OKR
Also known as: Objectives and Key Results, OKRs
OKRs (Objectives and Key Results) are a goal-setting framework where each objective (a qualitative ambition) is paired with 2–5 key results (measurable outcomes that prove the objective was achieved). Originated by Andy Grove at Intel in the 1970s, popularized by Google in the 2000s.
OKRs separate "what we want to be true" (the objective) from "how we will know we got there" (the key results). The framework forces teams to be specific: an objective without measurable key results is a wish list. The cadence is typically quarterly — each quarter, the company sets 3–5 objectives, teams set objectives that ladder up, individuals set objectives that ladder up to the team. Scoring at quarter-end measures how much of each key result was achieved (typically on a 0.0–1.0 scale).
Anatomy of an OKR
A well-formed OKR has one verb-driven objective and 2–5 measurable key results. The objective should be ambitious enough that landing all five key results at 1.0 feels unlikely — Google's convention treats 0.6–0.7 as a healthy score, with 1.0 signaling the objective was set too softly.
- Objective: Make the first 90 days the best part of joining the company
- KR1: Raise the average new-hire onboarding survey score from 3.6 to 4.3 out of 5
- KR2: Cut the time from signed offer to first productive day from 15 to 8 working days
- KR3: Have 90% of new hires finish every onboarding task by day 30
- KR4: Bring first-year voluntary turnover down from 22% to 15%
OKRs vs KPIs
KPIs (Key Performance Indicators) track ongoing business health — metrics you watch every week regardless of what you're working on. OKRs target a specific shift over a specific window — what changes this quarter. A KPI lives forever; an OKR retires when its quarter ends. The two complement each other: KPIs tell you the steady state, OKRs tell you the bet.
Common failure modes
- Sandbagging — setting KRs you know you can hit 100%. Defeats the framework.
- Output as outcome — counting features shipped instead of outcomes achieved
- Cascading too rigidly — every IC has 4 OKRs that all roll up. Bureaucracy at scale.
- Linking to comp — OKRs become political, sandbagging accelerates
- Quarter-long cycles for fast-moving teams — startups often run 6-week cycles instead