OKRs · Goal setting · Product management
OKR examples for a small product team, from first draft to final score
By T. Krause
Most lists of OKR examples stop at the writing. This guide follows one made-up product team through a whole quarter: a weak first draft, the three objectives they kept, how they scored each key result, and what the scores told them.
Search for OKR examples and you get long lists. Twenty objectives for marketing, twenty for engineering, each with three tidy key results. They show you the format. They don't show you which of your ideas belong in an OKR, what a bad first draft looks like, or what happens at the end of the quarter when someone has to put a number next to each key result.
This guide follows one team through all of that. It explains what OKRs are and how to write and score them, then works through a single example from the first draft to the final scores. The team and every number in it are made up, so you can check the arithmetic without anyone's real data.
A note before we start. We make Prompt Board, and the last section shows how to put your own OKRs on our OKR planning template. The method works the same on a spreadsheet or a wall of sticky notes.
What OKRs are
OKR stands for objectives and key results. An objective says what you want to achieve, in words. Key results say how you'll know you got there, in numbers. Andy Grove built the method at Intel, and John Doerr, who learned it from him, introduced it to Google's founders in 1999. Doerr sums it up as a sentence: "I will (objective) as measured by (key results)."
Most teams add a third layer, initiatives. These are the projects you expect to move the numbers. Keep them apart from the key results. A key result says where you want to end up. An initiative is your current bet on how to get there, and you can swap it mid-quarter without changing the goal.
Two more rules from the people who made OKRs popular. Google's re:Work guide says OKRs are public, so everyone can see what other teams are working on. And OKRs stay out of pay. Doerr's site calls them "divorced from compensation". That's one of the main differences from the older management by objectives approach.
How to write OKRs
- Start from the problem, not the backlog. Ask what has to be different by the end of the quarter. If the answer is a feature, ask what the feature is for. That answer is closer to an objective.
- Write the objective in plain words, with no number. "Make new groomers successful in their first week" says what matters. The numbers go in the key results.
- Give each objective about three key results. Google's guide suggests three to five objectives with around three key results each. A small team with one quarter usually does better at the low end. Our template has rows for three objectives.
- Give every key result a start and a target. "Activation from 30% to 50%" can be scored. "Better activation" can't. If you don't know the start, measuring it is the first job of the quarter.
- Check every key result for activity words. Google's internal OKR guide, published by What Matters, warns that key results with words like "consult", "help", "analyze" or "participate" describe activities. "Launch reminders" is an activity. "No-show rate from 12% to 6%" is a result.
- Decide which objectives are committed and which are aspirational. More on that below. Write it on the board so nobody argues about it at scoring time.
- List initiatives last. One or two per objective is enough to start the quarter. If an initiative is a product release, a user story map helps decide what goes into its first version. If it needs people from several teams, a RACI matrix settles who does the work and who signs it off.
Committed and aspirational OKRs
Google splits OKRs into two kinds, and the split changes what a good score looks like.
A committed OKR is a promise. The team expects to hit it in full and will move people and time to make sure it does. Google's guide says the expected score is 1.0, and anything less needs an explanation.
An aspirational OKR is a stretch. Nobody knows exactly how to reach it when the quarter starts. Google expects these to average 0.7, with a wide spread between them. The re:Work guide puts the sweet spot at 60 to 70 per cent. A team that scores 1.0 on every aspirational OKR set them too low.
Most teams need both. Keeping the product running is committed work. Changing how customers behave is usually aspirational.
A worked OKR example: a booking app for dog groomers
Here is the whole cycle for one team. The product, the team and the numbers are invented for this guide.
The team
Pawly is a booking app for independent dog groomers. A groomer signs up, sets up a booking page with their services and opening hours, and shares the link. Dog owners book a slot through that page. The team has six people: a product manager, a designer, three developers, and one person who handles support and marketing. They plan one quarter at a time.
The first draft
The product manager's first draft looked like this:
- Objective: Improve onboarding and launch SMS reminders
- Redesign the setup flow
- Launch reminders by February
- Interview 15 groomers
- Increase customer satisfaction
Everything on that list is worth doing, and it's still a weak OKR. The objective holds two unrelated goals. Three of the four key results are activities, so the team could finish all of them and change nothing for groomers. The fourth has no number, so nobody can score it. And it's all aspirational, while the thing groomers complain about most often, a slow booking page, isn't mentioned at all.
The OKRs they kept
After an hour of rewriting, the team settled on three objectives with three key results each. Two objectives are aspirational and one is committed.
Objective 1: Make new groomers successful in their first week (aspirational)
| Key result | Start | Target |
|---|---|---|
| New groomers who get their first online booking within 7 days of sign-up | 30% | 50% |
| Median time from sign-up to a published booking page | 3 days | 1 day |
| Setup questions to support per 100 new groomers | 40 | 20 |
Initiatives: a setup checklist on first login, and importing a client list from a spreadsheet.
Objective 2: Help groomers stop losing slots to no-shows (aspirational)
| Key result | Start | Target |
|---|---|---|
| No-show rate on Pawly bookings | 12% | 6% |
| Bookings where the dog owner confirms the reminder text | 0% | 40% |
| Rebookings made from the reminder text, per month | 0 | 300 |
Initiatives: a text reminder 24 hours before each appointment, and a rebook button in that reminder.
