Panda's Recap

What Makes a Good Performance Review

Last updated: August 4, 2026

Most advice treats the self-review as a summary you hand your manager. It is closer to evidence. Somewhere after you submit it, your manager sits in a room with their peers and argues for you, and what you wrote is most of what they have to argue with. That changes what belongs in it.

Who actually reads it

Your manager, obviously. But usually not only your manager.

At most companies of any size, ratings get compared across teams before they are final. Your manager goes into that conversation carrying your year, alongside other managers carrying theirs. Some of the people in the room have never met you. They are not being unfair; they simply have nothing to go on except what your manager can tell them.

This is the part that changes how you write. You are not trying to sound good. You are trying to hand someone the specific sentences they can repeat on your behalf. A manager who walks in with "she had a strong year and the team likes working with her" is bringing an opinion. A manager who walks in with "she cut the checkout failure rate from 4% to 0.5% and it held through the holiday peak" is bringing a fact, and facts survive a room.

So the test for every line is not "is this true" or "does this sound impressive". It is: could my manager repeat this to someone who has never met me, and would it land?

What a good entry looks like

The same shape as a good résumé bullet, aimed differently. A résumé is arguing you could do the job. A self-review is arguing you did.

None of the strong versions use a bigger word than the weak ones. They are just harder to write, because you have to know what actually happened.

What people leave out

The obvious wins go in without prompting. These are the ones that get forgotten and are often the ones that matter most for a promotion case.

• Scope that grew quietly. If you picked up a second client, absorbed someone's work when they left, or started being the person people ask, that is a change in your job. Nobody will write it down for you.
• Work that stopped things going wrong. The outage that did not happen is invisible by design. Name it anyway.
• The unglamorous thing nobody else would do. Documentation, a migration, cleaning up a mess you did not make. This is usually where the goodwill comes from, and it almost never makes it into the form.
• Things you taught someone. Not "mentoring" as a category. The specific person, and what they can now do without you.
• Work that got cancelled. You still did it. If a decision above you killed it, say what you learned and move on, rather than pretending the months did not happen.

Recency is the problem

Here is the uncomfortable part. Your manager does not remember your year either.

They remember the last two months clearly, one or two big things from earlier, and whatever went wrong. That is not a criticism of your manager. It is how anybody remembers twelve months of somebody else's work while also doing their own.

Which means the self-review is not really a summary of a year you both witnessed. For most of the year, you are the only witness. If March is not in your document, March did not happen.

And you are working against the same limit. You sit down in November to account for eleven months, and what comes back is November, plus a handful of things that were dramatic enough to stick. The steady, competent middle of your year is exactly the part that vanishes, and it is usually the part that makes the case.

This is why "keep a brag document" is the one piece of review advice that actually works, and why almost nobody manages it. It asks you to maintain a second document about your job, on top of your job, for a payoff eleven months away. Most people start one in January and abandon it by February.

What to skip

• Adjectives about yourself. Hardworking, passionate, detail-oriented. They are unfalsifiable, so they carry no information. Describe something you did and let the reader draw the adjective.
• Anything your manager cannot verify. If it is not checkable, it can only be taken on trust, and trust is what you are spending rather than earning.
• Grievances dressed as feedback. There may be a real problem with the team or the process. The self-review, which is attached to your rating, is the worst available place to raise it.
• False modesty. Underselling does not read as humble in a calibration room. It reads as a weaker year.
• A list with no shape. Twenty bullets of equal weight makes your manager decide what mattered. Decide for them.

That's what Panda's Recap is

It is a board where you run your work: tasks, timers, clients, projects. Track your work as you go, and the record builds itself as a by-product of doing the job, rather than as a separate document you have to remember to maintain. That is the difference between this and a brag document. There is nothing extra to keep up.

Come review season, it turns that record into a draft organised by quarter, client, project or area, with the hours and the outcomes you noted at the time already in it. You edit it into your own voice, which is a much smaller job than remembering March.

A review draft built from tracked work. The clients here are invented; the structure is what the record produces.

The board, the timers, the reports and the draft are free, with no account needed. Having it rewritten into a finished write-up is part of Pro.

Common questions

How long should a self-review be?

Long enough to make the case, short enough that your manager reads all of it. If your company gives you a box, fill it properly rather than treating the limit as a target. Nobody has ever been marked down for being easy to read.

What if I did not hit my goals?

Say so, early, with what happened and what you did about it. A missed goal you name yourself reads as judgement. The same goal, discovered by your manager in a document that never mentions it, reads as something you hoped nobody would check.

Should I include work that failed?

Include it if you learned something you then used. A project that was cancelled is not a mark against you; it is usually a decision somebody above you made. What matters is whether you can describe why it went the way it did.

What if most of my work is invisible maintenance?

Maintenance work is invisible precisely because it works, which makes writing it down more important rather than less. Describe what did not happen: the outages nobody had, the questions that stopped arriving, the process that no longer needs you standing next to it.

Is it a bad idea to use AI to write my self-review?

It depends entirely on what it is writing from. Given nothing, a model produces fluent paragraphs about collaboration and passion that could describe anybody. Given a real record of what you did, it is doing the part most people are worst at, which is turning notes into sentences. The judgement about what mattered is still yours.