Skills on AI 459 skills

Objectives and Key Results

Use when asked to write OKRs (Objectives and Key Results) — an inspirational, team-actionable objective paired with measurable key results — grounded in joelparkerhenderson/objectives-and-key-results, including known real-world pros/cons, as distinct from KPIs alone (see key-performance-indicators).

OKR (Objectives and Key Results) is a lightweight goal-setting method: the Objective states what you want to achieve; the Key Results state how you'll know you're making progress. A simple template: "I will (Objective) as measured by (this set of Key Results)."

History

Invented at Intel and championed by CEO Andy Grove (in his book High Output Management), later popularized further by venture capitalist John Doerr (Measure What Matters). Widely used at Intel, Google, Microsoft, Twitter, and Uber, among others.

What makes a good objective

  • Inspirational — provides a real sense of meaning and progress; skip anything as unambitious as "a small percentage gain."
  • Measurable by the team independently — no excuse like "that other team didn't measure it."
  • Actionable by the team independently — no excuse like "marketing didn't market it"; the objective has to be genuinely the team's own.
  • Relatable — plain language the team actually uses, not abstract jargon.
  • Timely — a clear sprint toward the goal, doable within roughly a month or a quarter; if it takes a year, it's probably strategy or mission, not an OKR-scale objective.

What makes a good key result

Ask: "how would we know if we met our objective?" Typically 1–3 key results per objective, always quantified — a key result is not a general or subjective planned action. Metrics commonly draw from growth, engagement, revenue, performance, or quality. Andy Grove's own bar: "to be useful a Key Result must contain very specific wording and dates, so that when the deadline arrives, there is no room for ambiguity."

Ground rules

  • OKRs are transparent by default — visible to the whole organization (rare exceptions for maximum-security/secrecy work). Visibility helps teams understand what others are doing and supports cross-team alignment.
  • OKRs nest: if a subordinate's objectives are met, the supervisor's objectives should be met as well — the hierarchy should be a genuine cause-and-effect chain, not independently-invented goals at each level.
  • OKRs are meant to pace a person (Grove's own framing: "a stopwatch in their own hand"), not to serve as the legal basis for a performance review on their own — mechanically grading someone against a rigid OKR ignores legitimate emerging opportunities the OKR didn't anticipate.
  • Aim for roughly 70% achievement as a good outcome on a genuinely stretching OKR — 100% attainment routinely suggests the objective wasn't ambitious enough.

How OKRs differ from strategy, KPIs, and task lists

  • Strategy vs. objectives — a strategy explains how a target will be reached; an objective is only the target itself. An organization can state an objective with no real strategy behind it, which leaves a hundred possible paths and no actual direction — see Strategic Balanced Scorecard for a framework that supplies the strategic layer OKRs alone don't.
  • Key Performance Indicators — KPIs are often used alongside OKRs as the next layer down: an objective's key result might be hard to hit directly, so KPIs track the constituent numbers that drive it, letting a team find the right combination empirically.
  • Task lists / work boards — OKRs deliberately capture only the top few goals and progress points; task lists capture a full enumeration of work items. OKRs answer "what matters most"; task lists answer "what needs doing."

Real-world tradeoffs (both sides, honestly)

OKRs have a large, genuinely mixed body of practitioner experience. Reported benefits: forced alignment conversations that surface real cross-team disconnects, visible priorities that help new hires orient quickly, and a shared vocabulary for saying no to work outside the current quarter's focus. Reported failure modes: OKRs imposed top-down with no bottom-up input become disconnected from day-to-day reality; tying compensation/performance review too tightly to OKR attainment incentivizes sandbagging or gaming; and OKRs treated as quarterly theater — written once, promptly ignored, backfilled with a plausible narrative at quarter-end — provide none of the intended benefit while still costing the time to write them. Whether OKRs work in a given organization depends heavily on genuine leadership commitment to the system, not on the OKR format itself being inherently good or bad.

Common pitfalls

  • Non-measurable key results — "improve X" without a number and timeframe fails Doerr's "as measured by" test and invites disputed, unfalsifiable claims of success later.
  • Objectives set with no strategic anchor — see the "strategy vs. objectives" distinction above; without it, OKRs measure activity against an arbitrary target rather than genuine strategic progress.
  • Tying OKR attainment directly to compensation/performance ratings — a well-documented way to turn OKRs into a gaming exercise rather than an honest stretch-goal tool.
  • Setting OKRs and never revisiting them until quarter-end — OKRs are meant to be looked at regularly (many teams: weekly), not written once and rediscovered at review time.
  • Purely top-down or purely bottom-up OKR-setting — top-down alone disconnects from operational reality; bottom-up alone isn't real strategic planning; a cycle combining both (context down, proposed key results up, then reconciliation) is what most mature practitioners report actually working.

Learn more

View objectives-and-key-results/SKILL.md on GitHub