The room where your rating gets decided

Hey Warblers,

Your performance review isn't solely decided by your manager.

It's decided in a room you'll never enter, by a group of managers negotiating ratings against an expected distribution, using information that may or may not reflect your actual work.

At Google, they call it calibration. At Meta, it's part of the Performance Summary Cycle. At Amazon, it's the Organizational Leadership Review. At Microsoft, calibration happens at each management level. The names vary. The mechanics are similar.

Your manager walks in with a proposed rating. They walk out with whatever the room decided.

Here's what makes this worth understanding: the process is designed to look meritocratic, but the real decisions happen through dynamics most employees never see. Peer feedback often arrives too late to matter. The calibration committee never reads your self-review. Your manager's summary of your work is all they hear, and that summary competes against summaries from every other manager in the room.

If you want to influence your rating, you need to influence the people who will be in that room. Today: exactly how to do that.

The Peer Feedback Problem

Most companies collect peer feedback. Most employees assume it matters. The reality is more complicated.

At Amazon, peer feedback is often due after the talent review meeting has already happened. The decisions are made before your peers even submit their input. The feedback becomes documentation, not decision-making input.

At Google, peer feedback is collected earlier, but here's the catch: the calibration committee never sees the raw feedback. Your manager summarizes it for them. The committee hears your manager's interpretation of what your peers said, filtered through whatever narrative your manager wants to tell.

At Meta, peer feedback gets consolidated into a "calibration package" that your manager presents. Better than Amazon's timing, but still mediated entirely by your manager's framing.

At Microsoft, peer "Perspectives" inform the process, but the actual numerical score that determines your compensation gets assigned in manager-only calibration meetings.

The pattern: Peer feedback exists, but it's either too late, filtered through your manager, or both. The peer feedback process gives employees a sense of participation while the actual decisions happen elsewhere.

This isn't necessarily cynical. Calibration meetings need to move fast. Managers can't read dozens of peer reviews for every employee being discussed. Summaries are practical.

But it means something important for you: if you want to influence your rating, don't rely on peer feedback to speak for you. Influence the people who will actually be in the room.

Who's Actually in the Room

The composition varies by company, but typically includes:

Managers at your level or one above. Your manager is there, plus their peers. These are the people who will debate ratings against each other.

Your skip-level. At most companies, the person running calibration is one or two levels above the employees being discussed. They see the full picture and often have final say on contested ratings.

Senior individual contributors (sometimes). At Meta and Google, senior ICs sometimes participate in calibration for technical roles. They can provide perspective your manager might lack.

HR (usually silent). They're there for process, not advocacy.

The people who can change your outcome are: your manager (who proposes), your skip-level (who approves or challenges), and peer managers (who can validate or dispute what your manager says).

Everyone else in the room is deciding someone else's fate.

The Three Alliances That Matter

If you want to influence what happens when your name comes up, you need relationships with three types of people.

Alliance 1: Your Manager

Your manager is still your primary advocate. They propose your initial rating. They tell your story to the room. They defend you when challenged.

But your manager alone isn't enough. Even strong managers face constraints: limited credibility if they're new, limited time if they have many reports, limited political capital if they're fighting for multiple people.

Your job is to make your manager's job easy. Give them ammunition that survives the compression of calibration.

What your manager needs:

1. A one-sentence headline. "Led the migration that reduced infrastructure costs by 40%." That's what survives when discussions move fast.

2. Quantified impact. The calibration room trusts numbers. Revenue, efficiency gains, customer metrics. Adjectives get challenged; data gets accepted.

3. Third-party validation. Quotes from stakeholders, emails from leadership, Slack messages praising your work. These give your manager evidence to cite: "The VP of Product called this the most successful launch of the year."

4. A clear narrative. Not a list of tasks, but a story: "This year, I went from X to Y, delivering Z impact." Stories are memorable. Lists are forgettable.

Alliance 2: Your Skip-Level

Your skip-level is almost certainly in the calibration meeting. They see ratings across multiple teams. They have context your manager might not. And they can validate or challenge your manager's proposal directly.

If your skip-level knows your work firsthand, they become a second voice. When your manager proposes a rating and your skip says "that matches what I've seen," the rating is harder to challenge. When someone questions your manager's assessment and your skip says "actually, I've worked with this person directly," the conversation shifts.

How to build this:

1. Request occasional 1:1s. Quarterly is reasonable. Frame it as wanting career guidance or understanding broader team priorities.

2. Update them directly after major milestones. A brief message: "Wanted to close the loop on the platform migration. Here's how it landed." Now they have firsthand knowledge of your impact.

3. Make your work visible in forums they attend. All-hands presentations, cross-team updates, technical reviews. Don't rely on your manager to translate your work upward.

4. Before review season, find natural ways to remind them of your contributions. Plant the seed so when your name comes up, they have a recent memory.

