Language: English
Start 14-day free trial

Official skill · Management

Quarterly OKRs from a strategy

v1.0.1 · Updated

Drafts quarterly OKRs: 1 to 3 objectives, measurable key results with baseline, target and owner, and a check that no KR is a task. Use it when you need to turn a strategy or goals into quarterly OKRs.

Add to my team's pod Download SKILL.md

Works in Claude, ChatGPT, and coding agents. On Team and Business, the official skill in your pod updates itself when we improve it.

When to use it

What you give it

What you get

A set of OKRs with baselines, targets, owners, committed or stretch labels, counter-metrics, and a quality check.

How it works

  1. Pick 1 to 3 objectives Our default for a startup team is 3 at most; Google's re:Work guide says to pick 3 to 5 and warns that more can over-extend teams and diffuse effort. Each objective is a qualitative endpoint a team member could repeat from memory ("Make onboarding something customers finish on day 1"), never ongoing activity ("keep hiring"). If the input has 5 goals, rank them by impact, keep the top 3, and list the rest under "Not this quarter".
  2. Write 2 to 4 key results per objective Our default; re:Work suggests around 3 per objective. Each KR is a metric moving from a baseline to a target by a date: "Day-1 onboarding completion from 34% to 60%". Use the form "[metric] from [baseline] to [target]". Key results must be measurable and verifiable at the end of the quarter.
  3. Test each KR: outcome or output? Ask "If we did this and nothing else changed, would we celebrate?" "Launch the new onboarding flow" fails: it's an output. Rewrite it as the change the launch should cause. Move outputs to an "Initiatives" list under the objective.
  4. Set ambition Mark each KR as committed (expected to be fully met) or aspirational, also called stretch (the path isn't known yet), the two types What Matters describes. Google's re:Work guide grades OKRs from 0.0 to 1.0, puts the sweet spot at 60 to 70%, and says someone who consistently fully attains their objectives isn't setting them ambitiously enough. If every KR is committed, say so; if every KR is stretch, nobody knows what's promised.
  5. Give each KR one owner A person, never a team. The owner reports progress; they don't have to do all the work.
  6. Check for gaps and conflicts A KR that can be gamed needs a counter-metric (raise trial sign-ups → watch trial-to-paid). Two KRs that pull against each other get a note on which wins.
  7. Run the quality check in the output and fix every failed line before sending.

The SKILL.md

The whole file, as your AI reads it. Download it

---
name: okr-draft
description: Drafts quarterly OKRs from a strategy, goals, or planning notes - 1 to 3 objectives, 2 to 4 measurable key results each that measure outcomes rather than tasks, with a baseline, a target, one owner per KR, and a quality check that flags tasks posing as key results. Use when a founder, manager, or team asks to write OKRs, set quarterly goals, turn a strategy into objectives and key results, or review draft OKRs.
---

# OKR draft

Good OKRs tell a team what to stop doing. An objective says where we want to be; a key result is a number that proves we got there, not a list of things we'll ship. If a key result can be "done" without anything changing for a customer or the business, it's a task. OKRs are also not a performance evaluation: don't tie them to anyone's rating.

## When to use

- Turning a strategy doc, board goals, or planning notes into quarterly OKRs.
- Reviewing a draft set of OKRs before a planning meeting.
- Cascading company OKRs into one team's OKRs.
- Not for a project plan or task list: OKRs sit above that. Use quarterly-planning for the full plan.

## What you need

- The strategy or goals for the period (company OKRs, a strategy memo, or a founder's notes).
- The team and quarter the OKRs are for.
- Current numbers for anything you'll measure. If a baseline is missing, write "[baseline: to measure in week 1]" and make measuring it the owner's first action. Never invent a baseline.
- Who is on the team, so each KR gets one owner.

## Steps

1. **Pick 1 to 3 objectives.** Our default for a startup team is 3 at most; Google's re:Work guide says to pick 3 to 5 and warns that more can over-extend teams and diffuse effort. Each objective is a qualitative endpoint a team member could repeat from memory ("Make onboarding something customers finish on day 1"), never ongoing activity ("keep hiring"). If the input has 5 goals, rank them by impact, keep the top 3, and list the rest under "Not this quarter".
2. **Write 2 to 4 key results per objective.** Our default; re:Work suggests around 3 per objective. Each KR is a metric moving from a baseline to a target by a date: "Day-1 onboarding completion from 34% to 60%". Use the form "[metric] from [baseline] to [target]". Key results must be measurable and verifiable at the end of the quarter.
3. **Test each KR: outcome or output?** Ask "If we did this and nothing else changed, would we celebrate?" "Launch the new onboarding flow" fails: it's an output. Rewrite it as the change the launch should cause. Move outputs to an "Initiatives" list under the objective.
4. **Set ambition.** Mark each KR as committed (expected to be fully met) or aspirational, also called stretch (the path isn't known yet), the two types What Matters describes. Google's re:Work guide grades OKRs from 0.0 to 1.0, puts the sweet spot at 60 to 70%, and says someone who consistently fully attains their objectives isn't setting them ambitiously enough. If every KR is committed, say so; if every KR is stretch, nobody knows what's promised.
5. **Give each KR one owner.** A person, never a team. The owner reports progress; they don't have to do all the work.
6. **Check for gaps and conflicts.** A KR that can be gamed needs a counter-metric (raise trial sign-ups → watch trial-to-paid). Two KRs that pull against each other get a note on which wins.
7. **Run the quality check** in the output and fix every failed line before sending.

