Here is the thing about OKRs that almost every explainer buries under the definitions: if your team hits 100% of its goals, you did something wrong. Not a moral wrong — a design wrong. You set the bar low enough to clear it, which is the one outcome the whole framework is built to prevent. Google grades OKRs on a scale where the target is a 70%, and treats a perfect score as a warning sign. That single inversion — where falling short is the plan and full marks is the failure — is what separates OKRs from every other goal system, and it’s where most teams get them wrong. So we’ll start there, not with the history.

Why a perfect OKR score means you aimed too low

At the end of a cycle, each key result gets a grade from 0.0 to 1.0. You average the key results to score the objective. So far this looks like any progress bar. The twist is the target: you are not trying to reach 1.0. Per Google’s own re:Work guidance, “the sweet spot for OKRs is somewhere in the 60-70% range,” and the expectation is “to get an average of 0.6 to 0.7 across all OKRs.” Google adds, with a straight face, that this “tolerance for ‘failure’ to hit the uncomfortable goals is itself uncomfortable” for teams new to the method.

Read the score like a diagnostic instrument rather than a report card:

• Consistently landing near 1.0 is not a triumph. It means the goals were sandbagged — you set targets you already knew you’d hit. You learned nothing about the edge of what’s possible.
• Consistently landing below 0.4 means you overreached or badly misjudged, and it’s worth a post-mortem on why.
• Landing around 0.7 means you set a real stretch goal and made strong, imperfect progress toward it. That is the win.

The logic is almost economic. A goal you’re certain to reach carries no information. A goal you reach 70% of the time sits right at the frontier of your capability, which is the only place where the number teaches you anything about how far you can actually push. The framework deliberately trades the comfort of a green dashboard for that information. If your OKRs are always green, you’ve optimized for the wrong thing — you’ve optimized for looking good instead of finding your ceiling.

This is precisely the part that generic goal advice — “set achievable targets,” “make it realistic” — gets backwards. SMART goals prize achievability. OKRs, in their aspirational form, treat guaranteed achievability as a defect.

Committed vs. aspirational OKRs: why one number can’t grade both

The 0.7 rule collapses the instant you apply it to the wrong kind of goal. “Ship the SOC 2 compliance feature by the audit date” is not a goal you want to hit 70% of. You want 100%, or the audit fails. So Google splits OKRs into two categories that are graded by completely different rules, and conflating them is the fastest way to break a team.

Committed OKRs are promises. They are essentially binary. In re:Work’s framing, a committed OKR is expected to hit 1.0 — pass/fail, no middle ground. Scoring 0.9 on a committed OKR is not “almost there”; it’s a miss that demands an explanation. Contractual SLAs, a compliance deadline, a launch a customer is counting on — these belong here. The whole point of labeling something “committed” is to say: we are staking our credibility on this, and we will replan around it if it slips.

Aspirational OKRs (also called moonshots or stretch goals) are the opposite bet. They’re supposed to be uncomfortable. Google grades them with an explicit color band rather than a single pass/fail line, as documented in the OKR scoring guidance on What Matters, the FAQ site run by OKR evangelist John Doerr:

Aspirational OKR score	Color	What it means
0.7 – 1.0	Green	”We’re on track; the things we’re doing are working.”
0.4 – 0.6	Yellow	”We’re at risk; we might need to adjust our approach.”
0.0 – 0.3	Red	”We likely won’t meet our KRs. Time for a hard look.”

Now watch what a 0.6 does. On an aspirational OKR it’s yellow — a middling-but-fine result on a hard goal. On a committed OKR it’s a full-blown failure of a promise. Same number, opposite verdict. That’s why the label isn’t decoration. If a team writes ten OKRs and doesn’t mark which are committed and which are aspirational, then at grading time nobody can say whether a 0.6 means “good stretch, keep going” or “we broke a commitment.” The score becomes noise. Most teams that complain OKRs “don’t tell us anything” skipped this step.

A practical ratio from Google’s practice: keep committed OKRs to the genuine must-dos, and reserve aspirational OKRs — where an average of 0.7 with high variance is healthy — for the ambitious bets. If everything is committed, you’ve just renamed your task list. If everything is aspirational, nothing is guaranteed and your customers notice.