Alliance 3: The Managers Who Will Be in the Room

Here's the underrated play: other managers in the calibration meeting can challenge your rating, up or down.

If you've worked cross-functionally with another team, their manager has observed you. They can speak up in calibration: "I worked with this person on the API integration. They were exceptional." Or: "That doesn't match my experience."

This is especially powerful in three scenarios:

When your manager is new. They don't have credibility in the room yet. A peer manager who knows your work carries more weight.

When your manager is stretched. With 12 direct reports, they can't fight hard for everyone. The people with backup advocates get prioritized.

When your manager is weak or checked out. Your backup advocates become your primary defense.

In extreme cases, multiple managers challenging a rating can effectively override it. If three people in the room say "that rating doesn't match my experience," the facilitator takes notice. Managers who consistently get overridden lose credibility. They learn to propose more defensible ratings.

How to build these alliances:

1. Excel at cross-functional work. When you partner with other teams, their managers form impressions. Make those impressions count.

2. Close the loop with their leadership. When a project succeeds, send an update that includes the partner team's manager. Now they have evidence of your impact.

3. Build relationships before you need them. Coffee chats, asking for advice, offering help. When your name comes up in calibration, you want them to have a face and a positive memory attached to it.

The Forced Distribution Nobody Admits To

Let's talk about the math that determines how many people can win.

Most high-performing tech companies have expected rating distributions. The exact numbers vary, but the pattern holds:

Top tier (10-20%): Maximum bonus, equity refreshes, priority for promotion.

Upper middle (25-35%): Good bonus, eligible for promotion, generally safe.

Lower middle (25-35%): Standard bonus, promotion unlikely, expected to improve.

Bottom tier (5-15%): Minimal bonus, no promotion, possible performance management.

Some companies enforce this strictly. Others treat it as guidance. But when leadership looks at aggregate ratings and sees 60% in the top tier, there's pressure to "recalibrate." Ratings come down until the distribution looks acceptable.

What this means: You're not just evaluated against an objective bar. You're evaluated against everyone else in your peer set. You could exceed every goal and still land in the bottom tier if your peers outperformed you, or were better advocated for.

The question isn't "did I do good work?"

The question is "relative to my peer set, how does my work compare, and do the people in the room know it?"

Positioning Above the Line

Understanding the system is step one. Positioning strategically is step two.

Know your peer set

You're being compared against someone. Usually people at your level in your organization, sometimes by function or tenure band.

Ask your manager directly: "Who am I being evaluated against?" If they won't say, observe who gets discussed together in planning meetings and team reviews.

Understand where the cut line is

In every distribution, there's a boundary between "safe" and "at risk." Your goal is to be clearly above it, not on it.

If you're on the margin, you become a bargaining chip. Managers horse-trade marginal cases to protect their top people. "I'll give you this one if you give me that one."

Create separation from your peers

What have you done that nobody else on the team has done? What impact would be lost if you left? What's the story that only you can tell?

If you and a peer both shipped projects, but yours drove revenue and theirs was internal tooling, that's differentiation. Make sure the room knows it.

Control the narrative

Work with your manager to craft a single, memorable sentence that captures your value. That sentence is what survives the fast-paced calibration discussion.

If the narrative is clear and repeatable, it sticks. If it's muddled, someone else simplifies it for you, and you might not like their version.

The Two Weeks Before

Calibration meetings happen on predictable schedules. Learn when yours is. Then use the lead-up strategically.

Two weeks before:

Send your manager a written summary of your accomplishments. Make their job easy. Headlines, metrics, stakeholder quotes. Frame it as "wanted to make sure you have everything you need."

Send brief notes to cross-functional partners thanking them for collaboration and highlighting shared wins. This refreshes their memory right before they might talk to their managers about you.

Have a direct conversation with your manager: "How do you think I'll be positioned? Is there anything I should be doing differently?"

One week before:

If you have a skip-level scheduled, use it. Mention your biggest wins naturally.

If managers you've worked with cross-functionally will be in the room, find a reason to interact. A quick Slack message, a follow-up on a previous conversation. You want to be fresh in their minds.

Ask your manager: "Is there anything specific I should know going into review season? Any concerns I should address?"

The Bottom Line

Your rating gets decided in a room where managers negotiate outcomes based on evidence, relationships, and expected distributions.

Peer feedback is often theater. It arrives too late or gets filtered through your manager's summary. Don't rely on it to speak for you.

Build three alliances: Your manager (give them ammunition), your skip-level (give them direct knowledge), and peer managers (give them positive experiences with your work).

The people who win in calibration aren't just the highest performers. They're the ones who made sure the right people knew about their performance before the meeting started.

When your name comes up, you want multiple voices ready to say "I've worked with this person." One voice can be dismissed. Multiple voices change outcomes.

Do great work. Then make sure the room knows.

~ Warbler