When are Google performance reviews? (2026)
Google assesses against explicit level expectations, with peer input and a calibration process that compares you to the level rather than to your team. Dates move, so the durable preparation is scope evidence mapped to the ladder.
Once a year
No specific review month is published for this company, because none could be sourced from the company's own material. What follows is the cycle shape, which is what is actually knowable. Your own manager is the authority on this year's dates.
How the cycle works
The defining feature of evaluation here is that you are measured against a written definition of your level, not against the person sitting next to you. Each rung has a description of the scope, ambiguity and influence expected at it, and the assessment asks how consistently you operated at that description. This is why the same body of work can read very differently for two people: one is a strong showing at their level, the other is expected behavior at the next one.
Peer input is a real input rather than a formality. Colleagues describe what you did and, more usefully, what changed because of it. Choosing who to ask matters more than most people treat it: the strongest peer input comes from someone who watched you resolve something hard, not from someone who likes you.
Promotion runs on evidence of already operating at the next level, assembled into a case that people outside your immediate team can evaluate. That is a documentation exercise as much as an engineering one.
Cycle dates vary and have changed across years. Confirm yours with your manager rather than assuming a schedule someone remembers.
What actually gets weighed
Scope of ownership is the primary axis. Not hours, not volume of code, not number of projects, but the size and ambiguity of the problem you can be handed and trusted to resolve without a safety net.
Measured impact is the second. This is a company with instrumentation everywhere, so a claim with a number attached is materially stronger than a claim without one, and a claim with the wrong number is worse than no claim at all. Know how your metric was computed and what it does not capture.
Cross-team influence is the third, and it is where senior cases are usually won or lost. Getting a decision made across three teams with no authority over any of them is the behavior the upper rungs describe.
What to have ready
- The written expectations for your current level and the one above it, with your own work mapped against each line.
- Metrics for anything you shipped, including how each one was measured.
- A shortlist of peers who observed you doing something difficult, chosen for what they saw rather than for how they feel.
- Design documents or decisions you owned, especially ones adopted outside your team.
- Ambiguous problems you scoped, since scoping is the behavior the ladder actually describes.