Why the words in a review matter more than they look
Most people assume feedback helps, so any review is better than none. The research says otherwise. In 1996, Avraham Kluger and Angelo DeNisi pooled 607 results from feedback studies in Psychological Bulletin. Feedback raised performance on average, but in more than a third of cases it made performance worse1.
The pattern they found is the useful part. Feedback worked less well the more it pulled attention away from the task and toward the person's sense of self1. "You are careless" is about the person. "Two of the last five invoices went out with the wrong tax rate" is about the work. The second gives someone a thing to fix. The first gives them a thing to defend.
Ratings also say a lot about the rater. A 2000 study in the Journal of Applied Psychology by Steven Scullen, Michael Mount and Maynard Goff looked at two large sets of managers, each rated by bosses, peers and direct reports. The rater's own habits explained 62% and 53% of the differences in scores. The managers' actual performance explained 21% and 25%2. Written evidence is how a review rises above one person's habits.
How to write review phrases that hold up
Describe behavior, not character. A trait ("proactive," "not a team player") is a label the person cannot act on. A behavior is something you saw or heard, and something they can repeat or change.
Use situation, behavior, impact. The Center for Creative Leadership's Situation-Behavior-Impact (SBI) model gives each comment three parts: when and where it happened, what the person did, and what it caused3. It works for praise as well as criticism.
Tie it to an outcome. Name the result the behavior produced or put at risk: a deadline, a customer, a number. This is what a manager needs later to argue for a raise or a promotion4.
End with the next step. A constructive comment should say what "better" looks like, so the person knows what to try next quarter.
A generic review and a specific one
The phrases below are for an invented account manager. The events and numbers are made up for illustration.
| Generic | Specific | |
|---|---|---|
| Strength | Great communicator and a real team player | In the March renewal with the largest client, sent a one-page summary of open issues before each call. The client renewed for two years and cited "no surprises" as the reason. |
| Strength | Goes above and beyond | Rebuilt the onboarding checklist on her own time in Q2. New accounts now reach first order in 9 days, down from 16. |
| To improve | Needs to be more organized | Three of eight quarterly business reviews were rescheduled on short notice. Next quarter, book all reviews by the first week and share the agenda two days before. |
| To improve | Can come across as abrasive | In the June pricing meeting, closed discussion before finance had spoken. Next time, ask each function for its view before proposing a decision. |
Every phrase on the right names what happened and what to do. The phrases on the left could be pasted into anyone's review, which is why they change nothing. Use this tool's output the same way: treat each phrase as a frame, and fill it with a real event from the person's year.
Vague language falls hardest on some people
Vague reviews are not only less useful. They are also handed out unevenly. Shelley Correll and Caroline Simard of Stanford studied over 200 reviews at one large technology company, reported in Harvard Business Review in 2016. Women's reviews had vague praise more often than men's (57% versus 43%). Men's developmental feedback was more often tied to business outcomes (60% versus 40%). And 76% of references to being "too aggressive" appeared in women's reviews4.
An earlier look points the same way. In 2014, Kieran Snyder collected 248 reviews from 180 people at 28 tech companies and wrote up the results in Fortune. Criticism of personality, such as "watch your tone," showed up in 71 of the 94 critical reviews women received, and in 2 of the 83 critical reviews men received5. The sample was people who chose to share their reviews, and it was not peer reviewed, so treat it as a warning sign rather than a measured rate.
A few checks catch most of this before a review is sent:
- Swap the name. Read the review as if it were about someone of a different gender or background. If a word now sounds odd, such as "abrasive," "emotional" or "bossy," replace it with the behavior you saw.
- Count the outcomes. Correll and Simard suggest naming a similar number of specific business results for everyone4. If one person's review has five and another's has none, the gap is in the writing, not only the work.
- Question style comments. Feedback about tone or manner is only fair if it names the moment, the effect and a different way to handle it.
What a phrase list cannot do
Good wording cannot fix missing evidence. The common mistakes happen before anyone writes a sentence:
- Relying on memory. A review written from recall leans on the last few weeks and the most dramatic moments. Keep brief notes through the year of what each person did and what it changed.
- Surprises. A review should confirm what was already discussed. Anything important enough for the review was important enough to raise when it happened, in a regular one-on-one.
- No shared standard. "Meets expectations" means little if the expectations were never written down. Agree on goals at the start of the period, for example with SMART goals, so the review can measure against them.
- Copying phrases whole. A sentence that fits anyone tells the person nothing. If you cannot attach a real event to a phrase, leave it out.
Kluger and DeNisi's finding is the reason to take care: a careless review can leave someone performing worse than no review at all1.
Frequently asked questions
- What are performance review phrases?
- Performance review phrases are ready-made sentences you can use or adapt when writing an employee review, describing strengths and areas to improve across competencies like communication, quality, and collaboration.
- Should reviews be only positive?
- No. The most useful reviews are honest and balanced: they recognize real strengths and name concrete areas to grow, so the person leaves knowing exactly where they stand and what to work on.
- Can I customize these phrases?
- Yes. Treat them as a starting point. Swap in specific examples and numbers from the person’s actual work to make each phrase land.
- How do you write a good performance review?
- Describe what the person did, not what kind of person they are. Name the situation, the behavior and its impact, tie it to a result such as a deadline, a customer or a number, and end with a clear next step. Research on feedback shows that comments aimed at the person rather than the task can lower performance, not raise it1.
- What words should you avoid in a performance review?
- Avoid personality labels such as "abrasive," "aggressive," "bossy" or "emotional," and empty praise such as "great team player." A 2016 Harvard Business Review analysis by Correll and Simard found that 76% of references to being "too aggressive" appeared in women's reviews4. Replace each label with the specific behavior you saw and what it caused.
- What is an example of specific feedback in a review?
- Instead of "needs to be more organized," write "three of eight quarterly business reviews were rescheduled on short notice; next quarter, book them all by the first week and share the agenda two days ahead." It names what happened and what better looks like.
- Can a performance review make performance worse?
- Yes. A 1996 meta-analysis in Psychological Bulletin by Kluger and DeNisi found that feedback improved performance on average, but in more than a third of cases it made performance worse, most often when it drew attention to the person rather than the task1.
Sources
- Avraham N. Kluger and Angelo DeNisi, "The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feedback intervention theory," Psychological Bulletin 119(2), 1996, pp. 254 to 284. doi.org
- Steven E. Scullen, Michael K. Mount and Maynard Goff, "Understanding the latent structure of job performance ratings," Journal of Applied Psychology 85(6), 2000, pp. 956 to 970. doi.org
- Center for Creative Leadership, "Use Situation-Behavior-Impact (SBI) To Understand Intent." SBI is a trademark of the Center for Creative Leadership. ccl.org
- Shelley J. Correll and Caroline Simard, "Research: Vague Feedback Is Holding Women Back," Harvard Business Review, April 29, 2016. hbr.org
- Kieran Snyder, "The abrasiveness trap: High-achieving men and women are described differently in reviews," Fortune, August 26, 2014. fortune.com
Written by Tom Olajide, Founder. Last reviewed September 24, 2026.