The Managers view answers a question a single cycle can’t: is a manager’s read on their team stable? One cycle tells you how a manager reviewed; several cycles tell you whether their standard holds — or whether their assessments swing between rounds, which is the signature of calibration drift.
It needs at least two cycles selected, with AI extraction complete on each, and only includes managers who have team review data in two or more of the selected cycles.
What this page answers
Which managers produce consistent team review quality cycle after cycle? Whose numbers swing the most between cycles? Whose teams are accelerating, and whose are declining?
What’s on the page
Three headline KPIs (top row)
Managers — How many managers have team data in at least two of the selected cycles.
Avg Team Quality — The mean of every manager’s average team review quality across the comparison.
Most Consistent — The manager whose team quality varied least between cycles.
Manager consistency
A sortable table with one row per manager, sorted by consistency — lowest variance first:
Manager — Click to open the employee detail flyout for the manager themselves
Department
Cycles — How many of the selected cycles this manager has team data in
Avg Team — Average team size across those cycles
Avg Quality — Average review quality across their team’s reviews, averaged over the cycles
Avg Sentiment — The same, for sentiment
Quality Variance — How much their team’s average quality fluctuated between cycles (a standard deviation — lower is steadier)
% Accelerating — The share of their team’s reviews, across all the cycles, where the employee’s trajectory reads as accelerating
% Declining — The same for declining
Where the data comes from
The page aggregates each manager’s per-cycle team metrics from the stored review extraction results — the same manager effectiveness numbers you see on the Reviews tab’s Managers page, computed for every selected cycle and then combined. Managers present in only one of the selected cycles are excluded, because consistency needs at least two data points.
How to read it
The top of the table is your calibration benchmark. Managers with low variance and healthy averages are assessing steadily — useful reference points when calibrating others.
High variance is a question, not a verdict. A manager whose team quality jumps around might be inconsistent — or their team changed substantially, or one cycle asked much more of reviewers. Check their team’s composition across the cycles before concluding.
Read % Accelerating and % Declining together with Avg Quality. A manager with high declining share and low review quality may simply be writing thin reviews that read as negative — a writing problem, not a team problem.
Sort by Avg Sentiment to find tonal outliers. A manager far more negative than peers across multiple cycles is a consistent-but-harsh reviewer; that consistency is exactly what single-cycle views can’t distinguish from a one-off bad round.
Tips
Small teams make noisy numbers. Use the Avg Team column as a confidence check — variance on a three-person team means much less than on a fifteen-person team.
Filter by department to compare like with like. Managers reviewing similar roles are the fairest comparison set.
Follow up on the Reviews tab. Once a manager stands out here, the Reviews tab’s Managers page shows their single-cycle detail — rating distribution, language patterns, and team specifics for each cycle in turn.