Feedback improves performance, except when it doesn't
Key takeaways
- Across 607 effects, feedback improved performance on average, but more than a third of feedback interventions made performance worse.1
- Feedback about the work helps; feedback about the person (praise, blame, comparisons) helps least.2
- In education research, specific, informative feedback had about four times the effect of rewards or punishment.3
- One person's rating tells you more about the rater than about the person rated, so combine several views.4
- Peer feedback works better when it is used for development, not for pay decisions.5
"More feedback" is one of the safest things a manager can say. It sounds modern, it sounds kind, and nobody argues with it. The research is less comfortable. Feedback does improve performance on average, but it also fails surprisingly often, and the way it fails is predictable.
A third of feedback made things worse
The study everyone in this field eventually runs into is Avraham Kluger and Angelo DeNisi's 1996 meta-analysis in Psychological Bulletin. They pooled 607 effect sizes from 131 papers, covering more than 12,000 people. The average effect of feedback on performance was positive and moderate (d = 0.41). But in more than a third of the cases, people who received feedback did worse than people who received none.1
Their explanation became known as feedback intervention theory: feedback works less well the more it pulls attention away from the task and towards the self. "Here is what went wrong in step three" keeps you on the work. "You're not a detail person" or "great job, superstar" makes you think about yourself, your image and your standing, and the task gets less of your attention.
A caveat worth stating: most of those studies were short tasks, often in labs, not colleagues giving each other feedback over months. The pattern is still the best starting point we have.
Information beats praise
Education researchers have studied feedback even more than organisational psychologists, and they land in the same place. John Hattie and Helen Timperley's review separates feedback about the task, about the process, about self-regulation and about the self, and finds feedback about the self to be the least effective. Mixing it in dilutes the useful part.2
A 2020 update by Benedikt Wisniewski, Klaus Zierer and Hattie pooled 435 studies. Feedback that carried a lot of information about the work had an effect of d = 0.99. Rewards or punishment had d = 0.24. Where feedback hurt motivation, it was mostly feedback that carried little information.3 These are classroom studies with a lot of variation between them, so treat the exact numbers loosely. The direction is very consistent: tell people something specific they can act on.
A single rating mostly measures the rater
The second problem is the score itself. In a study of more than 4,000 managers, each rated by bosses, peers, direct reports and themselves, Steven Scullen, Michael Mount and Maynard Goff found that the rater's own tendencies explained 53 to 62% of the variation in ratings. The manager's actual performance explained only 21 to 25%.4
That doesn't make ratings useless. It means a single rating from a single person is mostly noise about that person, and that averaging several raters, plus reading what they actually wrote, gets you much closer to the truth.
Peer feedback: small gains, bigger when it's for growth
What about feedback from colleagues specifically? James Smither, Manuel London and Richard Reilly reviewed 24 studies that tracked people before and after multisource (360-degree) feedback. Improvement was real but small, and on its own the peer-rating effect was close to zero (d = 0.05). The authors are blunt: don't expect large, widespread performance gains just from handing out feedback reports.5
Two findings in that paper matter more than the average:
- Improvement was about three times larger when feedback was used only for development (d = 0.25) than when it also fed into pay or promotion decisions (d = 0.08).
- People improved most when the feedback showed that change was needed, they reacted to it constructively, believed they could change, and set goals.
Other research points the same way. Simply asking for more feedback has only a small link to performance.6 And when peers compete for the same reward, their reviews of each other get less fair and agree less, at least in a well-known lab experiment.7
What this means in practice
Put together, the research describes feedback that helps:
- About the work. Rate skills and describe what you saw, not what kind of person someone is.
- Specific. A comment that says what happened beats a number, and a number beats a vague "good job".
- From several people. One rater is mostly noise; a pattern across colleagues is a signal.
- Developmental. Keep it separate from pay and from competition between colleagues.
- Regular, but not constant. In Smither's review, some ratings improved more when rounds were less than a year apart; nobody has shown that more is always better.
That list is also how we designed Reference Builder: colleagues rate a small set of skills the team chose, add a comment, and the app shows averages across raters rather than any one person's score. None of that guarantees results. It does keep feedback on the side of the research where it tends to help.
Sources
- Kluger, A. N., & DeNisi, A. (1996). The effects of feedback interventions on performance. Psychological Bulletin, 119(2), 254–284.
- Hattie, J., & Timperley, H. (2007). The power of feedback. Review of Educational Research, 77(1), 81–112.
- Wisniewski, B., Zierer, K., & Hattie, J. (2020). The power of feedback revisited. Frontiers in Psychology, 10, 3087.
- Scullen, S. E., Mount, M. K., & Goff, M. (2000). Understanding the latent structure of job performance ratings. Journal of Applied Psychology, 85(6), 956–970.
- Smither, J. W., London, M., & Reilly, R. R. (2005). Does performance improve following multisource feedback? Personnel Psychology, 58(1), 33–66.
- Anseel, F., Beatty, A. S., Shen, W., Lievens, F., & Sackett, P. R. (2015). How are we doing after 30 years? A meta-analytic review of feedback-seeking behavior. Journal of Management, 41(1), 318–348.
- Balietti, S., Goldstone, R. L., & Helbing, D. (2016). Peer review and competition in the Art Exhibition Game. PNAS, 113(30), 8414–8419.