What the grading system does that KPIs and SMART goals can’t

Strip the grading out and OKRs are just goals with metrics attached — which is what KPIs and SMART goals already are. The 0.7 target and the committed/aspirational split are the entire differentiator, and they do one specific job: they make ambition safe to attempt.

Every other framework punishes a miss. A missed KPI target is a red number somebody has to explain. A missed SMART goal — the “A” literally stands for Achievable — reads as a planning failure, because the goal was supposed to be reachable by design. Under those rules, the rational move is always to lowball. Promise what you know you can deliver, protect the green dashboard, never risk a public shortfall.

OKRs invert the payoff. By declaring up front that 0.7 is success and that aspirational goals are expected to fall short, the framework removes the penalty for reaching too far. A team can write “triple activation” knowing that a 0.6 won’t be read as failure — it’ll be read as a bold swing that made real progress. That psychological cover is the actual product. As Doerr puts it in Measure What Matters, OKRs are a vessel for setting goals that are “just beyond the threshold of what seems possible.”

So the honest one-line answer to “what are OKRs” is not “objectives with measurable key results.” Plenty of frameworks have that. It’s: a scoring convention that makes it rational to set goals you might not reach. Everything else — the two-part structure, the quarterly cadence, the cascading — is scaffolding around that idea.

What “OKR” actually stands for, in 60 seconds

With the point established, the mechanics are quick. OKR is Objectives and Key Results, a framework with two parts:

• Objective — one qualitative, memorable statement of what you want to achieve. Significant, concrete, action-oriented. It’s the direction, not the number.
• Key Results — the three to five measurable benchmarks that prove you got there, each specific, time-bound, and verifiable, per What Matters.

Doerr’s template fits it in one sentence: “I will (Objective) as measured by (Key Results).” Objective: “Win the enterprise segment.” Key result: “Close 20 accounts over $50k in Q3.” The objective inspires; the key result keeps you honest.

The rule that separates a real key result from a wish: a key result is a number, not a to-do. “Launch the new pricing page” is a task — you can check it off whether or not it moved anything. “Increase trial-to-paid conversion from 4% to 6%” is a key result. If you can complete it, it’s a task. If you have to measure it, it’s a key result. Hold that line, because the most common OKR failure is a task in a key result’s clothing, and we’ll come back to it.

How OKRs traveled from Grove’s Intel to Doerr’s Google

OKRs weren’t born at Google. They were built at Intel by Andy Grove, on top of Peter Drucker’s 1950s idea of Management by Objectives — agreeing on goals with people rather than dictating tasks. Grove added the missing half, measurable key results, and named his version iMBOs, “Intel Management by Objectives.” He wrote the mechanics down in his 1983 book High Output Management, where the whole method reduces to two questions: Where do I want to go? (the objective) and How will I pace myself to see if I’m getting there? (the key results). Grove’s blunt framing for why execution beat credentials, quoted by What Matters: “It almost doesn’t matter what you know. It’s what you can do with whatever you know.”

The bridge to the rest of the Valley was John Doerr, who sat through Grove’s course at Intel in the 1970s, tightened the name to “OKRs,” and became the framework’s traveling salesman. In the fall of 1999 — as a Kleiner Perkins VC who’d just put $11.8 million into a tiny search startup — he pitched OKRs to Google’s founders around a ping-pong table. Google adopted the method at roughly 40 people and never dropped it, scaling to tens of thousands with OKRs running the whole way. Doerr told that story in his 2018 book Measure What Matters, still the best-known OKR text.

Structurally, OKRs both cascade and cycle. Leadership sets a few company objectives; teams write OKRs that ladder up; a team’s key result often becomes the objective one level down. And they run on a quarterly rhythm — set at the start, checked (usually weekly) mid-quarter, graded at the end — short enough to stay urgent, long enough to move a real metric. Doerr’s own caution is that cascading should stay loose: the best OKRs mix top-down priorities with bottom-up ideas, not pure dictation. Which is a fine segue, because rigid cascading is one of the ways good OKRs rot.

OKR theater: the five ways good OKRs quietly rot

The framework is simple; the failures are predictable. After years of championing OKRs, product leader Marty Cagan of Silicon Valley Product Group now declines to recommend the technique to most companies he meets, because in practice it “ends up proving a waste of time and effort.” His diagnosis is that companies bolt OKRs onto an output-based, feature-factory culture and get “a contrived mashup of outcomes and features” — the ritual of OKRs without the substance. That’s OKR theater. It shows up in five recurring patterns.

