Performance Review Manager

What this tool does

CM8-269 runs a structured performance review cycle. For each person you record the objectives that were agreed, the results against them as facts, competency ratings, development goals, specific feedback and an overall summary — each with the employee's own comment alongside — then track the review from draft through shared, discussed and signed, and print a report for the file.

A review exists for two reasons: alignment — do this person and this organisation agree on what was expected and what happened? — and growth — what changes next period? Everything in the tool serves one of those two. What a review is not is the moment performance is discussed for the first time. If anything in the completed form comes as a surprise to the employee, the failure happened months earlier, in the conversations that were never had. A review summarises a period of conversations; it cannot replace them.

Everything runs inside this single file. There is no account, no upload and no network request of any kind. That matters here more than in most tools: a review contains candid judgements about named people.

How the register is organised

One row is one review item for one person in one cycle. A complete review for one person is typically four to eight rows:

• Objectives — the outcomes agreed for the period, each with a weight and a rating.
• Competencies — behaviours and skills rated on the same scale.
• Development goals — growth actions, usually left unrated.
• Feedback notes — specific observations, good or bad, with evidence.
• One summary — the overall rating and the paragraph that stands for the whole review.

Name cycles consistently — "2026 mid-year", "2026 year-end" — and give every row in a cycle the same cycle text. The tiles and the overview table read the latest cycle, which the tool takes to be the one containing the most recent review-meeting date.

Objectives that can be rated honestly

An objective that cannot be measured cannot be rated — it can only be argued about. "Improve quality" is a hope. "Reduce scrap from 4% to 2.5%" is an objective: at the review, the number either moved or it did not, and the conversation is about why.

Record the result in the detail field as a fact, before any rating: "finished the half at 2.8% against the 2.5% target." Facts first, judgement second. When the result is written down plainly, most rating arguments disappear, because both people are looking at the same sentence.

Weights (1–5) let you say that objectives are not equal. The scrap target that consumed half the year carries a 3; the tidy-the-shared-drive objective carries a 1. The weighted score then reflects what the period was actually about, rather than averaging the important with the trivial.

The rating scale — and the everyone-is-a-3 trap

The scale is anchored so that two managers using the same words argue less than two managers guessing what a number means:

• Rating — Anchor
• 1 — Not met — a significant gap against what was agreed
• 2 — Partially met — progress, but short of the agreement; developing
• 3 — Met — did what was agreed, solidly. This is a good rating
• 4 — Exceeded — clearly beyond the agreement
• 5 — Far exceeded — exceptional, and rare by definition

Two disciplines keep the scale honest. First, extreme ratings need evidence: the tool will not save a 1 or a 5 without something in the detail field, because "what happened?" is the first question anyone reviewing the file will ask. Second, watch the honesty curve — the rating distribution chart. Honest use of a five-point scale produces a spread. When almost every rating is a 3, the scale has collapsed into "fine", nobody learns anything, and the exercise has become appraisal theatre. The distribution tile flags when the share of 3s passes your threshold. A 3 should be a genuinely good rating — "did what we agreed" — not a euphemism for "I didn't want the conversation".

Competencies — the how

Objectives measure what was delivered; competencies rate how — leadership, safety behaviour, customer handling, craftsmanship, whatever your organisation genuinely values. Rate them on the same scale, and hold them to the same evidence standard: "team leadership: 4" means nothing, but "ran every shift handover, cross-trained two operators, covered absence without overtime" is a rating anyone can stand behind. Someone can hit every number in a way that damages the team; someone can miss a target while doing everything right in a bad market. Recording both dimensions separately is what stops one story hiding the other.

Development goals that are real

"Improve communication skills" is not a development goal; it is a horoscope. A real development goal is an action with a date and, where needed, money or cover behind it: "internal auditor course in October, then co-audit two audits", "shadow the service desk one day a month". Development goals are usually left unrated — the point is that the action happens, and the next cycle checks that it did. The development tile counts goals and the share of people who have at least one; a cycle where half the team has no development goal is a cycle that looked backwards only.

The employee's voice

Every item carries an employee comment field. Use it verbatim — their words, not your summary of their words. Agreement is worth recording; disagreement is worth recording even more, because a file that contains "I don't accept this rating and here is why" is a fairer and more defensible file than one that pretends consensus.

