A practical guide to better OKRs

OKR is short for objectives and key results. It is a simple way to agree what a team or organisation wants to change, and how it will know whether the change happened. The objective names the destination in words. The key results put numbers on it, so progress can be checked rather than guessed.

Each OKR has one objective and usually two to four key results, and it runs for a fixed period, most often a quarter. The period matters: it is long enough to achieve something real and short enough to learn and adjust. A team rarely needs more than three OKRs at once, because each one deserves attention every week.

A good objective is a short sentence that says what you want to be true by the end of the period. It uses words rather than figures, and it makes clear which way the team is heading. People should be able to read it once and explain it to a colleague without notes.

Leave the numbers to the key results. An objective written as a bare target, such as a figure to hit by a date, tells people what to count but not why it matters. Start from the reason, then let the key results show what success looks like. If a team cannot say why an objective matters this quarter, it is probably the wrong objective.

• 01Qualitative: it is written in plain words and describes a state the team wants to reach, not a number.
• 02Directional: it shows which way the team is heading and why that direction matters now.
• 03Short and memorable: it fits on one line, so people can recall it in a meeting without looking it up.

A key result is a measurable outcome that shows progress towards the objective. It describes a change in the world, such as fewer complaints or quicker deliveries, rather than a piece of work the team will do. If the work is finished but the measure has not moved, the key result has not been achieved.

Each key result needs a starting point and a target, and it should be possible to see where it stands at any time during the quarter. Two to four key results per objective is usually enough. Taken together, they should make it hard to claim success while missing the point of the objective.

• 01Move a number: take a measure from its baseline to a target, for example raising weekly active users from 1,200 to 1,800.
• 02Hold a threshold: keep a measure above or below a set line for the whole quarter, for example keeping service availability at or above 99.5%.
• 03Reach a milestone: something that is either done or not done by a date, for example a new supplier contract signed and in use by week 8.
• 04Improve a rate: raise or lower a percentage, for example lifting the share of orders delivered on the promised day from 88% to 95%.

Picture a customer support team that hears the same complaint every week: people wait too long for an answer and often have to ask twice. The team agrees that fixing this is its main focus for the next quarter. It writes one objective, ‘Customers get help quickly and only have to ask once’, and three key results under it.

Each key result tests a different part of the objective. The first covers speed, the second covers whether people need to come back, and the third guards against the team rushing replies to hit the other two. None of them names a task, so the team keeps the choice of how to get there, whether that means better help articles, new routing rules or more training.

An output is something a team produces, such as a report, a feature or a training session. An outcome is the change that output is meant to bring about, such as fewer errors or faster onboarding. Key results should describe outcomes, because an output can be delivered on time and still change nothing.

The work itself still matters. Projects and tasks that a team believes will move a key result are often called initiatives. List them under the key result they serve, look at them again as the quarter goes on, and drop the ones that are not helping. The key result stays fixed; the route to it can change.

A key performance indicator, or KPI, is a measure a team watches all the time to know that its work is healthy, such as on-time delivery or service availability. KPIs tell you whether things are running as they should. OKRs are for the few areas where you want something to change, and they last for a set period.

The two work together. When a KPI drifts out of its healthy range, or when a team decides a steady KPI must reach a new level, that measure can become a key result for one quarter. Once it is back where it should be, it returns to the list of measures the team keeps an eye on, and the OKR moves on to the next change.

Most organisations set a small number of company objectives that describe the changes that matter most across the organisation. Teams then set their own OKRs for the part of the work they can influence. Some people also set individual OKRs, usually for a piece of work that sits with them alone.

A team objective should link to the company objective it supports rather than copy it. Copying turns every team's OKR into the same sentence and hides what each team will actually do. Linking keeps the team's own wording and makes the connection clear, so anyone can follow how local work adds up to the company's aims.

