What is a Key Result (OKR)?

A Key Result is the quantitative, time-bound outcome used to measure progress toward an Objective in the OKR framework. Each Objective typically has three to five Key Results, each one a number that moves from a defined baseline to a target inside a quarter.

If you cannot tell whether it hit or missed without an argument, it is not a Key Result.

TL;DR

Measurable by definition: A Key Result names a metric, a baseline, and a target inside a fixed window, so success is unambiguous at quarter-end.
3 to 5 per Objective: Fewer means the Objective is under-defined; more dilutes focus and forces teams to optimize for tracking rather than outcomes.
Outcomes, not tasks: "Ship the new pricing page" is a task; "Increase free-to-paid conversion from 2.1% to 3.5%" is a Key Result. The difference is what predicts strategy execution.
Scored, not pass/fail: Most teams score Key Results on a 0.0 to 1.0 scale where 0.7 is the stretch sweet spot. Consistently hitting 1.0 usually means the targets were too soft.

Definition: A Key Result (OKR) is a quantitative and measurable outcome used to track and assess the success and progress towards achieving an Objective, a desired goal or aspiration within an organization.

How Andy Grove defined a Key Result

Andy Grove introduced Key Results in "High Output Management" (1983) as a discipline borrowed from manufacturing: every Objective needs a small set of measurable outcomes so that arguments about whether a goal was achieved become impossible. The structure spread from Intel to Google through John Doerr in 1999, and from Google to most modern tech companies through Doerr's 2018 book "Measure What Matters".

The key result has to be measurable. But at the end you can look, and without any arguments: did I do that or did I not do it? Yes? No? Simple. No judgments in it.

― Andy Grove, former CEO of Intel and inventor of the OKR framework

Grove's test still holds. If two reasonable people can disagree at the end of the quarter about whether a Key Result was hit, the Key Result is broken, not the people.

The two pieces of a Key Result: metric and target

Every Key Result has two components, and Key Results that drift in practice usually break at one of them.

The metric. A number that already exists, or can be measured cheaply, and that the team has at least partial control over. Examples: monthly active users, free-to-paid conversion rate, average response time, defect rate per release.
The target. A specific value the team commits to moving the metric to, expressed as either an absolute number ("reach 50,000 monthly active users") or a delta ("reduce response time from 4 hours to 90 minutes").

A Key Result without a baseline is not measurable in Grove's sense, because the team has no anchor to argue from. A Key Result without a target is just a metric the team is watching, which is a KPI, not a goal.

Aspirational vs committed Key Results

Google formalized two flavors of Key Result, both of which appear inside a normal OKR set. The distinction matters because the scoring rule for each is different, and conflating them is one of the most common OKR mistakes.

Dimension	Committed Key Results	Aspirational Key Results
Expected score	1.0 (must hit)	0.6 to 0.7 (stretch sweet spot)
Failure cost	High: a missed commitment damages trust across teams	Low: missing a moonshot is the expected case
Use case	Reliability, compliance, baseline product work	New markets, moonshot goals, bold bets
Typical share of OKR set	Roughly 30 to 40%	Roughly 60 to 70%
Review question	"Did we ship it?"	"What did we learn, and how far did we get?"

A healthy quarterly OKR set usually contains both. Pure aspirational sets create a culture where nothing reliably ships; pure committed sets remove the upside that makes the framework worth running.

What makes a Key Result actually work

A Key Result that survives a full quarter and produces useful information shares four properties. The first is non-negotiable, the others are where craft shows up.

It is a number, not a verb. "Launch a referral program" is a project. "Drive 8% of new signups through referrals by Sep 30" is a Key Result. Project work supports Key Results, but it is never the Key Result itself.
The team has a credible plan for the first 30% of the move. If the team cannot describe how they will move the metric in week one, the Key Result is a wish, not a goal.
It points at an outcome, not an activity. "Publish 12 blog posts" tracks output; "Grow organic search traffic from 18k to 30k monthly visits" tracks outcome. Outcome Key Results survive scope changes; activity Key Results do not.
It is owned by one team, even when it crosses functions. Shared accountability sounds collaborative and usually produces neither.

The fastest test in practice: read the Key Result out loud at the start of the quarter, and again at the end. If you can score it without checking a definition document, it works.

If you need a meeting to interpret it, rewrite it before the quarter starts.

When a Key Result is actually a task in disguise

The most common Key Result failure is not ambition; it is the milestone-as-metric pattern. A team writes "Launch the new pricing page" or "Complete the SOC 2 audit" and counts the Key Result as 1.0 when the work is done.