The status flow enforces the discipline: draft → shared → discussed → signed. Share the written review before the meeting, not at it — people cannot respond thoughtfully to a document they are reading for the first time in front of the person who wrote it. "Signed" means both people acknowledge that the document records the conversation that took place — not necessarily that they agree with every word.

The meeting

Evidence first, ratings second. Walk through what happened — the results, the observations, the employee's comments — before any number is discussed. When the meeting starts with the rating, the rest of the hour is a negotiation; when it starts with the facts, the rating usually settles itself.

Record the meeting date on the items, capture anything new that surfaced, agree the next check-in, and move the items to signed. The signature matters for one reason only: both people saw the same document, at the same time, and said so. That is what makes the record worth keeping.

Check-ins between cycles

Six or twelve months is far too long between conversations. Set a next check-in date on the summary — monthly or quarterly for most teams — and use it for a short, unscored conversation: how are the objectives moving, is the development goal actually happening, what has changed? After each check-in, move the date forward. The overdue tile counts check-in dates that have passed; the "needs attention" table names them. Regular check-ins are what make the year-end review a summary instead of an ambush.

Fairness and consistency

A rating is only meaningful if the same performance gets the same rating regardless of who is being rated and who is doing the rating. Use the same anchors for everyone; judge against what was agreed, not against the most memorable fortnight; and calibrate across the team with the distribution chart — if one manager's people are all 4s and another's are all 2s, the difference is usually in the managers, not the people. Where ratings feed pay decisions, consistency stops being a courtesy and becomes something you may have to defend.

One paragraph on confidentiality, because it matters: this file contains frank judgements about named people, and their own words. It lives in this browser's local storage and in any export you make — nothing is uploaded anywhere. Treat exports as confidential personnel records: store them where personnel files are stored, share them only with people who would be allowed to open the paper file, and remember that in many countries employees have a legal right to see what is written about them. Write every field as if the person will read it, because they may — and because they should be able to.

How the numbers are worked out

Unrated items are excluded from every average rather than counted as zero — a development goal without a rating is normal, not a failure. Where there is not yet enough data for a figure to mean anything, the tool shows "—" instead of a confident wrong number.

FAQ

How many objectives should one person have? Three to five, weighted. More than that and none of them is really an objective; the weights stop meaning anything because everything is a priority.

Should development goals be rated? Usually not in the cycle where they are set. Rate the outcome next cycle if you need to — this period, what matters is whether the action happened.

What if the employee refuses to sign? Record their objection verbatim in the comment field and leave the status at discussed. A documented disagreement is a legitimate outcome; a signature extracted under protest is worth less than no signature.

Can I run mid-year and year-end in the same register? Yes — use a distinct cycle name for each. The tiles and overview read the latest cycle; the full register keeps the history, which is exactly what you want when the year-end conversation asks what changed since June.

Someone's ratings are all 3s — is that a problem? For one person, possibly not; solid is solid. Across a whole team, almost certainly — check the honesty curve, and ask whether the 3 has become the rating that avoids conversations rather than the one that describes performance.

Saving your work

Review items, settings and the report header are written to this browser's local storage as you type, and the toolbar shows the time of the last save. That storage belongs to one browser on one computer: another browser, a private window, a second machine or a clean-up tool that clears site data will not have it.

Treat Export .json as the real save — one file containing everything, which Import .json restores anywhere. Export CSV gives you the register for spreadsheet work. Reset asks twice, then erases everything this tool has stored. There is no undo. These are personnel records — store exports accordingly.

Accuracy & disclaimer

The arithmetic here is simple and the tool does it faithfully. Everything that matters sits underneath it: whether the objectives were the right ones, whether the evidence is honest, and whether the same standards were applied to everyone. A rating is structured judgement, not a measurement, and comparing ratings between teams that used different anchors means nothing.

Employment-law requirements around performance documentation — what must be recorded, what the employee may see, how long records are kept, and what may lawfully feed a pay or dismissal decision — differ by country. This is a record-keeping and preparation aid, not legal advice, and a review is a record of management, not a substitute for it.