How to do an annual review
An annual review fails when next year's wish is written before the year that actually happened is on the page.
Goal
Finish with two artifacts:
- A past you could hand to someone who was not in your calendar: significant work, results you can show, unmet aims with causes.
- A next period that cites those gaps: what you will try, what support you need, and a future state a later review could check.
Include dates, artifacts, and outcomes. Cut mood summaries ("a busy year"), character verdicts, and goals that do not point at anything you already documented.
One sitting can cover the past. The future can wait until the past has evidence. If both must happen in the same afternoon, finish the past before you write a single next-year line.
Method
Three jobs compete in the same sitting: remembering, judging, and wishing. The wish is faster, so it wins. Last January's slogans come back with new adjectives.
The horizon is long. A two-week retro will not hold a year. The unit is you. A team debrief of one incident is a different sitting.
Label memory that has no artifact. If you cannot point at a doc, a date, a shipped thing, or a metric you already share, write "assumed" next to the line. Otherwise the past becomes a story you like, and the future copies it.
When last year had a stated goal (a number, a date, a quality bar), GRAI is the tighter spine: lock the goal, record the result, explain the gap, extract insight. When the year is already clear and you only need to file habits into continue / change / begin / drop, KISS is enough. Past and Future Self is the default when the year is still a pile.
Steps
1. Bound the period
Calendar year, last twelve months, or the cycle your company actually uses. Write the start and end dates at the top. Do not review "my career so far" in this pass.
2. Dump effort with evidence
List significant work and projects. Attach a date or an artifact to each line. Chronological or by category is fine. If memory is all you have, mark it assumed.
3. Name achievement, then growth gaps
Achievement is a result you can show. A growth gap is an unmet aim plus a cause you can act on, not a personality label. Keep those two lists separate.
4. Choose the writing spine
If the year is still a blur, stay on Past and Future Self (six parts). If a baseline exists and the review would stop at the score, switch to GRAI. If the past is already documented and you need action buckets, sort with KISS. See Frameworks.
5. Write the future last
Plan next period from the gaps. Name support in a reasonable scope. Describe a future state specific enough to review later. Do not start with the future state and backfill a heroic past.
- Period: [start] → [end]
- Past
- - Effort: [work + date or artifact]
- - Achievement: [result you can show]
- - Room to grow: [unmet aim + cause, not a character verdict]
- Future
- - Plan: [moves tied to those gaps]
- - Support: [what you need, in scope]
- - Future state: [a picture a later review could check]
Frameworks
How each framework helps with this problem.
Primary
Past and Future Self Review
Document a period with evidence of effort, achievement, and growth gaps, then set a next-period plan with needed support and a future state.
Alternative — goal-result gap
GRAI Review Framework
Review a period of work by comparing the original goal with the actual result, analyzing what produced the gap, and turning that into insight for the next period.
Alternative — action buckets
KISS Review Framework
Turn a review of recent work into four action buckets: what to keep, improve, start, and stop.
Example
Past and Future Self — a year that felt busy
Period: January–December.
Effort: two product launches (March, September); hiring loop for one engineer (April–June); on-call rotation most weeks.
Achievement: both launches shipped; the hire started in July; incident load stayed high through Q4.
Room to grow: no delegated owner for on-call, so leadership time sat on incident bridges. Cause: the rotation was never staffed beyond "whoever is free."
Plan: name one on-call owner by February; monthly 1:1 cadence with that owner. Support: manager air cover for the ownership change. Future state: fewer nights on incident bridges, same delivery bar.
The leadership wish is now tied to a documented constraint.
GRAI — when a number existed
Goal: ship mobile checkout by 30 November. Result: shipped 18 December. Analysis: design sign-off sat in review for three weeks in October; legal added a disclosure after the engineering freeze. Insight: keep the engineering slice as it ran; start a sign-off cutoff ten days before freeze; stop treating legal as a last-day checkbox.
The original date stays on the page, so next year's insight can be tested. "We need more urgency" would have dropped the baseline.
Mistakes
- Starting with next year's theme before the past has dates
- Treating a mood ("busy," "hard") as achievement
- Growth gaps that are character verdicts ("I'm bad at focus")
- Running a team After-Action Review as the whole annual method
- GRAI without restating the original goal
- KISS as personality Keep/Stop ("keep being a team player")
- A future state no later review could check
- Backfilling a heroic past to justify a wish
Next step
Bound the period and dump effort with evidence. If the year is still a blur, stay on Past and Future Self. If a baseline exists, write GRAI. If the past is already clear, sort with KISS. GRAI has the full four stages when you need them.
Apply this to your own context
Bring your situation into Advisor. It helps you choose a framework and work through the next step.