Not every person needs individual OKRs. For many roles, contributing to the team's OKRs is the clearer choice and avoids a long list of personal goals that nobody looks at. Individual OKRs make most sense for specialists and leads who own a distinct outcome of their own.

OKRs work best when the same few moments come round every quarter. A predictable rhythm means nobody has to wonder when goals are set, discussed or closed, and it keeps OKRs part of normal work rather than an extra exercise.

• 01Set: in the last two weeks of the previous quarter, draft, discuss and agree the objectives and key results for the next one.
• 02Check in weekly: each owner records progress and confidence once a week, which takes a few minutes per key result.
• 03Review at mid-quarter: around week six or seven, look at what is on track and what is not, and decide whether any key result needs a new approach.
• 04Close and reflect: in the final week, record where each key result ended, note what the team learned and carry those lessons into the next round.

Keep the rhythm the same from one quarter to the next. When teams know what is coming, less time goes on process and more on the work itself. Some organisations set company objectives for a year while teams keep to quarters, which works well as long as each quarter's OKRs link back to the yearly ones.

A weekly check-in records four things for each key result: the current value, how confident the owner is of reaching the target, anything blocking progress, and a sentence of context. The context matters, because a number on its own rarely explains itself. A line such as ‘waiting for the new supplier, expect a jump next week’ saves a meeting.

Short and regular beats long and rare. A two-line update every week gives colleagues an early signal, while a long report once a month tends to arrive after the moment to act has passed. If nothing has changed, say so; a check-in that reports no movement this week is still useful.

At the end of the quarter, each key result gets a score that shows how much of the target was reached, often on a scale from 0 to 1 or as a percentage. The score describes the goal, not the person who owned it. It helps the team see how realistic the target was and what got in the way, so the next set of goals is better judged.

It helps to agree in advance whether a key result is ambitious or committed. An ambitious key result stretches beyond what the team is sure it can do, and getting most of the way there is a good result. A committed key result is a promise others rely on, so anything short of full delivery calls for a calm look at what went wrong.

During the quarter, confidence is more useful than the score. Each week the owner says how likely they think it is that the target will be reached, for example on a scale from 1 to 5. A falling confidence is an early warning that invites help, not blame.

OKRs measure goals, not people. Keep them apart from performance appraisals, bonuses and pay. A key result that ends at 60% may reflect a bold target, a change in the market or a dependency on another team, and none of these says anything fair about one person's work.

When goals feed into a person's rating, people soon learn to set targets they know they can hit. Ambition drops, confidence levels stop being honest, and check-ins start to look better than reality. Keeping the two apart lets people aim high, raise problems early and treat a missed target as something to learn from.

Most early problems with OKRs come from a handful of habits that are easy to spot once you know them.

• 01Too many OKRs: five or more objectives per team spread attention thin, so pick the two or three that matter most this quarter.
• 02Tasks posing as key results: a line such as ‘launch the new form’ is an output, so ask what should change once it is launched and measure that instead.
• 03Business as usual: an OKR that describes work the team would do anyway adds paperwork without changing anything, so keep those measures as KPIs.
• 04Set and forget: OKRs written in the first week and opened again in the last cannot guide any decision, so build a weekly check-in into the routine.
• 05Top-down only: when every goal is handed down, teams lose the chance to say what they can realistically change, so let teams propose their own and agree them together.
• 06Tying OKRs to pay: linking scores to ratings or bonuses pushes people towards safe targets, so hold those conversations separately.

You do not need a large programme to begin. A small, well-run first quarter teaches more than months of preparation, and each later quarter builds on what the first one showed.

• 01Agree one to three company objectives with the leadership team and write them in plain words.
• 02Ask a few teams to draft their own OKRs and link each objective to the company objective it supports.
• 03Pick a fixed weekly check-in day and keep each update to a few lines.
• 04Hold a short review at mid-quarter to adjust the approach, not to rewrite the targets.
• 05Close the quarter with a team retrospective and use what you learn to set the next round.