The Managers page focuses on the people doing the reviewing rather than the people being reviewed. It’s the answer to “are our managers doing this well?”.
What this page answers
Which managers are writing high-quality reviews? How do their teams’ performance distributions compare? Are they rating consistently, harshly, or generously? Where are the calibration outliers?
What’s on the page
Three headline KPIs (top row)
Managers: count of managers with two or more direct reports in this cycle.
Avg Team Size: average direct report count across those managers.
Below Avg Quality: count of managers whose average review quality sits more than 0.10 below the org-wide average. These are the managers most likely to need writing support.
{Rating} distribution by manager
A heatmap with managers on one axis and the rating scale (or performance level) on the other. Each cell shows the count of that manager’s reports who landed at that rating. The pattern instantly shows you:
Managers with bimodal distributions (some great, some terrible) → calibrated
Managers everyone in one column → uncalibrated (all “Solid” is the most common pattern)
Managers with everyone in the top → likely over-rating
You can switch which rating drives the heatmap from the sidebar — pick AI Performance Level (default) or any custom rating dimension in the review.
Manager scorecard
A sortable table of every manager:
Manager: name (clickable — opens the manager’s employee detail popup on its Team tab)
Department
Team: team size
Avg Quality: mean review quality score across this manager’s reviews
Avg Sentiment: mean sentiment across their reviews
% Strong+: percentage of team rated Strong or Exceptional
Avg {Rating}: if you’ve picked a custom rating dimension, its average for this manager’s team
Where the data comes from
Every value on this page is rolled up from review extraction. Compass scores each review on quality, reviewer quality, sentiment, and performance level — the Managers page then groups by reviewer (manager) and aggregates.
The “Below Avg Quality” threshold is set at 0.10 below the org-wide average. It’s a relative measure: in a firm where reviews are generally strong, the bar to fall below average is high; in a firm where reviews are generally weak, you’ll see fewer below-average flags.
How to read it
Start with Below Avg Quality. That count is your shortlist of managers who’d benefit from training or templates.
Read the heatmap by row, not column. Each row is one manager’s distribution. A row that’s flat across all ratings is well-calibrated; a row that piles up in one column is not.
Compare team % Strong+. Differences across managers usually reflect both genuine team performance differences and rating tendencies — read alongside the heatmap to disentangle.
The sortable scorecard is your operational tool. Sort by Avg Quality ascending to find the managers needing the most coaching; sort by Team to identify managers whose load is too high or too low.
What “manager” means here
A manager appears on this page if they have two or more direct reports in the selected cycle. People with one direct report (or none) are excluded — there’s not enough signal to compare them meaningfully.
Tips
Manager quality scores reflect the writing, not the leader. A manager whose review-writing is poor isn’t necessarily a bad manager — but they probably do need help with the reviewing skill specifically.
Don’t use this page in isolation to judge people. It’s a starting point for a coaching conversation, not an evaluation.
Look for changes cycle-over-cycle. A manager who’s been below average for three cycles running needs different intervention than one who slipped once.