Performance reviews
Runs the review cycle with a written submission for every person and a calibration step before ratings are shared, so nobody is judged on one manager's memory of nine months ago.
Tools
Lattice, 15Five, Culture Amp, Google Calendar
Outcomes
Every review written and returned on time • Ratings comparable across teams • Gaps named before calibration starts • No employee surprised by their own review
Documentation
Instruction-ready detail below
Performance reviews fail the same way every year: the cycle opens, everyone is reminded twice, and a third of the reviews are written a week late from memory or never written at all. Calibration becomes a debate about who has better recall, and identical work is rated differently depending on which manager it sits under. This workflow opens the cycle against the current org chart, so somebody who has just left or joined is not quietly missing from a list. Each reviewer gets their own people, their own deadline, and a reminder when nothing has been started. Feedback given between formal cycles, and survey themes from the last round, are attached to the person, so nobody writes from scratch. A submission still open 48 hours out is listed with the reviewer, the person being reviewed, and the missing sections, which is usually enough to get it written. Calibration then compares ratings across teams against the same descriptors before anything is shared. A human still writes the review, decides what is worth praising, and runs the conversation that follows. The workflow makes the input consistent and the record durable; it does not decide who is good at their job.