OKR template with examples (based on Google's guide)
A free OKR template from Google's re:Work guide: writing objectives and key results, grading from 0 to 1, and examples for product, engineering, and HR.
A team sets nine objectives for the quarter, each with five key results, most of them tasks: "launch the new dashboard," "hold three customer workshops." By week six, nobody can name the objectives, and the end-of-quarter scoring is a debate about whether a workshop counts as done.
OKRs work when they're few, ambitious, and measurable. This template follows Google's re:Work guide to OKRs, checked on 16 November 2026.
The rules, from Google's guide
- Objectives are ambitious and may feel somewhat uncomfortable.
- Key results are measurable and easy to grade with a number.
- Three to five objectives, with about three key results each.
- Graded from 0.0 to 1.0, with an average of 0.6 to 0.7 as the sweet spot.
- Public, so everyone can see what others are working on.
- Not a performance evaluation, and not a shared to-do list.
The template
Works in Google Sheets, Excel, Notion, or a doc:
| Objective | Key result | Owner | Start | Target | Current | Score (0 to 1) |
|---|---|---|---|---|---|---|
| [Ambitious, qualitative goal] | [Measurable outcome] | [Name] | [Baseline] | [Target] | ||
| [Measurable outcome] | ||||||
| [Measurable outcome] |
For each key result, score = progress from start to target, capped at 1.0. Average the key results for the objective's score.
Writing good objectives
Google's tips:
- Pick three to five. More leads to over-extended teams.
- Avoid maintenance language like "keep hiring" or "continue doing X." It doesn't push for anything new.
- Use endpoints and states: "ship feature Y," not "work on feature Y."
Writing good key results
- About three per objective, each a measurable milestone that directly advances it.
- Outcomes, not activities. Google flags words like consult, help, analyze, and participate as signs of an activity. Its example: "publish customer service satisfaction levels by March 7th," not "assess customer service satisfaction."
- Evidence of completion that is available, credible, and easy to find.
Examples
Product
| Objective | Key results |
|---|---|
| Make onboarding effortless for new teams | Median time to first project from 3 days to under 1 day |
| Week-one activation from 42% to 60% | |
| Onboarding support tickets down 50% |
Engineering
| Objective | Key results |
|---|---|
| Ship with confidence every day | Deploy frequency from weekly to daily for 3 core services |
| Change failure rate under 10% | |
| Mean time to restore under 1 hour |
HR and people
| Objective | Key results |
|---|---|
| Make new hires productive fast | Time to first merged PR for engineers from 21 days to 7 |
| New-hire 30-day survey score from 3.6 to 4.3 out of 5 | |
| 100% of new hires with an assigned buddy on day one |
AI adoption
| Objective | Key results |
|---|---|
| Make AI coding agents a dependable part of how we build | 80% of engineers using an agent weekly |
| Agent-authored PRs merged without major rework from 55% to 75% | |
| Zero incidents caused by agents contradicting a recorded decision |
Grading
At the end of the period, score each key result from 0.0 to 1.0. Google's guide says to be honest and, above all, consistent. An average of 0.6 to 0.7 means the goals were ambitious enough; consistently hitting 1.0 means they weren't. Low scores are information for next quarter, not failure.
Common mistakes
- Too many objectives. Effort spreads thin and nothing gets done well.
- Tasks as key results. "Launch X" can be done without changing anything.
- Sandbagging. Easy goals scored at 1.0 teach nothing.
- Tying OKRs to pay. People set safe goals.
- Set and forget. Check progress at least monthly.
From OKRs to decisions
OKRs set direction. Reaching them takes a quarter's worth of decisions: which features to cut to hit the activation target, which service to migrate first. Those decisions are where an OKR succeeds or quietly fails, and they're usually made in meetings and threads nobody can find later.
Where Lockstep fits
Lockstep records those decisions where they're made, in PRDs, Slack, and pull requests, routes each to its owner for approval, and briefs engineers and Claude Code sessions on the ones that apply to their work, so a decision made to hit a key result reaches the people building toward it.
See Lockstep for product teams.
Sources checked 16 November 2026
Related reading
For turning goals into a plan, see the product roadmap template. For prioritizing work against them, see product decision-making frameworks. For engineering delivery metrics, see DORA metrics.
Frequently asked questions
- What is an OKR?
- Objectives and Key Results: an objective is an ambitious, qualitative goal, and key results are measurable outcomes that show whether you reached it. Google's re:Work guide describes objectives as ambitious and somewhat uncomfortable, and key results as easy to grade with a number.
- How many OKRs should a team have?
- Google's guide suggests three to five objectives with about three key results each. More objectives spread effort too thin.
- How are OKRs graded?
- Google uses a 0.0 to 1.0 scale. The sweet spot is an average of 0.6 to 0.7. Consistently scoring 1.0 means the objectives weren't ambitious enough; low scores are data for setting the next ones.
- What is the difference between a key result and a task?
- A key result describes an outcome; a task describes an activity. Google's guide says words like consult, help, analyze, or participate signal an activity. 'Cut onboarding time from 14 days to 5' is a key result; 'run onboarding workshops' is a task.
- Should OKRs be used for performance reviews?
- Google's guide says no: OKRs are not synonymous with employee evaluations. They can summarize what someone worked on and contributed, but tying pay to OKR scores pushes people to set safe goals.
Keep reading