Objective 3: Keep the booking page fast and reliable (committed)
| Key result | Start | Target |
|---|---|---|
| Booking page uptime, every month of the quarter | 99.5% last quarter | 99.9% or better |
| Booking page load time on mobile, 75th percentile | 2.6 seconds | under 2 seconds |
| No double-bookings reported by groomers (yes or no) | missed last quarter | met |
Initiatives: a load test before each release, and a rewrite of the calendar sync.
The interviews from the first draft didn't vanish. They became part of the work behind objective 1. "Redesign the setup flow" became the setup checklist initiative. The reminders moved under objective 2, where the key results measure whether reminders change what dog owners do.
The check-in at week 6
The team checked in every three weeks. At week 6 the rebook button still hadn't shipped, so the third key result of objective 2 stood at zero. They didn't change the target. They moved a developer from the spreadsheet import to the button, which shipped in week 9.
How to score a key result
Each key result gets a score from 0.0 to 1.0. When higher is better, subtract the start from the result and divide by the distance from start to target. When lower is better, divide how far you came down by how far you meant to come down. A yes-or-no key result scores 1.0 or 0.0. Anything past the target counts as 1.0.
The objective's score is the average of its key results.
The scores at the end of the quarter
Objective 1, aspirational
| Key result | Start | Target | Result | Score |
|---|---|---|---|---|
| First booking within 7 days | 30% | 50% | 44% | 0.7 |
| Median time to a published page | 3 days | 1 day | 2 days | 0.5 |
| Setup questions per 100 new groomers | 40 | 20 | 24 | 0.8 |
Average: 0.67.
Objective 2, aspirational
| Key result | Start | Target | Result | Score |
|---|---|---|---|---|
| No-show rate | 12% | 6% | 8% | 0.67 |
| Reminders confirmed | 0% | 40% | 28% | 0.7 |
| Rebookings from the reminder, per month | 0 | 300 | 90 | 0.3 |
Average: 0.56.
Objective 3, committed
| Key result | Target | Result | Score |
|---|---|---|---|
| Uptime every month | 99.9% or better | met in all three months | 1.0 |
| Mobile load time, 75th percentile | under 2 seconds | 1.9 seconds | 1.0 |
| No double-bookings reported (yes or no) | met | missed, one reported | 0.0 |
Average: 0.67.
What the scores told them
The two aspirational objectives average 0.61. That's a little under Google's 0.7 and inside the 60 to 70 per cent range re:Work calls the sweet spot, so the targets were about right. The weakest key result, rebookings at 0.3, had only one month to work because the button shipped in week 9. The team kept it for the next quarter instead of calling it a failure.
Objective 3 is the one that needs a conversation. It was committed, so 0.67 is a miss, and Google's rule asks for an explanation. The team's explanation was short. The calendar sync rewrite shipped in week 5 with a bug that let two dog owners book the same slot if they confirmed within seconds of each other. One groomer reported it, and the team fixed it two days later. The explanation went on the board next to the score, and a check that blocks a second booking for a taken slot became a committed key result for the next quarter.
Nobody's pay changed because of any of these numbers. That is what lets a team write 300 rebookings a month as a target when it has no idea yet whether dog owners will press the button.
Common mistakes
- Writing the backlog as key results. If every key result starts with "launch", "build" or "ship", you have a project plan. Ask what the launch is meant to change and measure that.
- Too many objectives. Five objectives for a team of six means each one gets a sliver of attention. Three is plenty, and one or two is fine.
- No start value. "No-show rate to 6%" can't be scored if nobody knows it was 12%. Measure first.
- Treating 0.7 as failure. On an aspirational OKR it's the expected result. On a committed one it isn't, which is why the board has to say which is which.
- Setting OKRs and opening them again in week 13. A check-in every few weeks is where initiatives get swapped while there's still time for that to matter.
Frequently asked questions
How many OKRs should a team have?
Google's re:Work guide suggests three to five objectives, each with around three key results. Small teams often do better with two or three objectives for a quarter. If you can't name who works on each objective this week, there are too many.
What is the difference between an OKR and a KPI?
A KPI is a number you track all the time, like uptime or monthly revenue. A key result often uses a KPI, but adds a target and a deadline: uptime of 99.9% or better in every month of this quarter. KPIs tell you how things are going. OKRs say what the team is trying to change this quarter.
What is a good OKR score?
It depends on the type. Google expects aspirational OKRs to average about 0.7, and committed OKRs to reach 1.0. A score of 1.0 on every aspirational OKR usually means the targets were too safe.
Should OKRs be tied to bonuses?
No. John Doerr's site calls OKRs "divorced from compensation", and Google's re:Work guide says OKRs are not employee evaluations. Once a bonus depends on the score, people set targets they know they can hit, and the stretch goals disappear.
Put your OKRs on a board
Open the OKR planning template. It has a row for each objective, three key result columns and a column for initiatives. The sample board belongs to a different made-up team, so replace its objectives with yours and write a start and a target into every key result. If you'd rather start from a sentence, our guide to briefing an AI whiteboard shows how to describe your team and quarter so the first draft is worth editing. When the quarter ends, a retrospective format gives the scoring meeting some structure.