1. Key results that are secretly tasks. Cagan’s central complaint, and the most common error by far: teams list deliverables as key results because “it is very easy to ship a deliverable, yet not solve the underlying problem.” “Redesign the pricing page,” “hire two engineers,” “launch the campaign” — all checkable, none of them outcomes. A real key result is a number that moves: conversion up, time-to-hire down, pipeline generated. When a whole OKR set reads like a project plan, the team is measuring effort and calling it impact.

2. Sandbagging. Setting targets you already know you’ll hit so the end-of-quarter score glows. What Matters defines it plainly as the instinct to “under-promise and over-deliver,” and traces it to a simple fear: “When goals are tied to compensation, employees start playing defense and stop stretching for amazing.” The 0.7 target is the specific antidote — if you’re always scoring 1.0, the diagnosis is sandbagging, not excellence.

3. Tying OKRs to compensation. This is the failure that produces the others. The moment a bonus rides on the score, every rational incentive flips toward lowballing and away from ambition. It’s a textbook case of Goodhart’s Law — when a measure becomes a target, it stops being a good measure. What Matters illustrates it with a GM executive who hit all 25-plus of his numerical targets while shipping a car he admitted was “really not that well” received: “I was handed my numerical goals… If I make them all, that’s success!” Every number green, the actual product a dud. Doerr and Grove both warned against welding OKRs to pay for exactly this reason. Use OKRs to inform performance conversations; never let them mechanically set the payout.

4. Cascading too rigidly. When every team’s OKRs are dictated top-down and mechanically decomposed from the level above, the framework becomes a compliance exercise. Teams lose the bottom-up half Doerr insists on, ownership evaporates, and the OKRs become something done to people rather than by them. Cagan’s broader point lands here: layering strict OKR cascades onto feature teams that don’t own outcomes just formalizes the dysfunction.

5. Set-and-forget. The quiet killer. OKRs written at a kickoff and ignored until the review detach from daily work and resurface, unmet, three months later. Without the weekly check-in, the cadence that makes the framework function never actually runs. The goals aren’t wrong — they’re abandoned.

Notice that none of these are the framework being flawed. Every one is a discipline problem: writing outcomes instead of tasks, keeping score honest, keeping pay out of it, keeping ownership local, and keeping the rhythm alive.

OKR examples that pass the cover-the-objective test

Here’s what well-formed OKRs look like once objective and key results are cleanly separated. Every key result is a number with a starting point and a target — never a task.

Function	Objective	Key Results
Product	Make onboarding something new users finish	KR1: Raise activation rate from 42% to 60% KR2: Cut median time-to-first-value from 9 min to 3 min KR3: Reduce day-1 support tickets from onboarding by 30%
Marketing	Become the obvious answer for our core keyword	KR1: Grow organic sessions from 40k to 70k/mo KR2: Rank in the top 3 for 15 target queries (from 4) KR3: Lift blog-to-signup conversion from 1.1% to 2.0%
Sales	Win the mid-market segment we’ve been missing	KR1: Close 25 accounts in the $25k–$75k band KR2: Increase average deal size from $18k to $30k KR3: Shorten sales cycle from 74 to 55 days
Customer Success	Turn renewals from a scramble into a formality	KR1: Raise net revenue retention from 96% to 108% KR2: Lift 90-day feature adoption from 55% to 75% KR3: Cut logo churn from 2.1% to 1.2% monthly

The test: cover the objectives, read only the key results, and you should still be able to tell whether each team succeeded. If you can’t — if the “key results” are things like “run the campaign” that could be done without moving anything — you’ve written tasks, and you’ve built OKR theater.

OKR vs. KPI vs. SMART goal, side by side

These three blur together constantly, and the confusion produces bad goals. They answer different questions.

