Skip to main content
OKRs were invented at Intel, popularised at Google, and are now used (and misused) everywhere. The idea is simple: set an ambitious Objective - a qualitative direction - and pair it with a small number of Key Results - quantitative measures that tell you whether you got there 🎯 John Doerr’s Measure What Matters is the canonical read if you want the full framework and the origin story.

​
The structure

Objective - aspirational, qualitative, time-bound. “Become the go-to tool for freelance designers” or “Make onboarding feel effortless for new users.” Should be ambitious enough to stretch, clear enough to be unambiguous. Key Results - measurable outcomes that signal the objective was achieved. Three to five per objective is typical. Not tasks or outputs - actual results. “Increase 30-day retention from 42% to 60%” is a key result. “Ship the new onboarding flow” is not. The distinction matters. Output-based key results let teams hit their OKRs without moving the needle that actually matters 💡

​
What OKRs are good for

  • Focusing a team on a small number of things that matter
  • Creating alignment between team-level and company-level goals
  • Making trade-off decisions easier - “does this help our OKRs?” is a useful filter
  • Separating aspirational goals (stretch OKRs) from operational commitments (business-as-usual metrics)

​
Where they go wrong

OKRs fail when they become a compliance exercise. Teams write them because they have to, track them in a spreadsheet nobody reads, and score them at the end of the quarter without changing anything. They also fail when every team has too many. If you have six objectives and twenty key results, you don’t have focus - you have a prioritised to-do list with extra steps 😅 The other failure mode: key results that measure output instead of outcome. Christina Wodtke’s Radical Focus is excellent on this - she makes the case that one well-chosen OKR per quarter, taken seriously, beats a stack of them treated as box-ticking.

​
OKRs and product

For product teams, OKRs work well when they’re set at the outcome level and the team has genuine autonomy to discover how to achieve them. That’s the model Marty Cagan describes in Empowered - the company sets the what, the team figures out the how. Lesson learned: the best OKR conversation I’ve been in lasted two hours and ended with the team agreeing to drop three of their five objectives. That discomfort was the whole point 🙌