What can I learn from this page?
The science behind Culture Amp's performance review templates
Who is this page for?
A group of Culture Amp’s People Scientists revamped our performance review templates in early 2021. This article includes:
Template design is an important part of the performance review process because it creates the experience your employees have and can reinforce key values and behaviors. If done well, it can also mitigate the common forms of bias that affect performance reviews and have become more common with employees working remotely. 2020 demonstrated the importance of taking the whole employee into account - their well-being, growth and feelings of inclusion in addition to evaluating performance. The templates have been redesigned with this in mind, along with incorporating new research, and our findings of how customers have used previous questions.
Our performance templates are a foundation to start from as we recognize that each organization has a unique performance management philosophy. There are many parts of your culture that you may want to incorporate like values, expectations, and specific competencies. In the course linked above, you’ll find a question bank that you can draw from. If you choose to develop your own questions, use the following tenets of design for the best result.
Tenets of design
Start with the positive. This helps employees be less defensive when confronted with constructive feedback later on.
Look to the past as well as the future. While evaluating performance is important for making important decisions like promotions and raises, giving future-focused feedback on how employees can do things differently leads to the ultimate goal of improving performance and developing their skills.
The more room for interpretation, the more room for bias. Where possible, focus on objective measures like metrics and goal attainment, over subjective measures like intentions. Guiding managers to review specific, objective things (like goals) results in less bias than having an open box which is filled with the reviewer’s natural biases.
Overall our templates intend to create a sense of progress, because progress has been found to be the single most effective thing in boosting emotions and motivations in a workday.
Rationale behind each component
Our templates include not only the manager review but also the other components we consider instrumental in a fair and equitable performance process.
A self-reflection: so employees can provide their viewpoint on their performance. Studies have found that:
Giving the manager an understanding of the employee’s own perceptions of their performance has been found to be a catalyst for deeper conversations about performance issues, with employees feeling more motivated to improve.
However, keep in mind the lake wobegon effect (that employees’ perceptions of their own performance is often inflated, though women, minorities, and those from collectivistic cultures are less likely to brag). And train managers not to change their rating based on an employee's inflated self assessment.
Peer & Upward Feedback: so the manager has a more complete view of the employee’s performance from the individual’s direct reports and peers. Studies have shown that:
Peers and subordinates can account for objective performance measures over and above managers
Since the evaluation is based on more than one individual’s feedback, it is less susceptible to idiosyncratic biases.
Rating scale recommendations
We’ve written extensively about the pros and cons of using performance ratings elsewhere, instead this section focuses on how to design your rating scale. Our data scientists recently undertook a study to identify what rating scale has the best utility for organizations by looking at the results of over 150 performance review cycles.
We found that for most companies a 4 point scale is the best choice because:
Removing the middle option eliminates central tendency bias (the instinct to select the middle option)
Having 2 options for high performance allows for differentiation
When using a 4 point scale, 1 should be associated with low performance, 2 with competent performance, and 3 & 4 with high performance.If you have a problem with leniency, meaning everyone receives a high rating, you could instead have 5 ratings to extend the high performance options to 3.
The anchors are equally as important as the number of options. We recommend the following anchors:
“Consistently Meets Expectations”
“Often Exceeds Expectations”
“Sets a New Standard”
Don't forget to add your own description!
We recommend avoiding:
Terms like Neutral, Average or Normal because they are difficult to define and assess. Our research found that the use of a ‘middle ground’ or neutral weighting impacted the spread of the scores.
References to compensation (e.g. Award), promotion (e.g. Mover) or increase in seniority/ level within the organization (e.g. Over performer). These labels also distorted the distribution.
Q: Which questions should be shared with the reviewee?
A: We recommend directly sharing qualitative feedback that will help the employee learn and grow, or create a sense of accomplishment. While rating feedback is better shared through a conversation. This is because rating feedback, or any that places employees into categories, puts employees into a fight or flight response. In turn making them less likely to digest the useful feedback, as well as more likely to get into a fixed mindset (“I am a 2 and always will be”). By sharing qualitative feedback, you can instead answer the growth mindset question of “what can I do to improve?”
Q: I’ve been using Culture Amp’s previous templates, should I switch?
A: The previous templates are not ‘wrong’ in any way. If those items are working for you, feel free to stick to them! However, if you’re looking to iterate on your approach, the new templates would be a good place to start. If you keep using the previous templates, we do recommend making sure your scales align to the new research.
Q: How long should the self-reflection/review/peer feedback template be?
A: This depends on the number of components in your process, and the average number of reviews/feedback processes each person is completing.While open-ended questions take longer to complete, they are more useful to the individual receiving feedback. In general, we recommend keeping open-ended questions to around 3-4, and capping it at 5. However, the best way to know if it’s too long is to test it out yourself - think of someone to write a review for and see how long your drafted template takes for you to complete.