	What it is	What it answers	Ambition level	Lives on a…
OKR	A framework: one objective plus 3–5 measurable key results	”What are we trying to change this quarter, and how will we prove it?”	Deliberately a stretch (0.7 = success)	Quarterly cycle
KPI	A single ongoing metric	”How is this thing performing right now?”	Steady-state health, no stretch	Standing dashboard
SMART goal	A checklist for writing one well-formed goal (Specific, Measurable, Achievable, Relevant, Time-bound)	“Is this individual goal written clearly enough to act on?”	Usually achievable by design	Any timeframe

The cleanest split, drawn from What Matters’ comparison: a KPI tells you how you’re doing; an OKR tells you why it matters to get this far, this fast. KPIs evaluate the status quo; OKRs drive change. A KPI is a gauge that runs forever — churn, uptime, monthly revenue. An OKR is a time-boxed campaign to move something.

They interlock. A KPI drifting out of a healthy range often becomes the seed of next quarter’s OKR: churn creeps to 3% (KPI), so “cut churn below 1.5%” becomes a key result (OKR). And a SMART goal isn’t a rival to either — it’s a writing test you can run on any single key result to confirm it’s specific and time-bound. The one place SMART and OKRs genuinely diverge is ambition: SMART’s “Achievable” wants a goal you can reach, while an aspirational OKR wants one you probably can’t. For a deeper look at that checklist, see SMART goals.

Where OKRs die between kickoff and review — and how Carly helps

Of the five failure modes above, set-and-forget is the one that kills the most OKR programs, and it’s the most fixable: the goals are fine, but nothing keeps them alive between the quarterly kickoff and the review. Carly is an AI executive assistant that works across your email, calendar, tasks, CRM, and 200+ connected tools, acting automatically on triggers you set. For OKRs, that means Carly can put the weekly check-ins and the end-of-quarter scoring session on the calendar as real recurring events, nudge each key-result owner before a deadline slips, and pull the underlying metrics from the tools you already use so progress stays visible instead of buried in a doc nobody opens. Carly keeps the rhythm; the leaders still set the OKRs, decide committed versus aspirational, and judge what a win looks like. Carly starts at $35/month.

FAQ

Why shouldn’t you score 1.0 on your OKRs? On an aspirational OKR, consistently hitting 1.0 means you set the bar too low — you sandbagged goals you already knew you could reach, so you learned nothing about your real ceiling. Google targets an average around 0.6–0.7 for exactly this reason. Committed OKRs are the exception: those are promises expected to score a full 1.0, pass/fail.

What’s the difference between committed and aspirational OKRs? Committed OKRs are binary promises graded pass/fail, expected to hit 1.0 — think compliance deadlines or contractual SLAs. Aspirational OKRs are stretch goals where roughly 0.7 is a win and 0.4–0.6 is an acceptable yellow. The same 0.6 score is a solid result on an aspirational OKR and a failed promise on a committed one, which is why you label each up front.

How many OKRs should a team have? Keep it small: roughly 3–5 objectives per team per quarter, 3–5 key results under each. If you have 8 objectives, you don’t have priorities — you have a backlog. Reserve the “committed” label for genuine must-dos; make the rest aspirational.

What’s the difference between an OKR and a KPI? A KPI is a single ongoing metric that tells you how something is performing right now (churn, uptime, revenue) and runs indefinitely on a dashboard. An OKR is a time-boxed framework — an ambitious objective plus measurable key results — designed to drive a specific change this quarter. KPIs measure the status quo; OKRs change it.

Should OKRs be tied to bonuses? No. Linking OKR scores directly to compensation triggers sandbagging — people set easy goals so the numbers look good — and violates Goodhart’s Law, where a measure used as a target stops being a good measure. Use OKRs to focus and align work; keep them out of the formula that sets pay.

Who invented OKRs? Andy Grove created the method at Intel, building on Peter Drucker’s Management by Objectives and calling his version “iMBOs.” He documented it in his 1983 book High Output Management. John Doerr learned it at Intel, coined the name “OKRs,” brought it to Google in 1999, and popularized it worldwide in his 2018 book Measure What Matters.

Related: SMART goals, how to build a 30-60-90 day plan, how to run effective one-on-one meetings, performance review examples, how to delegate tasks, the Pareto principle, time management statistics, and the best AI tools for daily planning.

Related Articles

Guides

SMART Goals: The 5-Part Framework With Real Examples

Jul 3, 2026

Guides

How to Write a 30-60-90 Day Plan (With Templates)

Jul 3, 2026

Guides

The Eisenhower Matrix: A Practical Guide to Prioritizing

Jul 3, 2026