You know the version of the performance review where the manager reads from a script, the employee nods, both parties say “great chat,” and three months later nothing has changed. You have probably sat through that review yourself. Now you have to run it.
Here is the harder truth most first-time managers do not want to hear: the form your company hands you is not the problem, your rating scale is not the problem, even your nerves are not the real problem. The problem is that most performance reviews are run as if they were a once-a-year ritual instead of one moment inside a year-long system. Once you see them as the latter, the whole thing gets easier. The conversation gets shorter. The follow-through gets real. Your team starts asking for the review instead of dreading it.
This guide is the complete playbook. By the end of it you will know exactly what to do in the 90 days before a review, how to write a review that respects your team’s time, how to run a 60 minute conversation without losing the room, which rating scales to use and which to avoid, and what to do in the months after so the review does not become a piece of paper nobody reads. Wherever a deeper article exists, you will find a link to it. Wherever a tool can do the work for you, the tool is linked too.
If you only remember one thing, remember this: a great performance review is the most honest career conversation your team member will have all year. The rest of this guide is just the operating manual.
1. Why most performance reviews fail (and what the working ones do differently)
Research from DDI’s Global Leadership Forecast 2024 found that only 40 percent of leaders rate their organization’s leadership quality as high, and the gap shows up most clearly in performance management. Gallup workplace research has put the number even more bluntly: only 14 percent of employees strongly agree that the reviews they receive inspire them to improve. Eighty six percent of reviews are time both parties will never get back.
The instinct, when you read those numbers as a new manager, is to assume the fix is technical. Better forms. Better rating scales. More questions in the template. That is not where the working reviews differ.
The working reviews differ in four boring ways:
- There are no surprises. The review is a summary of conversations the employee has already had, not the moment they hear hard feedback for the first time.
- The criteria existed before the review did. The employee knew what “great” looked like in October before being judged on it in March.
- The rating, if there is one, is not the point. The conversation is the point. The rating is a side effect.
- Something changes after the review. A development plan, a stretch project, a new accountability, a salary adjustment. If nothing changes, the review was theater.
Notice what is not on that list. Length of the review document. Number of competencies graded. Whether you used a 5 point or a 7 point scale. Those are the things new managers obsess over. The four things above are what actually separate reviews that move people from reviews that just happen to people.
Everything else in this guide is the practical application of those four principles. If something you are doing in your review process violates one of those four, stop doing it, even if HR sent you a 14 page template that says you should.
2. The four systems that make a performance review actually work
Treating the review as one event is the original sin of performance management. You cannot fix in 60 minutes what you have ignored for six months. The managers whose reviews land are running four small systems all year. The review is just where those systems publish.
System 1: Frequency that matches the work
If your company runs annual reviews and that is all you do, your team is operating with a 12 month feedback latency. Imagine driving a car where the steering wheel responds with a 12 month delay. That is what annual-only feedback feels like for the person trying to grow.
You do not need to fight HR to fix this. You need to layer in:
- A real weekly or biweekly one-on-one where small course corrections happen in real time. The corrections only work if the feedback in them is specific and behavioral, not vague — the constructive feedback examples library shows fifteen situations with the bad version and the good version side by side, which is the raw material these 1-on-1 corrections are built from.
- A short quarterly check-in where you and the employee write down (separately, then compare) what is going well, what is not, and what should change. Twenty minutes. No form.
- The annual review itself, which becomes a summary instead of a surprise.
When you run this cadence, the annual conversation gets dramatically easier, because nothing in it is new.
System 2: Criteria that exist before the review
A review without pre-defined criteria is a vibe check. The most common failure pattern for first-time managers is sitting down in February to write a review without having written down, in October, what success looked like for that person.
Fix it by setting clear team and individual goals at the start of the cycle. For each direct report, you should be able to answer four questions before the cycle starts:
- What outcomes does this person own?
- What behaviors do we expect of someone at their level?
- What does growth look like for them this year?
- How will we measure all three?
If you cannot answer those four in October, you have no business writing a review in February. Spend a 1-on-1 fixing that, even if it is awkward, because the alternative is grading without a rubric.
System 3: A rating philosophy you actually believe
Most companies inherit a rating scale from somewhere and pretend it is neutral. It is not. The scale shapes the conversation more than people realize. Section 3 below gets into specifics, but the system level point is: pick a rating philosophy and apply it consistently. Inflate everyone and the rating means nothing. Center everyone on “meets expectations” and your top performers get demoralized. Use a forced distribution (sometimes called stack ranking) and you will make enemies of half your team.
