Buy Sumizeit infographics

How to Build a Performance Review Process for Managers That Actually Resists Bias

By Sumizeit · Posted on

A performance review process for managers that collects real input and catches bias before it reaches your team.

Not sure what to read next? Take the 1-minute personal growth plan quiz Start the quiz →
Share this article

TL;DR

A good performance review process for managers starts months before you write a word. Set real goals at the start of the cycle, pull in self-reflection and peer input instead of relying on your own memory, and watch for recency bias and favoritism before they quietly decide the rating for you. Calibrate with other managers before anything becomes final. None of this requires a management degree. It requires borrowing a few ideas from people who've studied this for a living, which is faster than it sounds when you pull the idea from a book summary instead of the whole book. That's the gap Sumizeit fills.

Why most performance review processes break down under pressure

Most managers don't have a performance review process. They have a deadline and a blank form, and whatever happened in the last three weeks fills in most of the blanks. That's not a character flaw. It's just how memory works under time pressure, and it's the reason so many reviews read like they were written about the wrong month of someone's year.

The result is predictable. Someone who had a rough September but a brilliant Q4 gets graded on September. Someone who quietly shipped solid work all year without making noise about it gets a vague paragraph, because nothing dramatic happened to jog your memory. Neither outcome is fair, and both are completely avoidable with a little structure built in before review week ever shows up on your calendar.

What a real performance review process for managers looks like

A performance review process for managers worth having isn't one document you write alone the night before it's due. It's closer to four separate inputs that get combined into one picture. You write your own read on the person's work across the whole period, not just the recent part. The employee writes a self-reflection on their own progress and challenges. A couple of their peers weigh in with specific examples, not vague praise. And if your company uses ratings, those four inputs feed the number, rather than the number coming from your gut on a Tuesday afternoon.

You don't need all four pieces running on day one. Plenty of teams start with just the manager's write up and the employee's self-reflection, then add peer input once that rhythm feels normal. The point isn't to build a perfect system immediately. It's to stop treating the review as one person's memory dump and start treating it as a small, deliberate research project with your direct report as the subject.

If you want the mechanics of actually writing the words once you've gathered this material, how to write a performance review for your team when you have no idea what to say picks up right where this post leaves off.

Set the goals before the cycle starts, not during the review

You cannot fairly grade someone against a target you never actually agreed on. This sounds obvious and it's violated constantly. A huge share of mediocre reviews trace back to a goal that existed only in the manager's head, or a goal that got set in January and silently drifted by June without anyone updating it.

Fix this at the start of the cycle, not the end. Sit down with each report and write two or three goals you can both point to later and agree on whether they happened. Measure What Matters is the clearest short course on building goals specific enough to actually check later, instead of the kind of goal that's really just a vibe wearing a bullet point.

If the goals shift mid cycle, which they often do, write down when and why. That one habit alone will save you from the awkward review conversation where your report thought they were hitting the mark and you thought they'd missed it, because you were quietly grading against a target that changed three months ago.

Collect input from more than just your own memory

Your memory of someone's year is a biased sample, and not because you're a bad manager. You simply saw more of what happened near you, in meetings you attended, on projects you were close to. The work that happened without you in the room is real, and your review should reflect it even though you weren't watching.

This is why a short self-reflection from the employee matters more than it sounds like it should. Ask them to write down what they're proud of, where they struggled, and what they want to grow into next. You'll usually learn about at least one project or win you'd half forgotten, and you'll see where their own sense of their year differs from yours, which is exactly the gap worth talking through before the review becomes final rather than during the meeting itself.

Peer input closes the rest of the gap. Ask two or three people who actually worked alongside your report for one specific example of something that went well and one area they'd want to see grow, tied to an actual moment rather than a general impression. Multipliers is useful here too, since a chunk of it is about how much capability on a team goes unnoticed by the person at the top simply because they're not close enough to see it.

Watch for the three biases that quietly wreck reviews

Three patterns show up in almost every review cycle, and all three are fixable once you know to look for them.

Recency bias is the big one. You remember last month better than you remember March, so the review quietly becomes a review of last month with a full year's label on it. The fix is boring but it works. Keep a running note, even a messy one, of specific moments throughout the cycle instead of trying to reconstruct the whole year from memory in one sitting.

Favoritism is the second, and it's rarely intentional. Managers write longer, more specific reviews for people who talk to them often, and shorter, vaguer ones for quiet, steady performers who just do the work without narrating it. If one person's review draft is twice as long as another's, that's worth noticing before you submit either one.

The third is what shows up as grade compression, where almost everyone lands in the same safe middle rating because genuinely differentiating performance feels uncomfortable. Radical Candor makes the case directly that this kind of false kindness isn't actually kind. It just moves the discomfort from the review conversation now to a much worse conversation later, usually when someone's surprised by a promotion decision or a layoff they never saw coming.

Calibrate before you finalize anything

If your company has more than a handful of managers, ratings need a gut check beyond your own head. Calibration means a group of managers compares notes before anything becomes official, looking at the actual evidence behind each rating rather than taking anyone's word for it.

This matters because standards drift between managers without anyone intending it to. One manager's "solid, reliable work" is another manager's "exceeded expectations," and the person on the receiving end has no way to know which manager's scale they landed on. A short calibration conversation, even a rough one over thirty minutes, catches most of that drift before it becomes a pay or promotion decision someone can't take back.

The Making of a Manager is a good one to have in your head going into these conversations, mostly because it's written by someone who was new at exactly this and is honest about how uncomfortable calibration felt the first few times she did it. That discomfort is normal. It doesn't mean you're doing it wrong.

Turn this into a rhythm, not a fire drill every cycle

None of this works as a one time fix the week before reviews are due. It works as a loop. Set goals, check in quarterly, gather self-reflection and peer input before you write anything final, watch for the three biases, calibrate with other managers, then do it again next cycle with a little more muscle memory each time.

The honest barrier for most managers isn't motivation, it's time. You're not going to read five management books between now and your next review cycle, and you don't need to. Fifteen minutes with a book summary on your commute or over lunch is usually enough to walk into a calibration conversation or a goal setting meeting with one sharp idea instead of none. That's the entire reason Sumizeit exists, in text, audio, video, or infographic form depending on how your brain prefers to take things in. And if you're staring down this cycle's reviews right now, how to prepare for a performance review in 11 practical moves is worth a read too, written from your report's side of the table, so you know exactly what they're hoping to hear back from you.