OKR Examples: Objectives and Key Results by Department
Eight sets of objectives and key results by department, ten “bad KR → good KR” pairs, and the structure of a tracker to run them in.
700+ practitioners
Join the communityPeople read OKR examples for the shape of the thing: what an objective looks like, which key results sit under it, and how those results differ from the task list you are working through anyway. Below you will find eight blocks by department, a table of ten "bad KR → good KR" pairs, the structure of a working tracker, and answers to the questions that come up in the first workshops.
A warning before you see the first example: every number on this page is illustrative. We chose them to show the shape of the record — baseline, target value, unit of measurement, data source. These are not benchmarks. Your 18% conversion may be an outstanding result or a failure, depending on the market and the segment. The warning applies to every number below — in the explanations, in the departmental examples, and in the corrections table.
Peter Drucker described management by objectives (MBO), Andrew Grove reworked the approach at Intel and called his version iMBOs (Intel MBOs), and John Doerr brought the method to Google in 1999, presenting it to Larry Page and the team. Today Intel, Google, Amazon, Microsoft and the Government Digital Service run on Objectives and Key Results, and in Ukraine, Uklon and Monobank. We cover the mechanics of the method on the OKR page.
How to read an OKR example
Objective — a qualitative statement of where you are going and why. The team should read it and want to do it. Keep the objective to one sentence and put no number in it. If there is a percentage in the objective, you have written a Key Result and called it an objective.
Key Result — a measurable outcome, not an activity. It has a metric, a baseline, a target value, and a source you will check it against. "Run 12 interviews" is an activity. "Cut first-month churn from 19% to 11%" is an outcome.
Initiative — what you do to move a KR. A project, an experiment, a process change. You keep initiatives separately from the OKRs and change them within the quarter when a hypothesis does not pay off.
Here is how to tell them apart. Put the wording into the sentence "We did this, and now ___". If the ending sounds like "and now we have a new landing page", you are looking at an initiative. If it is "and now the pricing page converts at 3.1% instead of 1.8%", you are looking at a Key Result. A second check: name the report where you will see the number, and the person who will confirm it. No report, no Key Result.
The number of key results per objective, from our practice: 3–5. Fewer and you rest the objective on a single metric and lose it together with that metric. More and the team spreads itself thin and spends the check-in discussing a spreadsheet instead of the work.
OKR examples by department
Sales
Objective: We close the quarter without a scramble in the final week.
Key Results:
- Share of deals closed in the forecast quarter rises from 48% to 70% (source: the CRM funnel).
- Average deal cycle in the mid-market segment shortens from 74 to 55 days (source: the closed-deals report in the CRM).
- Qualified-lead-to-signed-contract conversion rises from 12% to 18% (source: the CRM funnel).
- Revenue from repeat sales to existing customers rises from UAH 2.1 to 3.4 million (source: the sales ledger report).
- Share of deals stuck in "under review" for more than 30 days falls from 26% to 10% (source: the weekly pipeline snapshot).
Initiatives: introduce a single qualification checklist in the CRM and block a deal from moving on until the fields are filled in; rewrite the commercial proposal separately for three segments and test it on 30 deals, comparing conversion against a control group.
We showed the data source in every KR in this block — below we drop it for brevity. In your own OKRs you always write the source.
Marketing
Objective: We bring in enquiries that sales picks up without arguing about quality.
Key Results:
- Share of enquiries from organic search rises from 18% to 35%.
- Cost per qualified enquiry falls from UAH 940 to UAH 600.
- Number of enquiries the sales team accepted as qualified rises from 40 to 90 per month.
- Share of assets producing at least one enquiry per month rises from 20% to 45%.
- Share of enquiries sales returns as off-target falls from 34% to 15%.
Initiatives: build 12 pages for bottom-of-funnel queries and interlink them to the hub; set up end-to-end analytics from click to signed contract, so that marketers and salespeople look at one number instead of reconciling two different reports in a meeting.
Product
Objective: A new user reaches first value on their own, without emailing support.
Key Results:
- Share of new accounts completing the key action within the first 7 days rises from 22% to 45%.
- Support tickets per 100 new accounts fall from 31 to 12.
- Day-30 retention rises from 38% to 52%.
- NPS among first-month users rises from 14 to 30.
- Time from sign-up to the first successful scenario shortens from 3 days to 40 minutes.
Initiatives: rebuild onboarding into three steps with in-product hints and a template for the first project; run 8 interviews with people who stopped at the second step and close the two most common reasons for stopping by mid-quarter.