A consistent philosophy beats a perfect scale. Your team can work with “Pat is strict but fair.” They cannot work with “Pat changes the rules every cycle.”
System 4: Follow-through with teeth
The fourth system is the one nobody runs. The review ends, both parties shake hands, the document goes in a folder, and life resumes. Eight weeks later neither person could quote a single thing from it.
Follow-through means three things, all of them small:
- A written development plan with no more than three concrete actions the employee will take in the next 90 days.
- One specific commitment from you. (“I will introduce you to the head of product so you can sit in on roadmap reviews.”)
- A check-in date on the calendar before the review meeting ends.
When follow-through is automatic, the review becomes a renewal of a contract. When it is not, the review becomes paperwork.
3. Rating scales: the ones to use, the ones to avoid, and how to actually use them
Almost every new manager underestimates how much the rating scale influences the conversation. A 5 point scale (“exceeds, meets, partially meets, does not meet, far below”) pulls you toward central tendency. A binary scale (“on track, off track”) forces decisiveness but loses nuance. A behavioral scale (“examples of what each level looks like”) is the most accurate but the slowest to use.
Here is how to think about the choice when you actually have one.
The 3 point scale: best for most teams
If you can choose, choose a 3 point scale: below expectations, meets expectations, above expectations. Three reasons.
First, 3 points is honest. Most people on most teams in most quarters are doing the job. The honest distribution is heavily weighted toward “meets expectations,” which on a 3 point scale is not an insult. On a 5 point scale, “meets expectations” reads as “the middle, which is bad.”
Second, 3 points forces real conversation. There is no comfortable middle to hide in. “Below expectations” means a hard conversation. “Above expectations” means specific evidence. There is no “I gave you a 3 out of 5 because, you know, you’re fine.”
Third, 3 points calibrates faster across managers. When eight managers in a department are rating 60 people, you can calibrate a 3 point scale in a 90 minute meeting. A 5 point scale takes three meetings and still ends up uncalibrated.
The 5 point scale: when you are stuck with it
You will probably be stuck with a 5 point scale at some point. When you are, beat the central tendency problem by writing the rating last, not first. Write the narrative. Write the evidence. Write the development plan. Then assign the rating that fits the story you just wrote, not the other way around.
Also, force yourself to use the full scale. If 80 percent of your reports are landing on the same point, you are not rating, you are sorting them into one bucket and calling it a day.
The “no rating” approach: harder than it looks
A growing number of companies (Adobe, Deloitte, Accenture, GE) have moved away from numerical ratings altogether. The pitch sounds great: focus on the conversation, not the score.
In practice, no-rating systems often shift the rating from explicit to implicit. The rating is now a paragraph instead of a number, but it is still a rating, and now your direct report has to interpret which paragraph means what. If your company runs a no-rating model, your job is to be more explicit, not less. Use clear language: “I would describe this as a strong cycle.” “I would describe this as a year where we did not hit the bar.” Write it down. Verbal-only no-rating systems are how memory drifts and salary discussions go off the rails.
Behavioral anchored ratings: the gold standard
If you ever get to design the system from scratch, use behaviorally anchored rating scales. Each level has a written example of what that level looks like in the actual job. (SHRM has good examples of how Fortune 500 companies build these.) The cost is real (someone has to write them). The benefit is calibration. Two managers reading the same anchor reach the same rating, plus or minus one tick. That is rare in this work.
4. The 90 day prep timeline (start before you think you should)
Most first-time managers start preparing for a review three days before the meeting. By then it is already too late. Use the calculator at Performance Review Time Calculator to get a realistic estimate of how many hours your review cycle actually takes when you account for all of your reports. The number tends to surprise people. For a team of five direct reports, a real cycle is 18 to 25 hours of total work, spread across a month.
Here is what those hours should look like.
T minus 90 days: pre-review hygiene
Three months out, you are not writing reviews. You are making sure you can write them when the time comes.
- Audit your notes. Open your one-on-one notes from the last quarter. If you have not been taking notes in your 1-on-1s, this is your wake-up call. Notes are the raw material of a fair review. No notes means you will write the review from memory, and memory is a recency biased liar.
