# Evaluating people fairly > Keep a fact log, spot halo, recency and similar-to-me bias, use the same criteria for everyone and ask for a second opinion. - Course: [From Specialist to Manager: Your First Year Leading a Team](https://productovi.com/courses/from-specialist-to-manager.md) - Module: Growing people, and a team that runs without you - Lesson 30/32, ~24 min - Source: https://productovi.com/courses/from-specialist-to-manager/evaluating-people-fairly ## The review season opens **Iryna:** Review season. By 31 August, one page per coordinator on the same form: results, work with the team, a plan for next year. I'll read all four. **Sofia:** Taras is the easy one. In July, when the payment page broke, he carried the queue. Leo thanked him in the company chat. **Alex Tutor:** Two biases are already talking. The **halo effect**: one strong impression colors everything else you think about a person. The **recency effect**: what happened last weighs more than what happened for months. **Sofia:** But the week was real. **Alex Tutor:** Real, and it belongs in his review. The question is how much of the review it should be. > **Exercise:** Since Leo publicly thanked Taras for the July week, you've started to think of Taras as the most careful coordinator on the team too. Which bias is this? ## One week out of thirty **Dataset (CSV):** [Klasna Support in July 2027, by week and coordinator](https://productovi.com/datasets/from-specialist-to-manager/evaluating-people-fairly/klasna-review-season-july-weeks.csv) Four July weeks, four coordinators. Two are on shift each week while the other two are on vacation. - `week_start`: Monday the week starts - `person`: Coordinator - `on_shift`: yes if the person worked that week, no if on vacation - `tickets_closed`: Tickets the person closed - `reopened`: Closed tickets the family reopened - `csat_responses`: Post-ticket survey answers - `csat_positive`: Positive survey answers **Alex Tutor:** Taras worked the weeks of 5 and 12 July. The second is the week of the payment-page fault. **Alex Tutor:** Work out his reopen rate for the week of 12 July: reopened tickets over tickets closed. **Sofia:** And then compare it with his other July week. **Alex Tutor:** Good habit: never judge a week on its own. > **Exercise:** Use the dataset. What share of the tickets Taras closed in the week of 12 July was reopened? Give a percentage to two decimals. **Alex Tutor:** 1.32%, and 24 of 25 survey answers positive. A great week. Now January to June: 9.51% reopened, 78.10% positive on 790 answers, above 8% every month. And 5 July: 9.68%. **Sofia:** So the week is the exception, not a new normal. **Alex Tutor:** One week out of thirty, 5.37% of his tickets this year. Counted in, it moves his seven-month reopen rate from 9.51% to 9.07%. And 25 survey answers is a small base. **Alex Tutor:** One more check on that week's numbers. > **Exercise:** In the week of 12 July, 24 of Taras's 25 survey answers were positive (96%). From January to June it was 78.10% of 790 answers. Why should the 96% carry **little weight** in his review? **Sofia:** Then where does the week go? **Alex Tutor:** Into the review, as a fact: how he handled a flood. It shows something real about him under pressure. It just can't replace six months of evidence. > **Exercise:** Taras's week of 12 July shows that his quality has improved, so his review should rest mainly on July. ## A fact log, not a memory **Alex Tutor:** Memory keeps the dramatic and the recent. A **fact log** keeps the rest: dated, observable facts and their impact, good and bad, for every person, across the whole period. **Alex Tutor:** One line is enough: "12 July, Taras: 152 tickets during the payment-page fault, 2 reopened." Not "Taras is a hero". **Sofia:** I haven't kept one since January. Is it too late? **Alex Tutor:** Rebuild what you can from the monthly facts, the incident log and your one-to-one notes. Then start one for next year: five minutes each Friday. **Alex Tutor:** And no hidden tracking. The log holds work facts the person could see and discuss, never anything about private life. > **Exercise:** Which of these is a good fact log entry? ## Like me, and one ruler for all **Alex Tutor:** Third bias: **similar-to-me**. We rate higher the people who work the way we do. **Sofia:** I was the fast one who grabbed hard tickets. Taras is fast too. Bohdan takes about 35 minutes to reply and argues with the AI reply assistant. **Alex Tutor:** And Bohdan logged the most mentoring hours this year, 36, while the seasonal hires learned the job. Quiet Maryna has the fewest reopens. None of it looks like you; all of it counts. **Alex Tutor:** Be honest with yourself here. > **Exercise:** **True or false:** a lead who was once the fast coordinator who grabbed hard tickets will tend to rate Taras, who is also fast, above Bohdan, who replies in about 35 minutes, before looking at any numbers. **Alex Tutor:** So write the criteria before you open anyone's numbers: quality for families, volume, help to the team, reliability. Same ruler, same period, for all four. **Sofia:** And who checks me? **Alex Tutor:** A second opinion. Ask each person for a short self-assessment first. Then Iryna reads your four drafts and asks, line by line: what's the fact? > **Exercise:** Which steps make the four evaluations fairer? Select all that apply. **Alex Tutor:** Last one before you write: what's behind Bohdan's slower replies? > **Exercise:** Bohdan's first replies are slower than Taras's. Before you write "slow" in his review, what should you check first? ## Your first line **Iryna:** Drafts by the 25th, please. I'll push back on any adjective without a fact. **Alex Tutor:** Start with the line that's hardest to write fairly. Here's the one you wrote on Friday evening. **Sofia:** "Taras was our star this summer and is the best coordinator on the team." **Alex Tutor:** Rewrite it so that Taras, Iryna and the other three would all find it fair. > **Exercise:** Rewrite your draft line, "Taras was our star this summer and is the best coordinator on the team", as two or three fair, fact-based sentences for his review. Facts: in the week of the 12 July payment-page fault he closed 152 tickets with 2 reopened; from January to July his reopen rate was 9.07%, and January to June 78.10% of 790 survey answers were positive. > **Key ideas** > > - Halo effect: one strong impression colors your whole view of a person. Recency effect: the last weeks outweigh the months before them. Similar-to-me: people who work like you get rated higher. > - One great week is a fact for the review, not the review: Taras's 12 July week (1.32% reopened, 25 survey answers) is 1 week of 30 and moves his seven-month reopen rate only from 9.51% to 9.07%. > - Keep a fact log all year: dated, observable work facts and their impact, good and bad, for everyone. No hidden tracking, nothing private. > - Write the criteria before you look at anyone's numbers, and use the same ones, over the same period, for every person. > - Ask for a self-assessment and a second opinion: every adjective needs a fact behind it. **Alex Tutor:** An evaluation looks back. Next, the part that looks forward: growth plans, and a team that runs while you're away. - Previous lesson: [Coaching conversations](https://productovi.com/courses/from-specialist-to-manager/coaching-conversations.md) - Next lesson: [Development plans, handover and being away](https://productovi.com/courses/from-specialist-to-manager/development-and-handover.md) --- Productovi Learning. Free, no sign-up needed to read.