Engineering
Objective: We ship changes daily instead of preparing for a release all week.
Key Results:
- Time from commit to production shortens from 6 days to 8 hours.
- Share of releases the team rolls back falls from 17% to 5%.
- Time to recover from an incident shortens from 3 hours 40 minutes to 45 minutes.
- Automated test coverage of critical scenarios rises from 35% to 80%.
- Releases per week rise from 1 to 10.
Initiatives: move the build into a separate pipeline with automatic rollback on error metrics and feature flags; introduce an on-call rota and a single template for reviewing every incident, so the team sees recurring causes rather than isolated cases and carries one fix into the next sprint.
HR
Objective: Everyone on the team understands where they are growing inside the company.
Key Results:
- Voluntary attrition in key roles falls from 19% to 10% annually.
- Time to fill an engineering vacancy shortens from 68 to 40 days.
- Share of employees with an agreed six-month development plan rises from 25% to 85%.
- Share of new hires passing probation rises from 74% to 90%.
- Share of internal moves among all filled vacancies rises from 8% to 25%.
Initiatives: describe the levels and promotion criteria for three tracks and publish them for the whole team; move the development conversation and the salary review conversation into different months, so that manager and employee discuss growth separately from money.
Finance
Objective: Executives make decisions on fresh numbers rather than on gut feel.
Key Results:
- The team closes the month on the 5th working day instead of the 14th.
- Share of costs allocated to cost centres rises from 60% to 95%.
- Variance of actuals against the cash flow forecast narrows from ±22% to ±8%.
Initiatives: move source documents into electronic document management with the two largest suppliers and bring contractor certificates there as well; agree a single cost item directory with the executives and close off the ability to create new items without the CFO's approval.
Manufacturing
Objective: We deliver orders on time throughout the month, without a rush in the final week.
Key Results:
- On-time order fulfilment rises from 71% to 92%.
- Defect rate at the end of the line falls from 4.3% to 1.5%.
- Unplanned equipment downtime falls from 46 to 15 hours per month.
- Line changeover time shortens from 95 to 40 minutes.
- Share of orders shop supervisors hand over in the last three days of the month falls from 45% to 20%.
Initiatives: start a morning supervisors' stand-up at the order board, reviewing yesterday's failures and deciding for the day; draw up a planned maintenance schedule for the four units that cause the most downtime, buy a quarter's stock of critical parts, and assign a responsible mechanic to each unit.
Public sector and communities
Objective: A resident receives the service online and does not come into the office.
Key Results:
- Number of services available online rises from 12 to 35 out of the 60 that exist.
- Average time to issue a certificate shortens from 9 to 2 working days.
- Residents' rating of service convenience rises from 3.2 to 4.3 on a five-point scale.
Initiatives: map the 10 most frequent services as journeys with their failure points and rewrite the document requirements in plain language; train 25 service centre staff on the new system and appoint two internal trainers to carry the training on after the quarter.
Bad KR → good KR: ten pairs
Half the poor wordings below are really KPIs dressed up as key results: a metric with no starting number and no date. Where the line runs is covered separately — how a key result differs from a KPI.
| Bad KR | Good KR | What was fixed |
|---|---|---|
| Make 40 cold calls a week | Raise cold-contact-to-meeting conversion from 4% to 9% | Activity metric → outcome metric |
| Improve customer satisfaction | Raise post-support CSAT from 3.9 to 4.5 on a five-point scale | Metric, scale and baseline added |
| Launch a new website and increase sales | KR 1: cut catalogue page load time on the new engine from 4.2 to 1.5 seconds. KR 2: raise pricing page conversion from 1.8% to 3% | Two goals in one KR split into two, both with a baseline |
| Strengthen the company culture | Raise eNPS from 41 to 60 in the March and June surveys | KR could not be verified → instrument and measurement dates named |
| Keep a register of contracts | Cut the share of contracts signed more than 5 days late from 30% to 8% | The author rewrote a job description as a KR → a shift appeared |
| Publish 20 vacancies | Raise the number of engineering vacancies filled by candidates who passed probation from 2 to 5 per quarter | Activity metric → outcome metric with a baseline |
| Optimise shop floor processes | Cut line changeover time from 95 to 40 minutes | Vague wording → a concrete metric |
| Cut costs by 10% | Cut logistics costs from UAH 1.4 to 1.1 million per quarter while holding on-time fulfilment at no less than 90% | No baseline and no subject → baseline, subject and a guardrail |
| Run 12 customer interviews | Cut first-month churn from 19% to 11% | Activity metric → outcome metric |
| Become the market leader | Raise the share of tenders won in the segment from 8% to 14% per Prozorro data | KR could not be verified → data source named |
You will have spotted the pattern: most of the corrections come down to three questions. What exactly will change in the world? From which value to which? Who will see it, and where?