- Pull the goals. Find the goals you set with each person at the start of the cycle. Are they still on file? Are they still relevant? If a goal is dead because the company pivoted, write that down. The goal does not count against the person, but it does need to be acknowledged.
- Run a temperature check. In your next 1-on-1, ask: “Anything you want me to know before review season starts? Anything you wish you could un-do or get credit for that I might miss?” This single question prevents 80 percent of “you didn’t notice X” surprises in the meeting.
T minus 60 days: evidence gathering
Two months out, you start collecting evidence. Not opinions. Evidence.
- Work outputs. What has this person actually delivered this cycle? Projects, problems solved, contributions to others’ work. Be specific. “Led the migration of three legacy reports to the new dashboard, on time, no rework needed” is evidence. “Strong contributor” is not.
- Cross-functional input. Reach out to two or three people who worked closely with this person. Not a full 360. A casual ask: “I am preparing for Jordan’s review. How has your experience working with them been this cycle?” You are looking for blind spots, not validation.
- Goal progress. For each goal, write down: target, actual, what shifted, what they did about it. This is the spine of the review. Everything else is commentary on it.
- Self-assessment. Two weeks before you write, send the self-evaluation prompt to your direct report. Their version of the year is data you cannot get any other way. It also flags any massive gap between how they see themselves and how you see them, which is the number one signal that you have not been giving enough feedback during the cycle.
T minus 30 days: writing
A month out, you write. Most new managers underestimate this part. A real review takes two to three hours per person to write properly. If you have five direct reports, that is 10 to 15 hours of focused writing time. Block it on the calendar. Treat it like a deadline.
How to write the actual review is in section 5.
T minus 7 days: calibration and pre-talk
A week before the meetings, two things.
First, calibrate with peer managers if your company does that. (If they do not, propose it.) Calibration is not about politicking. It is about catching cases where you are unconsciously grading “high” because you like someone, or “low” because they are quiet. Other managers will catch what you cannot.
Second, send a pre-meeting note to each direct report. “We are meeting Tuesday at 2pm to walk through your annual review. The full review will be in your inbox by Monday afternoon so you can read it before we meet. Plan for 60 minutes. Come with anything you want to discuss or push back on.” This kills 90 percent of the meeting anxiety on both sides.
T minus 24 hours: their review hits their inbox
Send the written review one calendar day before the meeting. Not less. They should have time to read it, react, calm down if they need to, and come into the meeting ready to discuss instead of ready to absorb.
If you cannot do this comfortably, the review is too long. A review your direct report cannot read in 20 minutes is not really a review, it is a research paper.
5. Writing the review: a structured approach that does not turn into a novel
A great written review has three sections and fits on two pages. That is it. New managers want to write everything they have ever observed. Resist that. The written review is a summary; the conversation is where nuance lives.
Section A: Summary of the cycle (one paragraph)
Three to five sentences. What did this person own this cycle. What did they accomplish. What was the headline. Write it as if your boss’s boss would read it (because she might).
Example: “Jordan owned the customer onboarding workstream this year, with a goal of reducing time-to-first-value from 21 days to 14. They delivered 12 days, with measurable impact on Q3 retention. They also stepped in to lead the data migration when Mei went on leave, which was outside their job description and was the kind of judgment call that earns trust across the team.”
Section B: Evidence by area (the body)
Pick three to five areas that matter for this person at their level. Common areas: outcomes, technical or domain skill, collaboration, communication, ownership, growth. For each area, write:
- What happened. Two or three concrete examples. Specific projects, specific moments. Names and dates if relevant.
- What worked. What they did well, said in a way that is useful to them. (“Your willingness to take the migration when Mei went on leave was a leadership move. Even though it cost you on your own goals, it was the right call.”)
- What is the next level. What growth would look like in this area for the next cycle.
Two to three short paragraphs per area is plenty. Bullet points are fine. Prose is also fine. What matters is specificity. If you find yourself reaching for the same generic phrasing across reviews, the cluster article on 150 performance review phrases by competency is a working library of specific phrasing you can adapt to your direct report’s actual situation, organized by ten common competency areas.
Section C: Forward look (one paragraph plus a development plan)
What you want for them in the next cycle. What they have told you they want. What growth opportunities exist. The development plan is three concrete actions, with a who/what/when. (“By end of June, you will lead a kickoff for the new feature flag system, with Mei as your shadow.”) If you cannot write three concrete actions, your development plan is wishes, not a plan.