The shortcut feels productive. It is also the fastest way to build an OKR culture that tracks activity instead of impact, because the metric stops being a number and starts being a checklist item.

Three signals that a Key Result is a task in disguise:

It can only be scored as 0 or 1. Real Key Results have a continuous scoring range. If "done" and "not done" are the only outcomes, the work is a milestone, not a measure.
The owner is the team that does the work, not the team that benefits. Engineering "launching the pricing page" tracks delivery. Marketing "lifting conversion on the pricing page" tracks value.
It would survive a successful launch with no business change. If the page ships, the audit passes, the system gets deployed, and nothing measurable about the business shifts, the Key Result was tracking the wrong thing.

Tasks are still essential. They belong inside the team's execution plan underneath the Objective, not as the Key Results themselves.

Scoring Key Results at the end of the cycle

OKR scoring is what turns Key Results from a tracking tool into a learning tool. Most teams use Google's 0.0 to 1.0 scale, scored at the end-of-cycle review.

0.7 to 1.0 on aspirational Key Results: the target was set correctly. The team stretched and delivered most of the move.
0.4 to 0.6: real progress, but the target was either too ambitious or the plan missed a constraint. Useful learning either way.
Below 0.4: the Key Result was poorly chosen, the priority shifted mid-quarter, or the team ran into a blocker nobody surfaced early enough.
Consistent 1.0 across the board: sandbagged targets. The framework is being used as a status report, not a goal-setting system.

A Key Result that scores 0.7 is doing its job. A Key Result that scores 1.0 every quarter is not aspirational.

According to the OKR Impact Report 2022, 83% of companies using OKRs report a positive impact on their organization, and the most successful adopters review Key Result progress at least bi-weekly through structured check-ins.

Using Key Results inside your OKR cycle

The cleanest pattern for embedding Key Results into a working quarterly cycle:

Draft. Each team writes 3 to 5 Key Results per Objective during planning week, with both metric and target named explicitly.
Pressure-test. A peer team or OKR coach reviews the draft against three questions: Is it a number? Is there a credible plan for the first 30%? Does it map to a strategic goal?
Commit. Publish the Key Results visibly to everyone who depends on them. Hidden Key Results don't drive alignment.
Check in. Every one to two weeks, the owner posts a current value, a confidence level (red, yellow, green), and the next step. No theater, no slides.
Score and retire. At quarter-end, score each Key Result, write a one-paragraph reflection, and decide whether the underlying problem is solved, persists, or has shifted. A Sears Holding study found teams using OKRs consistently across all quarterly cycles saw an 8.5% increase in sales per hour, while inconsistent users (one to three cycles per year) saw only 3% (Lamorte, 2018).

The retire step is where most OKR programs leak value. Key Results that quietly carry forward from one quarter to the next, unscored and unreflected on, accumulate into the kind of OKR fatigue that ends programs by year two.

How many Key Results should an Objective have?

Three to five is the working range used by Google, Intel, and most modern OKR programs. Fewer than three usually means the Objective is under-defined; more than five splits team attention and forces tracking ceremony to replace real focus. If you need more Key Results to express the Objective, the Objective probably needs to be split into two.

What is the difference between a Key Result and a KPI?

A KPI is a metric you track continuously to monitor health, with no target attached. A Key Result is the same kind of metric with a baseline, a target, and a deadline attached, used to drive a specific change inside a quarter. Many KPIs become Key Results when a team commits to moving them.

Can a Key Result be a project or a milestone?

Strictly, no. A Key Result names an outcome the team is trying to produce, not an activity the team is doing. "Launch the new pricing page" is a milestone; "Lift free-to-paid conversion from 2.1% to 3.5%" is the Key Result the new page is supposed to drive. Milestones belong in the execution plan underneath the Objective, not as the Key Results themselves.

What's a good score for a Key Result?

On aspirational Key Results, 0.7 on a 0.0 to 1.0 scale is the target zone. On committed Key Results, anything less than 1.0 counts as a miss. Consistent 1.0 scores across all Key Results usually mean the targets were sandbagged, which is why Google explicitly designs aspirational Key Results to score below 1.0 on purpose.

Who owns a Key Result?

A single team or person owns each Key Result, even when delivering it requires work from several functions. Shared ownership sounds collaborative but typically produces neither accountability nor follow-through. The owner is the person responsible for posting the current value, confidence level, and next step at every check-in.

How often should Key Results be reviewed?

Bi-weekly is the most common cadence among high-performing OKR programs, and the OKR Impact Report 2022 found that the most successful companies have 28% higher communication intensity around Key Result progress than the comparison group. Weekly tends to over-instrument; monthly lets Key Results drift off track for too long before anyone surfaces it.