OKR template: which columns to set up
Here we close the "OKR template" request without a file to download — build the sheet yourself in Google Sheets or Excel, and it becomes your OKR tracker.
Set up one row per Key Result and these columns: Period (quarter), Level (company / team), Objective, Objective owner, Key Result, Metric, Data source, Baseline, Target, Current value, Progress, Status, Initiatives, Check-in comment.
How to fill it in. You set the baseline once at the start of the quarter and do not move it afterwards — without it you cannot calculate progress. In the "Data source" column write specifics: the name of the report, the dashboard, the survey number. The owner is one person, not a department.
How to score it. Calculate progress as (current − baseline) / (target − baseline), giving a result from 0 to 1. The owner sets the status by hand: on track, at risk, stopped. A mismatch between high progress and an "at risk" status is a normal situation, which the owner explains at the check-in.
We recommend keeping the check-in within 15–30 minutes and holding it weekly or every other week. In it the owner updates three columns — current value, status, comment — and names what they are changing in the initiatives.
Agree on the type of goal before you start. Andrew Grove rebuilt Drucker's management by objectives at Intel into a system where the team adds measurable results to every goal. The split into committed and aspirational goals is Google's practice: John Doerr describes it in "Measure What Matters", and Google itself repeats it in its re:Work materials. A committed goal there means a commitment: the team plans resources so as to close it fully or almost fully. An aspirational goal the team deliberately sets above the achievable, and for those goals the Google re:Work guide names 0.6–0.7 as the benchmark. These are two different contracts with the team, and mixing them within one sheet is not a good idea. We have gathered further materials in the OKR UA library.
Why OKR examples from the internet don't work in your company
The first trap is someone else's context. You take the wording, and with it you take someone else's market, someone else's stage, someone else's baseline. A team that sets day-30 retention at 52% knows its current value and the reasons for churn. You do not know those — and the copy turns into a wish.
The second trap is translation from English-language blogs. Along with the metrics you take instruments and roles you do not have, and with them the illusion that you already know how to measure this.
The third is an example without initiatives. You see a handsome objective and a few KRs, but you do not see what the team was going to use to move them, which hypotheses it tested, and what it dropped in the second month.
The fourth is a top-down cascade without a conversation. The leader breaks the company goals down across departments, the department heads accept them in silence, and over the quarter nobody remembers why these particular metrics were chosen.
The fifth is an example someone assembled for ideal data. The author of such an example has analytics in place, while you discover at the start that half of your metrics are not being counted by anyone. You then often spend the first quarter learning how to measure, and that is a normal outcome.
Take the shape from the examples, and mine the substance from your own data and from a conversation with your team. We have described how we run that conversation in the How we work section.
FAQ
How many Key Results should there be per objective?
From our practice, 3–5. That is the number a team holds in its head and gets through at a check-in. If you end up with more metrics, check the list: some of them are initiatives rather than results.
Can a Key Result have no number?
A binary KR is acceptable as an exception: "certification passed — yes/no". You verify it by fact and date. But if most of your KRs are binary, you have written a project plan rather than results. Look for a metric even where it is inconvenient.
Should OKRs be set for an individual?
Set them for the team. Set OKRs on a person and sooner or later you will tie them to a bonus, and that tie makes people lower their goals. Discuss individual contribution separately, in the development conversation.
What do we do with a KR delivered at 40%?
First take two versions apart at the retro: was the hypothesis wrong, or did the initiatives fail to work. Depending on the answer you change what you do next quarter. If you agreed on a committed goal, this is a shortfall, and the team examines the reasons. If the goal was aspirational, 40% is a signal to check whether you set a metric you do not influence.
How does an example differ from a template?
In an example you see filled-in content for a specific department. In a template you see an empty structure with columns and a scoring rule. You read an example once, to understand the shape. You keep a template running all quarter.
How often should the sheet be updated?
We recommend once a week or once every two weeks, in a 15–30 minute check-in. More often and you risk reacting to noise in the data. Less often and you find out about a failure once the quarter is already over.
If you would like to go through your draft OKRs with practitioners, come to the community or get in touch.
Next
Ready to set your own objectives?
Examples show the shape. You write the objectives — we carry you through the first cycle so it does not stop at a list of intentions.