# Role Rating Sheet **Version:** 1.0 **Date:** 2026-08-22 **Type:** Instrument — a decomposable 1–10 rating, one page per person per period --- ## 1. What this is A single number, one to ten, for how someone is doing in their role — built the way a football coach rates a player after a match. One number, so it can be compared and tracked. But **the number is always shown with the parts it was made from**, because a score nobody can take apart is not a judgement, it is an opinion wearing a number. This sheet is filled by the person's manager, on a fixed schedule, and it is shown to the person it is about. If it cannot be shown to them, it is not usable here. --- ## 2. The rule that makes this work > **The overall is never written on its own.** Wherever the 1–10 appears — a file, a conversation, a spreadsheet — it appears with the component scores that produced it. A 7 is not a 7. A 7 is *"7 = reliability 9, guest handling 7, pace 6, standard 6."* This is the whole design. A rating that can be taken apart can be argued with, and a rating that can be argued with is one people will act on. A rating that cannot be taken apart gets treated as a verdict about a person, and defended or resented instead of used. Two consequences, both non-negotiable: - **The scale is 1 to 10 and nothing leaves it.** No 11, no "off the charts", no 9.5 written in as a favour. One number outside the scale makes every other number on the page unreliable, and people notice immediately. - **A component with no evidence is left blank, not guessed.** A blank component is visible in the arithmetic and forces the question "what would I have needed to see?" A guessed component disappears into the total and cannot be recovered. --- ## 3. Components are weighted BY ROLE, not the same for everyone A striker is judged on goals; a defender on the same scale would rate poorly for reasons that have nothing to do with how good a defender they are. The same trap applies here: rating a room attendant and a front desk agent on one common list produces a number that mostly measures which job someone has. **So each role carries its own weights, set once, written on the sheet, and the same for everyone in that role.** Weights must total 100. | Component | What it means | Weight | |---|---|---| | Reliability | turns up, on time, on the days agreed | ___ | | Output | the work actually completed in the time worked | ___ | | Standard | the work holds up when someone checks it later | ___ | | Guest handling | what happens when a guest is in front of them | ___ | | Independence | how much of the shift runs without being directed | ___ | | Teamwork | what happens to everyone else's shift because of them | ___ | | **Total** | | **100** | **Set the weights before rating anybody**, and date them. Weights chosen after seeing the people are weights shaped around the people. ⚠ **The weights are yours to set and they are not a formality.** If one component is weighted so heavily that nothing else can move the total, you have not built a rating, you have built that component with extra steps. If every component is weighted equally, you have said nothing about what the role is for. --- ## 4. The sheet **Role:** ______________ **Period:** __________ **Rated by:** __________ **Date:** ______ | Component | Weight | Score 1–10 | Weighted | Evidence — what was actually seen | |---|---|---|---|---| | Reliability | | | | | | Output | | | | | | Standard | | | | | | Guest handling | | | | | | Independence | | | | | | Teamwork | | | | | | **Overall** | **100** | | **___ / 10** | | **One thing to keep doing:** _______________________________________ **One thing that moves the lowest component up one point:** _______________________ **Next review:** __________ **Seen and discussed with the person:** Y / N, date ______ --- ## 5. What makes a dishonest sheet visible **The arithmetic closes.** Weighted scores must sum to the overall, and the weights must sum to 100. An overall written first and components fitted around it rarely closes, and the failure is visible on the page to anyone who adds it up. **Every score carries evidence, or the score is blank.** The evidence column is not a justification written afterwards — it is what was seen. A row with a score and an empty evidence cell is an unfinished row, exactly like an unmarked line on a stock sheet: it reads as "nobody looked", never as "fine". **The lowest component drives the next-step line.** The form does not let a manager write a development note about a strength. It must address the lowest-scoring component, because that is the one the total is losing points to. A rating that produces no specific next step has told the person nothing they can act on. **The same person, rated by the same manager, at the same score, for three periods running, with no change in the next-step line** — that pattern is the form telling you the review has become a ritual. Nothing is wrong with the person; the process has stopped doing anything. --- ## 6. What this is for, beyond the review Once several periods exist, the sheet answers questions no individual review can: - **Which components does this property score badly on across everyone?** That is not a people problem. That is training, staffing levels, or equipment. - **Which components predict someone staying?** Worth knowing before hiring for them. - **What does a genuinely good person in this role look like, component by component?** That profile is what a role requirement sheet should be built from — and it is measured from your own people rather than imagined. That last one is the real payoff, and it only works if the components were honest when nobody was looking for a pattern in them. --- ## 7. Before first use 1. **Are the weights written down and dated before anyone is rated?** If not, stop here. 2. **Can the person doing the rating actually observe each component?** A manager who never works the same shift cannot score reliability from evidence, only from report. 3. **Will the person see their own sheet?** If the answer is no, the evidence column will quietly become optional and the instrument is finished. 4. **What happens to a low score?** If the answer is nothing, ratings will drift upward within two cycles, because there is no cost to generosity and a real cost to honesty. 5. **Employment law is jurisdiction-dependent.** Anything used in a pay, promotion or termination decision has local requirements this sheet does not address. Check them. --- ## What this instrument does not do It does not measure potential, attitude, or fit — those are not observable and putting them on a scored form makes bias look like arithmetic. It does not compare people across different roles: two overalls built from different weights are different measurements that happen to share a scale, and treating them as one is a category error. It cannot fix a manager who does not observe. It makes that visible in the evidence column, which is as far as any form can go. The sheet publishes blank. Every property's filled sheets are its own.