share it out by rating and by how each person's pay compares with the market.

Explain the decision. Every person should know why they got what they got, in one or two sentences.

Here is a simple worked example for a team with a ₹1.2 crore annual salary bill and an 8% increase budget, so ₹9.6 lakh to share:

The exact numbers depend on your budget, market pay and each person's current salary. If someone is paid well below market for their role, fix that separately from the rating.

In India, remember that an increase to basic pay also changes provident fund and gratuity costs. Our explainer on CTC vs in-hand salary shows how. For senior people, consider equity alongside cash: see ESOPs in India.

Running the review conversation

The written review is half the work. The conversation is the other half. A simple order:

1. Start with their self-review. Ask what they'd like to talk about most.
2. Share what went well. Be specific, and mean it.
3. Talk about what didn't. Use examples. Ask what got in the way before suggesting a fix.
4. Agree the focus areas. Two or three, written so you'll both know whether they happened.
5. Ask what they need from you. Write it down.
6. End with the one thing. If they remember only one sentence from the meeting, what should it be?

Send a short written summary the same day. It's the record you'll both look at in six months. Keep it to the four questions, the agreed focus areas and anything you promised to do. If you change something later, update the summary rather than writing a new one.

A review, filled in

Here is what a short, useful manager review looks like for a customer success lead. It fits on one page.

What went well. Took over 40 accounts in May with no handover and kept every one. Built the onboarding checklist the whole team now uses. Customers mention you by name in renewal calls.

What didn't go well. Two large renewals were flagged late, in the final week. The cause was that account health lived in your head, not in a shared place. Some weekly reports were skipped during the busy months.

Focus for the next six months.

1. Every account above ₹5 lakh a year has a written health score, reviewed every month.
2. Hire and train one more customer success person by January.

What you need from me. A clear list of which accounts sales expects to expand, and a monthly half hour with the product team.

Rating: strong in the role.

Notice what it leaves out. No adjectives without examples, no long list of competencies, no surprises. Every line is something the person already heard during the year, written down so you both remember it.

When someone is not meeting the role

Most people in a small team will be doing well. Occasionally, someone isn't. Handle it early and kindly.

• Don't wait for the review. If there's a serious problem, raise it in the next one-to-one.
• Write a short improvement plan. Three or four specific things, how you'll both know they've improved, the support you'll give, and a date to check, usually 60 to 90 days away.
• Check in every two weeks during the plan.
• Be honest about outcomes. If things improve, say so and close the plan. If not, have the harder conversation with respect.

If the plan ends with someone leaving, handle their exit and final settlement properly. It matters to them and to everyone watching.

What not to do

• Don't copy a 15-page appraisal form. Nobody will fill it in honestly.
• Don't rate people on traits. Rate work and behaviour you can describe, not "attitude".
• Don't save feedback for the review. If it's important, it's important now.
• Don't let the loudest person win the biggest rise. Decide pay against the budget and the ratings, then explain it.
• Don't skip the founder's own review. Ask your team, and your board or co-founders, the same four questions about you. It sets the tone.

Common mistakes

• Reviews only once a year, with no one-to-ones in between.
• Too many goals, so none of them get done.
• Ratings that everyone gets the same, which makes them useless for pay.
• Pay and feedback in one meeting.
• No written summary, so nobody remembers what was agreed.

Questions people ask

How often should a small team do performance reviews?

Twice a year works well for teams of 10 to 50. Once a year is too far apart to fix anything, and quarterly reviews take more time than a small team can spare. Between reviews, keep short monthly one-to-ones so nothing in the review comes as a surprise.

Should a startup use ratings in reviews?

Use a simple scale of three or four levels, with plain descriptions of each level. It helps when you decide pay rises, and it forces a clear answer. Avoid five-point scales with a middle option everyone drifts to, and never force people into a fixed curve in a team this small.

Should the review decide the salary increase?

The review should inform the pay decision, but keep the two conversations apart. Hold the review first, about the work and the next six months. Share the pay decision a week or two later, with the reason. When both happen in one meeting, people only hear the number.

What if someone disagrees with their review?

Listen first, and ask for specific examples you might have missed. Change the review if they show you something real. If you still disagree, write down both views and agree what each of you will look for over the next three months. Offer them a conversation with a second person, such as the founder.