A note on tone
Write the review the way you would talk to a peer you respect. Not the way HR talks. Not the way a textbook talks. If you would not say a sentence out loud to this person across a table, do not put it in the document. Bureaucratic distance reads as lack of care. Direct, specific, human language reads as respect.
6. The 60 minute conversation: a framework that actually leaves room for the human
The meeting is where most new managers go wrong, even after great prep. They either over-script the meeting (reading the document aloud while the employee stares at a screen) or under-script it (winging it and hoping nothing weird comes up). Neither works.
Use this 60 minute structure. Adapt as needed. Do not skip steps.
Minutes 0 to 5: orient and disarm
Two sentences. “I want this hour to be a real conversation. I sent you the review yesterday, so you have already read it. I am not going to read it to you. I want to use this time on the parts that matter most to you and the parts I want to dig into.”
Then ask: “What stood out to you when you read it?” That question hands them the floor. It also surfaces immediately whether they agree, disagree, or are confused about anything in writing. Whatever comes up here is the most important thing in the meeting, even if it is not what you planned to discuss.
Minutes 5 to 25: discuss the cycle
Walk through the body of the review. Not section by section. Theme by theme. Spend time where they want to spend it. If they push back on something, take it seriously. (“Tell me more about how you saw that project. I want to make sure I am not missing context.”)
This is where most managers either over-defend their writing or fold immediately. Do neither. Listen, restate what you are hearing, and then either update your view or hold your ground with new specificity. (“I hear you that the timeline got pushed because of the security review. That is fair. I also still think the risk should have been flagged earlier. Both can be true.”)
Minutes 25 to 45: forward conversation
Pivot the meeting forward. “Now let me ask the more important question: what do you want next?” Listen to their answer for as long as it takes. Then connect it to the development plan you wrote. Adjust the plan together. (You can absolutely change the development plan in the meeting; it should not be a plan you delivered, it should be a plan you built together.)
This section often runs over. Let it. Forward conversation is where reviews compound. Backward conversation is where they end.
Minutes 45 to 55: the difficult parts (if there are any)
If the cycle was hard (performance was an issue, or you needed to deliver a low rating, or there are concerns about fit), this is where you address it directly. Not first, not last. In a position where there is room to talk, but not so late that you ran out of time.
Be specific. Be calm. Be done. If a PIP is on the table, this is the moment you say so, not in an email next week.
Minutes 55 to 60: close with action and date
End every review with three things stated out loud:
- The summary in one sentence. “I would describe this as a strong cycle with one specific area we are going to work on together.”
- Their next step. “You are going to share the draft of your development plan with me by Friday.”
- Your next step. “I am going to set up the intro to the head of product by end of next week.”
- The date. “We are going to check in on this on June 15. I will put it on your calendar.”
If you do these four things, the review you just had cannot become “we had a great chat and nothing happened.” The conversation is now a contract.
7. The bias problem: how good intentions still write bad reviews
Even managers who do everything else right get tripped up by rating bias. The NeuroLeadership Institute’s research on performance management is worth reading in full, but here is the operating-level summary you need.
Six biases will quietly write your review for you if you do not catch them.
Recency bias
You will overweight what happened in the last three weeks compared to the last 12 months. The last presentation, the last fire drill, the last late deliverable. The fix is mechanical: when you write, force yourself to cite at least one example from each quarter of the cycle. If you cannot, your view of this person is built from a partial year.
Halo and horns effects
If a person is great at one thing, you will assume they are great at everything (halo). If they have one weakness that bothers you, you will see it everywhere (horns). The fix is to rate areas independently. Force yourself to write each section without reading the previous one. Then read the whole thing and ask: would I write this differently if I did not know how I scored the other sections?
Central tendency
You will be tempted to give everyone a “meets expectations” because it is comfortable and avoids hard conversations. This is the bias that, more than any other, makes reviews useless. The check: if you cannot identify the strongest 20 percent and the weakest 20 percent on your team, you are not rating, you are surviving review season.
Leniency bias
Most new managers grade high. You like your people. You understand their constraints. You feel responsible for any gap. The result is a team where everyone is “above expectations,” which is mathematically impossible and tells your direct reports nothing. The check: would your boss describe these ratings the same way? If your boss would say “no, half of these are inflated,” they are.
Similar-to-me bias
You will subtly rate people higher who think like you, work like you, and communicate like you. You will subtly rate people lower who do not. This is one place where the playing favorites quiz is genuinely useful as a self-check; it surfaces patterns you do not see by yourself.
