Performance Review Templates: Questions and Rating Scales

A usable performance review template has nine sections, filled in the same way for everyone: role and period, goals and results, competencies, strengths, areas to develop, overall rating, development plan, employee comments and sign-off. Copy the version below, choose three to five questions per reviewer, and pick one of three rating scales.

The rating scale needs the most care. Steven Scullen, Michael Mount and Maynard Goff analysed two large samples of managers in 2000, each rated for development by two bosses, two peers, two subordinates and themselves, and found that rater quirks explained more of the variation in scores than performance.¹ No form removes that, so the template below pairs written anchors with a short practice session for the managers who use them. This page covers the form and its wording; the conversation itself belongs to the guide on how to hold a performance review meeting, one of the site’s talent development guides.

The performance review template, ready to copy

The template follows the cycle described by the US Office of Personnel Management (OPM): expectations set in advance, progress monitored, development planned, and a periodic rating that summarizes performance against what was agreed.² Paste it into a document and replace each bracketed prompt; sections 1, 2 and 3 are best filled in at the start of the period, when the goals are agreed.

A filled-in goal line shows the level of detail that works. A support team lead might write: “Goal: answer new tickets faster. Measure: median first reply under four working hours by June. Result: reached in May, held through September except during the August system change. Rating: 4.” Anyone reading that line later can see what was agreed, what happened and why the rating sits where it does. A line such as “Good progress on response times” gives the same reader nothing to check.

What should be included in a performance review template?

A performance review template should include who and what period it covers, results against agreed goals, a short set of competencies, strengths, areas to develop, a rating if one is used, a dated development plan, the employee’s comments and a sign-off. US OPM guidance asks that the elements and standards behind any rating be measurable, understandable, verifiable, equitable and achievable.²

Every section is there for a later reader, whether that is HR setting pay, the employee looking back or a new manager taking over the team. The purpose line in section 1 matters more than it looks. The UK’s Chartered Institute of Personnel and Development (CIPD) found in a 2016 evidence review that ratings in developmental appraisals tend to be stricter, and drawn from a wider range of examples, than ratings used for pay and promotion, and it recommends separating the two conversations.³ Saying which kind of review this is tells both people how candid the form can be.

Section 8 puts the employee’s own view on the record beside the manager’s. Guidance from the UK Information Commissioner’s Office (ICO) uses a performance review as its example: a manager’s opinion that someone is underperforming is not inaccurate just because the worker disagrees, and the manager could add a note recording that disagreement.⁴ A comments box makes that note routine. A designer who thinks a missed launch date was caused by late supplier files can say so on the record, beside the manager’s view.

Performance review questions for managers, employees and peers

Good performance review questions ask about specific work and behaviour in the period, one topic at a time. Fewer questions lose little, because ratings of separate qualities tend to move together: a 2005 meta-analysismeta-analysis: A study that combines the results of earlier studies on the same question into one overall estimate. Pooling makes the estimate more precise, but it cannot repair the studies it pools: a meta-analysis of surveys is still survey evidence.Full entry in the glossary by Chockalingam Viswesvaran, Frank Schmidt and Deniz Ones found that halo, a rater’s overall impression spilling across dimensions, substantially inflated the links between ratings of different dimensions.⁵

A form that asks for twelve traits mostly collects one impression twelve times. Three to five competencies, each with an example, give each answer a chance to say something new. Pick a few from each bank below.

Manager questions

• What did this person deliver against each goal, and what is the evidence?
• Which piece of work from this period best shows their strengths? What made it work?
• Where did a result fall short, and what got in the way?
• What did they do this period that they could not do a year ago?
• What one change in how they work would make the biggest difference next period?
• What support from me or the team would help most?

Self-evaluation questions

• Which result from this period are you most pleased with, and what made it possible?
• Which goal went less well than planned, and what would you do differently?
• What did you learn, and where have you used it?
• What got in the way of your work that I could change?
• Which parts of the job would you like to spend more time on next period, and which less?

Expect self-ratings to run higher than a manager’s. Self-ratings were more lenient than supervisors’ ratings in Western samples, according to Heike Heidemeier and Klaus Moser’s 2009 meta-analysis, and the size of that gap depended partly on the scale’s format and content.⁶ Questions that ask for examples, as these do, give both people the same facts to look at, so a disagreement turns into a discussion about a specific piece of work.

Peer questions

• What has it been like to work with this person on a shared project this period?
• What do they do that makes your work easier?
• What is one thing they could do differently that would help the team?
• Can you describe a specific moment when they handled something well, or less well?

Peers often see parts of the work a manager misses, such as how someone shares the load on a joint project. Ask them for examples rather than scores, and label every comment by who gave it, so the manager can weigh one colleague’s view as one view.

Three rating scale options

The three common choices are a 3-point scale, a 5-point scale with behavioural anchors, and no score at all with a written summary instead. Writing for SHRM in 2014, Stephen Miller reported the consultant Dick Grote’s view that 3-tier and 5-tier scales are the most prevalent.⁷ Choose by what the rating will be used for.

