Performance reviews, explained
How the corporate performance review actually works — the process, the self-assessment, calibration, and how to prepare so the whole year counts, not just the three weeks before it.
A performance review is your employer's formal verdict on how you did over a set period, usually a year, ending in a rating that feeds pay, promotion, and — in a bad year — whether you keep the job. What decides how yours goes is rarely how hard you worked. It is how well the period is documented when your manager, and later a room of managers, tries to remember it. Recency bias means the last month counts for more than it should, so a dated record beats a good memory every time.
What a performance review actually is
Strip away the software and the five-point scales and a review is one thing: a written judgment of your work over a cycle, made mostly by your manager, that the company then uses to set pay, promotions, and priorities. Most corporate reviews bolt together three parts — a self-assessment you write, your manager's assessment and rating, and goals for next cycle. The rating is the part that travels: into the comp spreadsheet, the promotion committee, and your permanent file.
It is worth being honest about how well the ritual works, which is: not very. Gallup finds only 14% of employees strongly agree their reviews inspire them to improve. Gartner puts manager dissatisfaction with the process at 95%. The review is not going anywhere, but nobody is calling it a clean measurement. Treat it as a high-stakes memory test with a rating attached and you will prepare for the right thing.
The corporate review process, end to end
The steps vary by company. The shape barely does:
- You write a self-assessment. You summarise the cycle, usually against your role's criteria or last cycle's goals.
- Your manager writes their assessment and a proposed rating. Often they gather peer feedback first.
- Calibration. Managers meet to compare ratings across the team or org and adjust them so they are consistent — this is where a proposed rating can move up or down.
- The delivery meeting. Your manager walks you through the final rating, the reasoning, and next-cycle goals.
Two of these you can actually influence. The self-assessment is your one direct input. Calibration is where the evidence you handed your manager either survives the room or quietly does not. Everything else rides on how well those two are backed.
Annual vs mid-year reviews
Most companies run a full annual performance review once a year and a lighter mid-year review in between. The annual one sets ratings, pay, and promotions. The mid-year is a checkpoint: it confirms you are on track, catches gaps early, and quietly seeds your manager's memory for the round that counts.
The mistake is treating mid-year as a rehearsal. What you put on the record in June is what your manager still has in December, long after the specifics have gone fuzzy. Use it to write your first-half wins down while they are fresh, so they are not competing with last week for attention.
How to write your self-assessment
Your self evaluation is the highest-stakes thing you write all year, because it frames the entire conversation. The rule is boring and it works: turn every task into a statement of impact — what changed, for whom, how much — and cover the whole cycle, not just the months you remember. "Ran the weekly reports" is a task. "Rebuilt the weekly reporting so the team stopped losing two hours every Monday" is impact. Same work, different rating.
What to include — and what to cut
- Outcomes across the whole cycle, framed as impact
- Numbers, links, or named results as evidence
- Scope you grew into during the period
- One honest growth area with a concrete plan
- Only the last month or two of work
- A diary of tasks with no consequences
- Self-criticism with no path forward
- Empty superlatives — "smashed it", "crushed it"
Performance review goals examples
Good performance review goals are specific, tied to a number, and mapped to the next level of your role. Vague intentions do not survive to the next cycle. A few to adapt:
- Impact goal: "Cut average support resolution time from 14 hours to under 8 by Q3, owning the triage rewrite."
- Scope goal: "Lead the migration end to end and unblock two teammates, moving from contributor to owner on the platform."
- Development goal: "Present at two cross-team reviews to build the visibility my next level requires."
- Collaboration goal: "Mentor one new engineer to independent shipping within the cycle."
The pattern does not change: a verb, a number, a level. If you cannot evidence it at the next review, it is a wish, not a goal.
Areas of improvement, without sinking your rating
Every review asks for areas of improvement, and skipping it reads as no self-awareness. Name a real one, then pair it with a plan and a first step you have already taken. "I want to get better at cross-team influence; I have started presenting at the monthly review and I will run the next planning session" reads as maturity. A bare confession with no plan just hands the reviewer a weakness to weigh. Name one, own it, show it is already moving.
What calibration is — and why your record matters
Calibration is the step most employees never see and the one that most often moves a rating. Once managers have drafted their scores, they meet to compare everyone and force consistency, so a "strong" means the same thing under every manager. In that room your manager is defending your rating from memory — against peers, against the clock, against whatever you shipped most recently looming largest.
This is the entire case for keeping a record. A manager armed with dated, specific evidence can hold the line. A manager working from a warm blur cannot. You do not get to attend calibration, but you do get to decide what your manager walks in carrying.
How to prepare for a performance review
- Gather the whole cycle first — pull wins from month one, not just the last sprint.
- Reframe every task as impact — consequence, not activity.
- Map each item to your level's written criteria so the rating has somewhere to land.
- Bring evidence — numbers, links, named results. "It went well" is forgettable.
- Give your manager the ammunition — hand them the record before calibration, not after.
The annual review is mostly a memory test
Here is the opinion the whole guide rests on: the annual review measures your work far less than it measures how well your work is remembered — a critique HBR laid out years ago. Culture Amp calls the culprit the "what have you done for me lately" bias — recent events loom larger than earlier ones, even when the earlier ones mattered more. It is part of why only 14% of people find these reviews useful. The fix is not a heroic Q4. It is to stop trusting memory, yours or your manager's, and write the cycle down as it happens.
When a performance review doesn't matter
Not every review earns the effort. If yours is a formality with fixed raises and no promotion path, or you have already decided to leave, do the minimum and spend the energy elsewhere. Preparation pays when the rating actually feeds pay, a promotion, or your standing in a calibration room. When it feeds nothing, treat it as the box-ticking it is.
How Workfied helps here
Workfied holds the whole cycle, so recency bias works for you instead of against you — your strongest win from January is still on the page in December. You walk into your self-assessment writing from a dated record, not from memory, and your manager walks into calibration with specifics they can defend. Built for you, never your company. Encrypted in transit and at rest.
Evidence
The problem
Only 14% of employees strongly agree their performance reviews inspire them to improve.
Source: Gallup
The bias
Managers weigh recent work over the full year — the "what have you done for me lately" effect.
Source: Culture Amp
The fix
Evidence from across the cycle survives calibration where a warm impression cannot.
Workfied house view