Contrast effect
If the person before in your queue was a star, the next person will read as weaker than they are. If the person before was struggling, the next person will read as stronger. The fix: rate each person against the criteria, not against the last person you reviewed. If you find yourself thinking “well, compared to so-and-so,” stop and reread the criteria.
A small but powerful habit: after you write a draft of all your reviews, read them in random order, and read each one as if you have never met the person. The drafts that do not survive that test are the drafts shaped by bias, not evidence.
8. After the review: the 80 percent of the work nobody does
Every guide on performance reviews you have ever read ended at “have the conversation, take notes, send the document.” That is where 80 percent of the actual value is left on the table.
The follow-through plan is small. It does not take much. But it has to happen, and it has to be visible.
Week 1 after the review: the recap
Within five working days, send a short written recap. Not a re-litigation. A summary of what you both committed to, with dates. Three to five bullets, plus the next check-in date. This recap becomes the artifact you both refer back to, instead of two slightly different memories of an hour-long meeting.
Weeks 2 to 4: integrate the development plan into 1-on-1s
The development plan should not live on a separate page. It should live inside your weekly 1-on-1 agenda. Add a standing “development plan” line item to your shared 1-on-1 doc. It does not need 20 minutes every week. It needs five minutes most weeks, where you both glance at where things are.
If you do not have a 1-on-1 cadence yet, start with the 25 best one-on-one meeting questions and pick three that connect to growth. That is enough.
Mid-cycle: the formal check-in
Three to four months after the review, hold a 30 minute mid-cycle check-in. This is not a mini-review. It is a calibration: are the things we said would happen happening? Has anything in the world changed enough that we should change the plan?
Most development plans get derailed not because of bad intent but because of drift. Quarterly check-ins prevent drift cheaply. Without them, you will arrive at next year’s review with a plan that was abandoned in month two and you will both pretend that is not what happened.
Throughout the cycle: feedback that is not “review feedback”
The biggest mistake first-time managers make in the months after a review is treating feedback as a review-time activity. It is not. The whole point of building the four systems in section 2 is that feedback is happening continuously. The annual review is the moment you summarize it, not the moment you finally do it.
If you find yourself in November storing up feedback for February, give it now. The review will be easier because there will be nothing in it your direct report has not already heard.
Salary, promotion, and the things the review cannot decide alone
A common new manager mistake is letting the review become the place where compensation and promotion decisions land for the first time. By the time the conversation happens, those decisions have usually already been made by your boss, your boss’s boss, and your HR partner. The review is where you communicate the result, and where you make the case for the next cycle.
If you want to advocate for someone’s promotion, the should I promote this employee quiz is a useful pre-review gut check, but the real promotion case is built across the year: in your bi-weekly with your manager, in calibration meetings, in the projects you put this person on. By review week, the promotion is either already in motion or it is not. The review is the explanation, not the decision.
The same applies to underperformance. If a person is genuinely off track, the review is rarely the place that conversation should happen for the first time. Use the should I put this employee on a PIP quiz before the review to check whether you are heading toward a coaching conversation or a formal plan, and use the cycle leading up to the review to make sure your direct report has heard the concern out loud, with examples, before they read it in writing. If the right next step is coaching rather than a PIP, the cluster on coaching an underperforming employee back to meets expectations walks through the four-step framework. If the right next step is a formal plan, the cluster on whether you should put this employee on a PIP covers the decision-tree, the working template, and the conversation that opens it.
9. Performance reviews for remote and hybrid teams: what changes
If your team is fully colocated, sections 1 through 8 apply as written. If your team is remote or hybrid, three things shift.
Evidence is harder to gather, so you must work harder
In an office, you absorb evidence by accident. You see how someone handles a hallway question. You overhear them with a peer in the kitchen. You watch them present in a room. None of that happens remote. Your evidence is what gets written down, what gets recorded, and what gets shared with you on purpose.
The fix is to make evidence collection deliberate. Ask people to share their work artifacts in a shared place. Cross-functional input becomes more important, not less, because your direct line of sight is narrower. Schedule the cross-functional asks; they will not happen by accident.
The conversation needs more structure, not less
Remote review conversations are easier to derail and harder to read. You cannot see body language as well. Silences feel longer. Pushback feels sharper.