## Output format

```markdown
# OKRs — [team], [quarter]

## Objective 1: [qualitative, memorable sentence]
Why now: [one line linking to the strategy]

| KR | Baseline | Target | Type | Owner |
|---|---|---|---|---|
| 1.1 [metric] | [number or "to measure in week 1"] | [number] by [date] | committed / stretch | [person] |

Counter-metric: [metric that must not get worse, if a KR can be gamed]
Initiatives (outputs, not KRs): [project], [project]

## Not this quarter
- [Goal left out] — [why]

## Quality check
- [ ] 3 objectives or fewer, 2–4 KRs each
- [ ] Every KR is a number with a baseline and target
- [ ] No KR is a task ("launch", "build", "hire", "write")
- [ ] Every KR has one named owner
- [ ] Mix of committed and stretch
```

## Example

**Input:** "Growth team, Q3. Strategy: we're losing trials because setup is confusing. Goals: improve activation, launch new onboarding, hire a lifecycle marketer, get more trial sign-ups, start a referral program. Activation (connected first data source in 7 days) is 31%. Trial sign-ups ~900/month. Team: Mina (PM), Joel (lifecycle), Ana (eng lead)."

**Output:**

```markdown
# OKRs — Growth, Q3

## Objective 1: New trials reach their first result in the first week
Why now: confusing setup is the main reason trials don't convert.

| KR | Baseline | Target | Type | Owner |
|---|---|---|---|---|
| 1.1 Trials that connect a data source within 7 days | 31% | 50% by 30 Sep | committed | Mina |
| 1.2 Median time from sign-up to first connected source | [baseline: to measure in week 1] | under 1 day by 30 Sep | stretch | Ana |
| 1.3 Setup-related support tickets per 100 trials | [baseline: to measure in week 1] | half of baseline by 30 Sep | committed | Joel |

Counter-metric: trial-to-paid rate must not drop below current.
Initiatives (outputs, not KRs): new onboarding flow, setup emails rewrite.

## Objective 2: More of the right people start a trial

| KR | Baseline | Target | Type | Owner |
|---|---|---|---|---|
| 2.1 Monthly trial sign-ups | ~900 | 1,150 by Sep | stretch | Joel |

Counter-metric: share of sign-ups that activate (KR 1.1) must not fall.
Initiatives: referral program pilot.

## Not this quarter
- "Hire a lifecycle marketer" — a task, not an outcome; tracked in the hiring plan.

## Quality check
- [ ] 3 objectives or fewer, 2–4 KRs each — not yet: Objective 2 has 1 KR; add a second or merge it into Objective 1.
- [x] Every KR is a number with a baseline and target (2 baselines measured in week 1)
- [x] No KR is a task — "launch new onboarding" moved to initiatives
- [x] Every KR has one named owner
- [x] Mix of committed and stretch
```

## Checks before you send

- [ ] No KR starts with launch, build, ship, hire, write, or complete.
- [ ] Every baseline comes from the input or is marked "to measure in week 1".
- [ ] Every KR has exactly one named person as owner.
- [ ] Each objective has 2 to 4 KRs, or the gap is flagged in the quality check.
- [ ] Goals left out are listed under "Not this quarter" with a reason.

## Common mistakes

- Writing the team's project list as KRs. That's a roadmap.
- Targets with no baseline ("increase NPS to 50"). Nobody can tell if 50 is ambitious.
- 5 objectives. That's a list of everything, which tells the team nothing about priority.

Share it with your team

Add it to an AI Skill Pod and everyone on the team, contractors included, uses the same version from their own AI. Read the best ways to share AI skills with a team

Add to my team's pod

Changelog

Sources

  1. Google re:Work - Guides: Set goals with OKRs (read 2026-11-04)
  2. What is an OKR? OKR Meaning, Definition & Examples - What Matters (read 2026-11-04)