How-to: Write a Self-Review
A self-review is not a marketing brochure and not a confession booth. It is a structured account of outcomes, how you worked with others, what you learned, and where you want to grow next. Managers calibrate across people; your job is to make evidence easy to find and judgment easy to apply. Clear beats clever. Specific beats adjectival.
Time box: 2–3 hours across two sittings (gather evidence, then draft and cool-edit). Rushing in one anxious night produces either fluff or self-owning that helps nobody.
Use with the Individual Contributor hub and New Manager hub. Pairs with receiving hard feedback and everyday feedback in the moment. Follow your company’s form and ratings language first; this how-to fills the blanks with substance.
When to use / when not
Use when:
– Your company runs a performance cycle with a self-review field
– You are preparing for a promotion conversation and need a packet of evidence
– You changed managers mid-cycle and must reconstruct impact
– You want a personal annual archive even if the form is light
Do not use when:
– You are asked for a legal statement or HR investigation narrative (different process)
– You plan to litigate coworkers by name in a shared form
– You invent metrics or claim solo credit for team outcomes
– You are in an active formal plan with specific HR instructions that supersede this template
If the cycle intersects with accommodations, leaves, or concerns about bias, talk to the right channel early (how HR works).
Prep
- Download or copy the exact prompts and rating scales
- Pull calendar, tickets, docs, launch notes, goals, and kudos from the cycle period
- List 5–8 candidate accomplishments; you will keep 3–5
- Gather peer input you already have (do not spam the company for compliments last minute)
- Note constraints: headcount, incidents, market shifts, inherited messes (facts, not excuses)
- For managers, gather team outcomes and people practices separately
If your work tied to revenue or cross-functional launches, reconcile stories with partners so you do not contradict CRM or program truth (how sales works, how departments work together). Money claims should match how finance talks about results (finance for non-finance managers).
Steps
- Skim the form and rubrics (15 min). Write to their questions, not your favorite essay shape.
- Build an evidence table (40–50 min). Outcome, your role, collaborators, proof link, business relevance.
- Choose headlines (20 min). Three to five accomplishments with verbs and objects. Cut hobby projects unless they mattered.
- Draft outcomes section (30–40 min). Lead with results, then your contribution, then how you know.
- Draft collaboration and values section (20 min). Concrete behaviors: mentorship, handoffs, incident response, inclusive practices.
- Draft growth and goals (20 min). Honest gaps plus a forward plan. Reference feedback you received and what changed.
- Cool down and edit (20–30 min next day). Remove emoting, hedging piles, and unearned superlatives. Check word limits.
- Optional peer read (15 min). Ask one trusted colleague: “Where is this unclear or overclaimed?”
- Submit on time and keep a personal copy.
New hires in a short cycle should say what they learned and owned, not apologize for tenure (New hire hub). Directors should emphasize org-level outcomes and succession, not task lists (Director / VP hub).
Script (what to say / send)
Accomplishment bullet pattern:
[Outcome]. My role: [led / co-owned / contributed]. Evidence: [link or artifact]. Business relevance: [customer, risk, efficiency, revenue enablement]. Collaborators: [teams/names].
Example shape (illustrative structure only):
Shipped the redesigned onboarding checklist used by the new-hire cohort. I drafted the runbook, ran two pilot weeks with People Ops, and folded feedback into v2. Manager dashboard now shows completion in one place. Collaborators: People Ops, IT buddy program.
Growth paragraph starter:
Feedback I received on [theme] led me to [specific behavior change]. I still want to improve at [skill]. Next cycle plan: [practice, mentor, project stretch], success signal [observable].
Ask to manager before ratings lock:
I submitted my self-review. Happy to walk evidence for 20 minutes if useful, especially on [ambiguous item]. I also want your read on growth areas so my next-cycle goals match how you evaluate.
If you disagree with a likely rating (calm):
I would like to understand the criteria and examples behind my rating. Here is evidence I may not have surfaced well: [2–3 bullets]. What would stronger performance look like next cycle?
Peer input request (early, not night-before):
If you have two minutes: what is one outcome we shared this cycle that I should include accurately, and one growth edge you have observed? Short bullets are perfect.
Output artifact
Self-review working file:
| Accomplishment | Role | Proof | Partners | Notes for form |
|---|---|---|---|---|
Draft sections:
- Goals vs. outcomes
- Top accomplishments (3–5)
- Collaboration / leadership / values
- Learnings and feedback applied
- Goals for next cycle
- Optional: support you need (resources, clarity, scope)
Keep private notes separate from the submitted form when they include sensitive interpersonal detail. Promotion hopes belong in a dedicated conversation with a packet, not hidden inside self-review poetry.
Watch-outs
- Invented precision. Do not fabricate percentages or rankings. Use real artifacts and qualitative proof when numbers are unavailable.
- Invisible “I” in team wins. Name your part without erasing others.
- Only chores, no outcomes. “Attended meetings” is not performance.
- Only outcomes, no learning. Growth sections show coachability.
- Sandbagging or peacocking. Both hurt calibration.
- Surprise complaints about your manager in the form. Use live dialogue or HR paths when appropriate.
- Em dashes, sarcasm, or jokes that read wrong in a packet that lives for years. (Keep tone plain.)
Done looks like: a form that a busy calibrator can skim, evidence that can be checked, growth that sounds adult, and a copy in your own archive. Your self-review cannot guarantee a rating. It can guarantee you showed up as a clear historian of your work.