Compensate with structure. Send the document earlier. Send a clear meeting agenda alongside it. Use video, not phone. Pause more often than you would in person and ask “How is this landing?” instead of assuming. Take notes during the meeting in a shared doc that you both can see; this slows things down in a useful way and creates a real-time artifact.
The companion guide to running a remote 1-on-1 covers more of this; the same instincts apply at review scale.
Follow-through has to be more visible
In an office, the follow-through happens by osmosis. You bump into someone, you ask about the project, you nudge. Remote teams have no osmosis. If the development plan is not in a shared document with shared deadlines, it will quietly disappear.
Use a real tool, even a simple one. A shared Notion page, a shared Google doc, a single line item in a project tracker. The medium does not matter. Visibility does.
10. Tools, templates, and what to do this week
You do not need new tools to run a great review cycle, but a few will save you hours.
- The Performance Review Time Calculator gives you a realistic estimate of the hours your cycle requires. Use this to negotiate calendar time with your boss before review season starts.
- The Should I Promote This Employee quiz is a pre-review gut check for any direct report who might be ready for the next level. Do it two weeks before you write.
- The Should I Put This Employee on a PIP quiz is the equivalent on the other end. If you find yourself avoiding writing one of your reviews, this is usually the reason; the quiz forces the question. Once the quiz says yes, the PIP Process Cost calculator shows the real operational cost (manager hours, HR time, productivity loss, replacement risk) so you do not start a process that costs more than the alternative.
- The Am I Too Soft quiz catches leniency bias before it lands in a draft. Take it once before you write the first draft of any review.
- The Am I Playing Favorites quiz catches similar-to-me bias.
If you want to read deeper on the mechanics, the first performance review guide walks through a single review at the tactical level. This page you are reading sits one layer up; it is the operating system, that one is the launch checklist.
If you take only one step from this guide today, take this one: open your calendar and put the review prep work on it before anyone asks you to. Block two hours a week, four weeks out. The managers whose reviews actually work do not work harder than you in the two weeks before. They work earlier, in smaller pieces, with less drama.
The review is not a deadline. It is a system. Build the system once and run it every cycle. Your team will start asking for the conversations instead of dreading them. That is what 14 percent of managers actually do differently. Now you do too.
Frequently asked questions
How long should a performance review document be?
Two pages, plus a development plan. If your review is longer, you have stopped writing a review and started writing a research paper. Use the conversation to add nuance; use the document to summarize.
How long should the review meeting itself be?
Sixty minutes for a standard review. Ninety if the cycle was unusual (large new role, mid-year change, performance concerns). If your meeting consistently runs to 30 minutes, you are not having a real conversation. If it consistently runs over 90, you are not preparing well enough up front.
How often should I do performance reviews if my company does not require them?
Once a year is the floor. Every six months is better. Combine an annual full review with quarterly 20 minute check-ins and a real weekly or biweekly one-on-one cadence. The four systems in section 2 are about distributing the review work across the cycle rather than concentrating all of it in one window.
What if my direct report disagrees with the review?
Take it seriously, even when you are confident in your view. Ask for their version. If they raise context you did not have, update the document and tell them you are doing so. If they raise a feeling without new context, hold your ground with new specificity rather than backing down or doubling down. “I hear you, and I still see it the way I wrote it. Let me tell you why.” The disagreement is information either way.
Should I rate myself before reviewing my team?
Yes, briefly. Two questions. Where did I add value to this team this cycle? Where did I get in their way? You do not need to write this down formally, but if you cannot answer the second question honestly, you are probably going to underweight your part of the issues your team is having.
What if I just inherited the team and have no notes from before I arrived?
Acknowledge it. The first review you run for an inherited team is partly a baseline-setting exercise, not a true performance evaluation. Tell people that out loud. Use the cycle to start the four systems from scratch so the next review is honest. Do not pretend you have a year of evidence you do not have.
How do I review someone who is performing but I do not personally like?
Rate the work, not the person. Force yourself to write evidence for both their strengths and their growth areas with the same level of specificity. If you find your “growth area” examples come fast and your “strength” examples are vague, flag it as a bias signal and do another pass. The am I playing favorites quiz is genuinely useful here as a check on yourself before the review goes to anyone.
What if HR wants a long template I do not believe in?
Fill in their template. Then, separately, write the two-page version of the review you actually want to use in the conversation. Lead with your version in the meeting and submit theirs to the system. You are not allowed to skip the template; you are allowed to not let the template run the conversation.