OKRs (Objectives and Key Results)
A quarterly goal-setting framework that aligns organisations around measurable outcomes. Developed at Intel by Andy Grove, popularised at Google by John Doerr, and now widely used across technology companies.
Structure
An OKR consists of:
- Objective: An inspiring, motivating statement of what the team is driving towards in the quarter. Should make people want to get out of bed; neither too fluffy (“delight our customers”) nor too granular (“ship feature X”).
- Key Results (typically three): Measurable outcomes that answer “how do we know we succeeded?” — not tasks or milestones. Triangulation heuristic: one hardcore metric, one quality/retention signal, one revenue or growth signal.
The derivation question for key results is “how do we know?” — applied to the objective, it generates a set of measurable signals. A 10-minute brainstorm of every possible metric before selection often surfaces unexpected but insightful choices.
The full hierarchy
Mission (5-year aspiration — what are we doing and why?)
→ Strategy (strongly held hypothesis about how to win; resolves business model, channel, segment)
→ Annual themes (rough quarterly sequencing)
→ Quarterly OKRs (objective + 3 key results)
→ Weekly commitment (3 P1 initiatives; Monday status email)
Strategy is the often-missing layer. Without it, OKR objectives float free with no anchor — teams cannot use them to make independent decisions.
Cadence
- Monday: commit — each person/team sends a short status: confidence level on each KR (rough %, not precise), 3 P1 initiatives for the week, blockers
- Friday: celebrate — “What was the most awesome thing that happened this week?” ritual; builds psychological belonging
- Weekly status email: distributed to whole company if possible; creates transparency and fast connection between people who should talk
The OKR review meeting should take ~10 minutes once the process is internalised. If it is boring or takes hours, the team is reviewing tasks, not outcomes — something is broken deeper (unclear strategy, insufficient psychological safety, or a leader who won’t let go).
OKRs as vitamin, not medicine
Christina Wodtke’s framing: OKRs amplify healthy organisations; they reveal and accelerate dysfunction in broken ones. Prerequisites:
- Empowered teams
- Psychological safety (teams must be able to say “this isn’t working” without fear)
- A real strategy (not just a vague mission)
Common failure modes
| Failure | Root cause |
|---|---|
| Key results are tasks or milestones | Teams don’t understand outcome vs output distinction |
| OKRs too fluffy or too boring | Objectives not tied to a meaningful strategy |
| Boring review meetings | Leaders reviewing IC tasks, not strategy health |
| Company-wide rollout fails | Implemented too fast without understanding; should pilot with best team first |
| Approval process kills cadence | Replace hierarchical sign-off with 3-peer 24-hour review |
Grading and learning
~70% success rate on key results is healthy; 100% means targets were too easy. Retrospective matters more than precise scoring — focus on why you didn’t hit and what to change. The quarterly grade → retrospective → next OKR cycle is the often-overlooked mechanism that builds cumulative strategic knowledge in the organisation.
Rollout sequencing
- Pilot with the best multidisciplinary team (they will figure out how to adapt OKRs to your culture)
- Check in after one quarter
- Expand to two more teams; refine the template
- Bring to management last
Never use OKRs to fix a broken team; it will make them perform worse and hate OKRs.
Sources
- Christina Wodtke on OKRs, Radical Focus, and Why They Go Wrong — primary source; full treatment of structure, cadence, failure modes, rollout
- Chandra Janakiraman on An Operator's Guide to Product Strategy — the strategy layer OKRs depend on; Chandra’s small S process is one method for producing that hypothesis
- Bill Carr on Working Backwards, Input Metrics, and Amazon's PR/FAQ Process — Amazon’s input/output distinction maps onto Wodtke’s tasks-vs-outcomes distinction