Review completion, counted honestly

Mid-cycle, leadership asks how reviews are going, and the delivered answer is a percentage with no denominator and no names. The honest tracker partitions the eligible population and tells you who is stuck at which step.

RuleTrack the cycle from process events: every eligible worker in exactly one status, statuses summing to the total, stale steps named.

Neverreport a completion percentage without its denominator and chase list. The number that moves reviews is a name with a manager attached.

A percentage with no denominator and no names is a mood. The partition plus the names is a tracker.

◆ The tracker, week 3 of the cyclestatuses that sum, chase list below.

Status	Workers	Share
Completed	214	68.59
In manager review	51	16.35
Self-assessment pending	32	10.26
Not started	15	4.81
Eligible population	312

The 312 eligible derive from the headcount snapshot’s 351 minus the ineligible classes, so the denominator is a query, not an estimate. The 15 not started are the chase list by manager; the 32 pending self-assessments get a different nudge than the 51 sitting with managers, because the event fact knows which step each review is actually at. Sample values are illustrative, never client data.

◆ Step timingswhere cycles actually stall.

With every step dated in the event fact, the cycle’s bottleneck is a grouping: median days in self-assessment, in manager review, in calibration. Cycles stall in predictable places, usually calibration scheduling, and the timing table turns next cycle’s window planning from folklore into arithmetic. The same elapsed-time discipline as recruiting’s clock, applied to reviews.

◆ After completionthe handoff to the shape.

When the cycle closes, completed reviews become rating rows, and the audit moves from counting to the distribution: 296 of the 312 completed by close in this sample, and their shape is the next page. Completion is the process metric; the distribution is the product.

Use case

Problem

Cycle progress is a percentage with no denominator, and nobody knows which step each stuck review is at.

What we build

The cycle tracked from dated process events: statuses partitioning the eligible population, step timings computed.

What you get

A tracker with names and steps, bottlenecks measured, and a clean handoff to the distribution when the cycle closes.

Which fifteen people have not started their reviews?

We ship the tracker that answers with names and managers.

Talk to us

Terms on this page

eligible population: Everyone the cycle covers. The denominator, derived.
partition: Every worker in exactly one status, statuses summing to the total.
self-assessment: The worker’s own step, before the manager’s.
chase list: Not-started reviews grouped by manager for follow-up.
step timing: Days spent in each review step, from dated events.
calibration: The alignment session between manager review and completion.
close: The cycle’s end: statuses freeze, ratings become rows.