The Complete Goal-Setting & Planning Manual

OKRs are not the problem. OKRs as a one-time exercise in January are the problem. Every company that has 'tried OKRs and they didn't work' actually tried OKR-setting but not OKR-running. The framework is fine; the cadence is what most people skip. This guide is the full operating system: how to set them, how to score them, when to change them, and how to close out a quarter without it being a panic.

Part 1 — Why most quarterly plans fail by week 6

Three patterns kill plans: too many objectives, key results that measure activity not outcome, and no scheduled check-in until the quarter ends. The combination is fatal — by week six, the team has lost the connection between weekly work and the larger objectives, no one has scored anything, and the original plan starts to feel naive. By week ten, half the objectives are quietly abandoned. By week thirteen, the team sets new ones with the same flaws.

The three killers

• Too many objectives. Five means at least one is fake. Three is the sweet spot. Four is the max.
• Activity-measuring KRs. 'Ship the new pricing page' scores 1.0 regardless of whether pricing improved anything. Bad KR. 'Improve trial-to-paid by 15%' is the outcome version.
• No scheduled check-in. Setting them in week 1 and scoring in week 13 means nothing about the quarter changes mid-flight.

Part 2 — The cadence over the framework

Anyone who's read Measure What Matters can write good OKRs in two hours. The 90 days between writing and scoring is where the value lives. Run those 90 days with a cadence and you'll hit your objectives; run them without and you'll wonder why this framework is supposedly so good.

The 4-loop cadence

• Weekly: each owner posts a 3-line update on their KR — actuals, blockers, asks. Async, 5 minutes max.
• Bi-weekly: 30-minute leadership sync. Each owner self-scores 0.0-1.0; anyone under 0.7 explains the gap.
• Mid-quarter (week 6): 90-minute deep review. Anything tracking under 0.5 — decide to invest, descope, or pivot. Written-down decision.
• End-of-quarter (week 13): 2-hour public scoring. Each KR scored on a slide, in public, with one paragraph explanation.

Part 3 — Setting fewer objectives

The single hardest part of OKR planning isn't writing the objectives — it's the things you don't choose. Every objective rejected feels like neglect; every objective chosen requires the courage to say no to the others. Most leadership teams duck this courage and end up with seven objectives that get the same shallow attention they'd give to fifteen.

The forcing function

Constrain yourself: exactly 3 objectives. No 'plus one nice-to-have'. No 'we're also tracking X informally'. The constraint is what makes the choice real. If you can't pick three, you don't have a strategy — you have a wish list.

Part 4 — KRs that measure outcomes, not output

Output is what you do. Outcome is what changes because of it. 'Ship 5 features' is output. 'Lift user retention 3 percentage points' is outcome. KRs should always measure outcome, even when it's harder to phrase. Outcome KRs force you to predict whether your output will matter — which is the strategy question disguised as a measurement question.

The output-to-outcome rewrite

• Bad: 'Ship the new pricing page.' → Better: 'Increase trial-to-paid conversion by 15%.'
• Bad: 'Hire 6 engineers.' → Better: 'Reduce time-to-merge of P0 fixes from 2 days to 4 hours.'
• Bad: 'Refactor the auth layer.' → Better: 'Cut auth-related incidents by 75%.'

Part 5 — Tying objectives to projects

Objectives without supporting projects are vibes. Projects without parent objectives are busy work. Every objective should link to 1-3 projects that are how the work actually gets done. Every project should link back to an objective it serves. When you find a project that doesn't ladder up to any objective, that's a decision moment — keep doing it (and acknowledge it's not strategic), or stop doing it.

Part 6 — The 14-day check-in

Bi-weekly. 30 minutes. Every owner self-scores 0.0-1.0 on each KR, in writing, before the meeting. Anyone below 0.7 has 90 seconds to explain why and what changes this week. Anyone above 0.9 explains whether the KR was too easy. Anyone exactly 0.7 explains what's keeping it there. The point isn't blame — it's collective situation-awareness.

The score-explain template

Each KR update is three lines: current actual vs target, score (0.0-1.0), and one sentence about what changes this week. That's it. Three lines × 6 KRs = 18 lines posted by the owner the day before the meeting. The meeting itself is 30 minutes of discussion, not 30 minutes of presentation.

Part 7 — Mid-quarter pivots done right

Sometimes a KR is just wrong. The world changed, the data came in worse than expected, or the strategy needs to flex. Changing a KR mid-quarter is sometimes right — but doing it silently is never right. Write a one-paragraph memo: what we set, why it's not working, what we're changing it to, what the new target is. Then update the OKR page with the change history visible. Future-you needs to see both numbers to learn from this.

Part 8 — Closing the quarter

End of quarter score in public. Every owner presents their KRs on a slide, in front of the company, with the actual final score (0.0-1.0) and one paragraph: what we learned. The 0.7s are celebrated. The 1.0s are scrutinized (were they ambitious enough?). The 0.4s are discussed without shame (what did we get wrong about the world?). The honesty of the scoring is what makes the next quarter's OKRs better than this one's.

The lessons-into-next-quarter ritual

End-of-quarter scoring is the start of next-quarter planning. The two events should happen within the same week, not a month apart. Score Friday; plan Monday. The lessons from this quarter — what we got right, what we underestimated, what we never expected — are still vivid. Use them.

Closing

Plans don't survive contact with reality. Cadence does. Build the cadence, and the plan becomes a living thing — adjusted weekly, pivoted at mid-quarter, scored publicly. That's the loop. Everything else is paperwork.

Set goals you can land at 0.7 if you do your best. Not 1.0. Not 0.5.