This article reports on Cornell-led experimental research examining how the format of performance feedback — numerical-only, narrative-only, or a combination — affects employees' perceptions of fairness and motivation to improve. The study, published in the Academy of Management Discoveries, was conducted by Emily Zitek, Joonyoung Kim, and Caitlin Stroup, who designed four experiments involving 1,600 participants receiving identical evaluations in varied formats. The central finding is that narrative-only feedback was perceived as the fairest and most conducive to understanding how to improve future performance. Contrary to the researchers' initial hypothesis, combined feedback did not outperform narrative-only feedback. An exception was identified for highly positive feedback with mentions of monetary bonuses, which was well-received regardless of format. The authors also note that purely numerical feedback — even mid-range scores — tended to make recipients feel negatively evaluated. The article concludes by acknowledging practical trade-offs: while narrative-only reviews may improve fairness perceptions, numerical ratings retain administrative utility for decisions around bonuses and promotions. Key insights: Narrative-only performance feedback was perceived as fairer than numerical-only or combined formats across four experiments with 1,600 participants. Even mid-range numerical ratings can cause employees to feel negatively evaluated, reducing the perceived fairness of the review. Highly positive feedback with mentions of monetary bonuses was well-received regardless of format, suggesting format effects are moderated by feedback valence. Practical takeaways: Organisations weighing the removal of numerical ratings from performance reviews face a trade-off: narrative formats may improve fairness perceptions but reduce the administrative utility needed for bonus and promotion decisions. For employees receiving average or below-average performance feedback, narrative framing may reduce the experience of feeling negatively evaluated compared to numerical scoring alone.