Build a Two-Cycle Evidence File for Meta's Checkpoint Reviews

If you want the top rating at Meta, stop polishing a lone hero project. Meta's Checkpoint program will launch in mid-year 2026 with four rating buckets. A top rating is a pattern, and a pattern needs repeated proof.

Checkpoint will run two review cycles, one mid-year and one at year-end, with the same scale in both and bonuses paid twice a year. Your early cycle is the opening part of the record your manager will use later.

A hero project will not carry you

A hero project can be impressive, but it can also be a lone event. When a manager writes a case for a top rating, the question is whether the result would still be true if the project disappeared. If the answer is no, the file is thin.

The new Meta grading system can pay the highest bonus at 300% of base bonus. A new Meta Award will pay a 300% individual bonus multiplier to a limited group of employees whose impact is described as truly exceptional. The difference is what your file has to show.

The expected spread is: Outstanding covers roughly one in five people with a 200% bonus multiplier; Excellent covers about 70% of staff with a 115% multiplier; Needs Improvement covers about 7% with up to a 50% multiplier; Not Meeting Expectations covers about 3% with no bonus. The file must be complete.

The file needs repeated proof

A performance packet is the file your manager can use in a review. It should answer the questions before the manager has to interpret anything for you. The checklist below is the minimum.

• Measurable impact: Can you state the number, baseline, and result in a short line?
• Scope beyond role: Did it affect a team, product, process, or stakeholder outside your title?
• Repeated proof: Does the same pattern appear in separate cycles, not a lone project?
• Manager-ready language: Can your manager quote your evidence without adding interpretation?

Measurable impact is the floor. A number without a baseline is a claim. A baseline without a result is a story. The manager needs the change, not the activity.

Scope beyond role is where the org chart matters. If your work only moved your own tasks, it may be good work, but it is not yet a top-rating argument. The strongest cases show influence across a team, a product decision, a process, or a stakeholder who did not report to you.

Repeated proof is the part that matters most. The file is the safest way to show it. A strong cycle can be luck. Repeated cycles make a trend.

Manager-ready language is the part most people skip. Write the evidence so your manager can lift it into a review without translating it. If you have to explain the context at length, the file is not ready.

Annual equity refreshers will still be paid once a year, but the amount will reflect the average of your two yearly ratings. A weak cycle can drag the number down, and a strong cycle cannot fully repair it. It has to hold up over time.

Starting in 2026, Meta's performance reviews will treat AI-driven impact as a core expectation. If your work uses AI to improve speed, quality, or tooling, put that in the file. If it does not, say what you are building toward.

The file has to survive scrutiny

A manager does not need more adjectives. The best files are short, dated, and tied to outcomes. They name the problem, the action, the result, and the scope. They also leave room for the human part: what you learned, what you would do differently, and where you want to go next.

When you sit down to write it, start with the review cycles. List the outcomes that matter most. For each item, write the number, the baseline, the result, and the people or systems affected. Then read it as a skeptical manager would. If a sentence needs a follow-up question, rewrite it.

A top rating is a judgment that your impact was repeated, measurable, and larger than your title. Build the file so a manager can justify that judgment, a bonus, an equity case, or a promotion case. Then send it before the cycle closes.