What a 5-point rating scale means

A 5-point rating scale in a performance review places each person on one of five ordered levels, usually from unacceptable to outstanding, with level 3 as the standard expected of someone doing the job well. Current US rules for federal agencies use that structure: level 1 is Unacceptable, level 3 is Fully Successful or equivalent, and level 5 is Outstanding or equivalent.⁸

The middle level causes the most trouble. Grote warned that employees given a 3 can hear mediocrity, as if it were a C grade at school, and suggested defining a 3 as a good, solid performer.⁷ Write the label and the description so a 3 reads as success.

Myth

A 3 on a 5-point scale is an average, middling score.

Fact

On a well-written scale, 3 describes someone doing the job fully and well. Label it that way so employees do not read it as a C grade.

The no-score option follows a trend CIPD describes: less focus on ratings and lengthy processes, more on high-quality conversations about current challenges.⁹ Grote noted that organizations keep appraisals because pay and promotion decisions need the data, so an organization that drops scores has to record a judgment for those decisions somewhere else.⁷ Whether to drop ratings altogether is a debate with its own guide coming.

How to write anchors that mean the same thing to every manager

An anchor is a short description of what performance at each point of the scale looks like in this job, written before anyone is rated. The best-known method, the behaviorally anchored rating scale (BARS), is usually traced to a 1963 paper by Patricia Smith and Lorne Kendall; it builds each anchor from real examples of effective and ineffective work collected from people who know the job.¹⁰¹¹

“Good communication” means one thing to a manager who prizes brevity and another to one who prizes detail. A description such as “sends the weekly client update on time, with risks flagged and a next step for each” gives both the same picture.

OPM’s guidance on writing standards offers a practical test: decide whether quality, quantity, timeliness or cost matters for each element, and where a number fits, define the number.¹² Where performance can only be described, name who is best placed to judge it and what they would look for.¹²

Anchors alone are not a cure. Paul Kingstrom and Alan Bass reviewed all the published studies comparing BARS with other formats in 1981 and found the formats differed relatively little, if at all, on leniency, halo and agreement between raters, though methodological problems made firm conclusions difficult.¹³

1. Bare numbers: an illustration, not data; with no description, each manager decides alone what a 2 or a 4 means
2. Anchored and discussed: written descriptions plus practice on examples; in a 2012 meta-analysis, training raters in a shared standard improved rating accuracy

The wording on the page seems to matter less than whether managers share the same picture of each level, and that is something a team can practise.

The study

Moderate evidence

A shared standard, taught: the 2012 update on frame-of-reference training

Sylvia Roch and colleagues updated the research on frame-of-reference training, which teaches raters to judge performance against a common standard. With over four times as many studies as the previous meta-analysis, they found it improved rating accuracy, most of all on two of the accuracy measures studied, one of them accuracy about behaviours.¹⁴

The method can be scaled down for a team of managers. In one US medical-education version, assessors rated several recordings of the same task done at different skill levels on their own, then compared ratings as a group and discussed where they differed; its authors note that the gains from such training in medical education have varied.¹⁵ A manager team can do the same with three anonymized examples of real work before review season: rate them alone, compare, and rewrite any anchor that produced a spread.

Before you use the scale

• Three to five competencies chosen for this role
• An anchor written for every level, describing what someone does
• Level 3 worded as the standard for doing the job well
• Anchors tested on two or three real examples by several managers
• Any anchor that produced different ratings rewritten

How do you write a performance review for an employee?

Write a performance review section by section from evidence gathered through the period: the result against each goal first, then one dated example for each competency rating, then strengths, areas to develop and a development plan with owners and dates. Leave section 8 for the employee and read their self-review before you finalize your own.

Each comment works best with three parts: what happened, its effect, and what to keep or change. “The March supplier report arrived a week late, which delayed the pricing decision; next quarter, flag a slip by the Wednesday before” gives the person something to act on. The site’s guide to the situation-behavior-impact feedback model shows the same structure for spoken feedback, and how to set SMART goals helps with the development plan. The same anchoring logic runs through structured interviews, where answers are scored against descriptions agreed in advance.

Records, access and the rules where you work

A completed review is a record about a person, and in many countries data protection law applies to it. In the EU, the 2016 GDPRGDPR: The General Data Protection Regulation, the European Union's main data protection law, applied since 2018. It sets rules for how organizations collect and use personal data about people in the EU, including organizations based elsewhere, and gives people rights over that data.Full entry in the glossary gives people the right to access personal data held about them, and lets member states set more specific rules for employment.¹⁶ The UK ICO, in guidance it is reviewing after the UK’s Data (Use and Access) Act, lists appraisal and performance review records among employment records and says workers can request a copy of their information through a subject access request.⁴

Write every comment as if the employee will read it, because they may. Rules on how long to keep reviews, who may see them and whether they can be recorded differ by country and sometimes by collective agreement, so check with HR, your national data protection authority or an employment lawyer where you work.

The bottom line

Managers are most likely to read the rating scale in different ways, so give it the most preparation. Before the next review cycle, cut the competency list to three to five, write an anchor for every level of your scale, test it with other managers on a few real examples, and leave room for the employee’s own words.