What are OKRs? Let me explain with teeth, a rocket, and sex toys

Lately my circle of friends and acquaintances has expanded again, and conversations keep drifting toward metrics — including OKRs. It turns out that plenty of people, even in IT, have never met the thing.

That is hardly a hanging offence. OKRs came out of company and product management; if someone did not grow up in big tech or spend their best years trapped in meeting rooms with dry-erase markers, blaming them would be odd.

So here is what OKRs are and why companies use them, explained through dentistry. Almost everyone has teeth; a quarterly strategy is less evenly distributed.

At some point, a company stops being a product held together by three people and one heroic spreadsheet. A serious finance team appears, then sales, then a full engineering department with processes of its own. Everyone already has goals, reliability and quality metrics, reports, and one person who knows where the last working cable is buried.

Yet sales is barely connected to engineering, despite depending on new features and competitive advantages. Finance can poke engineering with a stick, on a good day. Marketing brings leads, couriers move orders, support fights fires, and everyone optimizes a private square foot of reality.

Management eventually realizes that office parties are lovely, but the business still needs to scale and climb. The company needs shared goals that will not end up as “Be the best” printed on a mug.

Enter OKRs.

OKR stands for Objectives and Key Results. It is a methodology that connects a clear, meaningful goal to several measurable outcomes.

No, it is not OCD. Though some people come close after spending three days trying to write a single KR without the words “improve,” “increase,” or “make better.”

Some people have 32 teeth; some, like me, have 28. Suppose 18 are perfectly healthy and 10 need crowns, cavity treatment, or replacement fillings. We already have a number: 10 problem teeth.

The Objective could be: “Return my mouth to a state where the dentist can look at the X-ray without a heavy sigh.”

But “go to the dentist ten times” is a poor KR. You can sit in the chair ten times and leave reality mostly undisturbed. We need an outcome, not a collection of disposable shoe covers.

If five of ten teeth are treated, the first KR is at 50%. That is not “good” or “bad”; it is a fact: we covered half the planned distance. Then we ask why — not enough money, time, enamel, or courage.

One important detail: health is not a quarterly report. A real treatment plan comes from a dentist, not a person holding an OKR template. The method can show progress; it cannot diagnose you.

“Fly into space” is both ambitious and complicated. If we watch only the final launch, the result becomes brutally binary: it flew, everyone is brilliant; it did not, everyone face-down in the dust, sprinkling regolith over their heads.

But a mission has many proofs of readiness. One Objective can, and should, have several Key Results.

Suppose the rocket did not launch this period, but five of six results were achieved: the stages were tested, defects closed, permits issued, and the full rehearsal completed. The final KR is technically zero, but saying the team “screwed up everything” means throwing most of reality in the bin.

We were probably too ambitious about the launch date. That is not a disaster if we can now see what is ready, what is not, and why. That is the point of OKRs: not to paint the report green, but to align movement and look honestly at the remaining distance.

Engineering, procurement, legal, logistics, and even the cleaner understand why metal shavings cannot be left near the test stand. The whole team sits in one barrel and knows where it is rolling. Uphill, naturally. Why we are in a barrel and who parked it on a slope can go on the retro agenda.

“Become number one” sounds powerful. So powerful that, without a definition, it means absolutely nothing.

Number one where: the UK, Europe, or among three shops next to one Tube station? By what measure: revenue, units, brand awareness, repeat orders, reviews, or the sheer size of the rubber inventory? Who is the benchmark, and where did their numbers come from?

First, lock the segment down: for example, online sales of own-brand, mid-priced adult products in the UK. Then choose the data sources and measurement period.

Now we can see how one goal reaches the whole company. Engineering improves the catalogue and checkout. Marketing grows reach and awareness. Researchers learn what customers actually want, rather than what occurred to the director after a third espresso. The warehouse sends the right items, the courier does not deliver a box that looks as though it survived the Somme, and support does not answer a human crisis with a canned paragraph.

Even the cleaner affects the process when production or warehouse hygiene is part of safety. But do not strap an OKR to every individual. A shared result comes from a system, not a hundred private spreadsheets in which everyone heroically scores 0.73.

A KPI is not necessarily binary. It measures steady operational health: service availability, conversion, delivery time, defect rate. A KPI asks: “How well is the system working?”

An OKR asks a different question: “What meaningful change do we want to make, and which results will prove that we moved?”

For example, 99.95% platform availability can be a standing KPI. The Objective “Make the service suitable for large dental clinics,” with the KR “Reduce critical incidents from 12 to 2 per quarter,” describes a change the team is making now.

For a metric that should increase, a simple score looks like this:

If repeat purchases rise from 18% to 24% against a target of 30%, you have covered 50% of the distance: (24 − 18) / (30 − 18).

For a metric that should decrease, reverse the direction. If there were 10 problem teeth, now there are 5, and the target is 0, you have also covered 50% of the distance.

Sometimes 70% on a genuinely ambitious OKR is a strong result; sometimes 100% means the goal was too easy. There is no sacred universal threshold. Agree beforehand which OKRs are committed, which are aspirational, and how the scale will be interpreted. Otherwise the number becomes a religion and the spreadsheet its stern, deaf god.

• One goal, one clear direction. Do not weld growth, quality, hiring, and an office move into one Objective.
• Usually 2–5 Key Results. If you have 18, you probably have a work plan, not a set of key results.
• Every KR has a baseline, target, deadline, and data source.
• A KR describes an outcome, not an activity. “Run user research” is a task; “Validate demand among 60% of the target sample” is a result.
• Teams can see dependencies. Sales does not promise what engineering is not building, and engineering knows which change the work is meant to create.
• Progress is discussed regularly. Opening the OKR on the final day of the quarter is not management; it is archaeology.
• A missed goal is examined, not hidden. Otherwise every goal in the next cycle will be the size of that horseshoe nail.

What are OKRs in simple terms?

OKR is a goal-setting method. The Objective defines a meaningful direction, while Key Results show in numbers how far the team has moved toward it.

What does OKR stand for?

OKR stands for Objectives and Key Results. The Objective answers “where are we going, and why?”, while Key Results answer “how will we measure progress?”

How are OKRs different from KPIs?

A KPI usually measures the health and stability of an existing process. An OKR describes a meaningful change and the results to achieve within a chosen period. They can work together; no emperor needs to make them fight to the death.

How many Key Results should one Objective have?

Usually two to five. Too few can give a one-sided view of the goal; eighteen is probably a task list wearing someone else’s helmet.

Is 70 percent a good OKR score?

For a genuinely ambitious OKR, it can be. Committed goals are judged more strictly. Agree on the goal type and scoring rules in advance, rather than digging them out from under the table on the last day of the quarter.