People & HR
Employee performance review
Six competencies on one shared four-point scale, two open questions, and a single overall rating.
Why this survey gets sent
“Review season means eleven documents, each a different shape, and by the third one I am rating people against my memory of the second. Then somebody asks me to compare the team and I cannot — I never asked them the same questions.”
A review is only comparable when every person is judged on the same list against the same scale. That is what the grid does: six competencies and one four-point scale, read once and applied six times, so the manager makes six judgements instead of six decisions about what the scale means. The two open questions then carry everything a grid cannot — and they come after it, once the manager has been forced to be specific.
What is in it
- Employee name, department and review date
- Six competencies in one grid: quality, deadlines, communication, teamwork, initiative, reliability
- A four-point scale, from needs improvement to outstanding
- What went particularly well this period (long answer)
- Where more support would help (long answer)
- One overall rating for the period
About 4 min to answer
Employee performance review
About 4 min
How to write this survey
Write competencies as behaviour you can observe
The six competencies here — quality of work, meeting deadlines, communication, teamwork, initiative and reliability — are written as things a manager can remember examples of. That is the only useful test when choosing them: if you cannot attach the rating to something that actually happened during the period, the line is written in the language of adjectives rather than behaviour, and it needs rewriting or removing.
Beware of lines that describe the employee's character rather than their work. “Enthusiasm”, “loyalty” and “a positive personality” cannot be measured, and rating them ends up as the manager's impression of the person. Replace them with behaviour: “proposes solutions before being asked” instead of a vague initiative, “flags a delay early” instead of an abstract reliability — if your team needs that level of detail. The short labels in this template work when everyone agrees in advance what they mean; when they do not, spell the line out in full.
Four points, every one labelled, and no middle
The scale here has four points: needs improvement, meets expectations, exceeds expectations, outstanding. The even number is deliberate. A scale with a middle collects everybody the manager has not made their mind up about into a single column, and the review comes back as a row of middling grades that says nothing. Four points force a lean one way or the other, which is what a review is for.
Label every column with words rather than numbers. A 3 out of 4 means “very good” to one manager and “slightly short of the mark” to another, while “exceeds expectations” means the same thing to both. Then keep the same wording from one period to the next: the whole value of a review is in comparing it with the last one, and changing the column labels erases that comparison without warning you.
Two open questions, not ten
After the grid come exactly two questions: what went particularly well this period, and where more support would help. The first forces the manager to name a specific event instead of offering a general impression, and it is the part the employee will read and remember. The second is a deliberate replacement for “weaknesses”: the same wording moves the conversation from a flaw in the person to a gap in the support, and it produces answers you can act on — training, a tool, a different split of the work.
Their position after the grid, not before it, is part of the design. A manager who has just rated six competencies has been forced to be specific, so what they write is built on the judgements they have just made rather than on a general impression formed before thinking. And two is the limit: every extra essay question lowers the chance that the manager writes anything useful in any of them.
Rate the period, not last week
The biggest fault in performance reviews is letting the last month outweigh the whole period. The review date field is not decoration: write the period the review covers into the survey's opening paragraph, and ask the manager to recall examples from its beginning as well as its end before they start. Anyone who keeps short monthly notes writes a more accurate review in less time.
Use the same template for everybody on the team in the same period. The real value of a shared grid is that it makes comparison possible — who needs training in communication, who can be relied on for a critical task — and that comparison is destroyed entirely if half the team is rated on one list and two extra lines are added for the other half.
The overall rating is a judgement, not an average
The last question is a single overall rating for the period, and its place at the end is deliberate. It is not the arithmetic mean of the six competencies: somebody average on five lines and outstanding on the one their job rests upon is not an average employee. Leave it as an explicit judgement by the manager, and let the grid explain that judgement rather than produce it.
Whatever the number, the review should never be the first time the employee hears any of this. The survey documents a conversation that happened during the period; it does not replace it. Agree in advance who sees the results and what will be built on them, and say so in the opening paragraph — this template's intro states outright that the results are used to plan support and development rather than to rank people against each other, and that one sentence changes how it is answered.