── ── Mental model
OKRs (Objectives and Key Results)
OKRs separate ambition from measurement: an Objective is qualitative and aspirational; Key Results (3-5) are quantitative outcomes proving the objective was reached. KRs must be outcomes, not activities. Calibration rule: 70% achievement is success — routine 100% means goals were sandbagged. Developed by Andy Grove at Intel (1971); introduced to Google by John Doerr (1999).
Run OKRs (Objectives and Key Results) on a real problem
Bring something you're actually deciding — free, in the browser.
How it works
Step 1 — Write the Objective. Qualitative, aspirational, time-bounded, one sentence. Test: "If achieved, would this objectively matter to the business?"
Step 2 — Write 3-5 Key Results. Quantitative outcomes (not activities), time-bounded, 70%-success-calibrated. "Ship v2.0" = activity (wrong). "Reach 100k WAU on v2.0 within 30 days" = outcome (right). If all KRs happen and the Objective is not achieved, the KRs are wrong — iterate.
Step 3 — Calibrate ambition. >85% probability per KR → sandbagged, stretch it. <30% → unreasonable, compress. Target: 50-70% probability.
When to use it
- user says 'set OKRs', 'write objectives and key results', 'our goals aren't measurable', 'teams are hitting targets but the business isn't moving', 'we need stretch goals', 'align teams around quarterly goals', or mentions Doerr / Measure What Matters
When not to use it
When the decision is routine and reversible, applying a formal method costs more than it returns.
Worked example
OKRs in an AI-Native Org (2024–2026)
By 2024–2026, "add AI" had become the default objective for nearly every product team. The trap is structural: an objective like "ship AI features" or "increase AI usage" invites exactly the vanity and Goodhart metrics OKRs exist to prevent. This is a composite, illustrative case — a small, fast product team (roughly 8 engineers plus a PM and a designer) putting an AI assistant into an existing SaaS product — chosen to show how the six Process steps discipline AI bets. The numbers below are…
Install this skill (free, MIT)
npx skills add deciqAI/knowledge-skillsUseful? Star the repo — stars help other builders find it.
Related mental models
Colonel John Boyd (1927-1997) derived the OODA Loop from Korean War air-combat data: the F-86 Sabre achieved a ~10:1 kill ratio over the technically superior MiG-15…
People remember experiences not by averaging all moments but by sampling two: the peak (highest emotional intensity) and the end.
A power-law distribution is a statistical distribution where probability of size x is proportional to x^(−α): large events are rare but far more probable than a…
One party (the principal) delegates to another (the agent) whose interests differ and whose actions can't be fully observed — producing agency cost: monitoring spend +…