qa-aman Overview
Based on "Measure What Matters" by John Doerr, who brought OKRs from Intel to Google. Doerr's core distinction: Objectives are the "what" (inspirational, qualitative), Key Results are the "how we know" (measurable, binary at quarter end). CFRs - Conversations, Feedback, Recognition - are what make OKRs work in practice.
Two types of OKRs (Doerr):
- Committed OKRs - Must be achieved 100%. Missing them requires explanation.
- Aspirational OKRs - Stretch goals. 70% attainment is success. 100% means you aimed too low.
Workflow
Step 1: Identify the type
Ask: is this a committed OKR (must hit) or aspirational (stretch)? This changes how you write the Key Results.
Step 2: Write the Objective
- Inspirational and qualitative
- Memorable - a team should be able to recite it
- Time-bound to the quarter
- Answers: "What do we want to be true in 90 days?"
Doerr's test: "Does this objective make you uncomfortable in a good way?" If it's easy, it's not aspirational enough.
Step 3: Write 3-5 Key Results per Objective
Each KR must:
- Be measurable with a specific number
- Be outcome-based, not output-based ("30% reduction in churn" not "launch retention feature")
- Have a baseline and a target
- Be independently verifiable
Format: [Metric] from [baseline] to [target]
Step 4: Apply the Doerr CFR check
For each OKR set, identify:
- Conversation: What weekly check-in question will track this?
- Feedback: How will the team know mid-quarter if they're on track?
- Recognition: What does hitting this KR mean for the team?
OKRs without CFRs become forgotten spreadsheets.
Step 5: Check vertical alignment
Doerr's "cascading" principle: team OKRs should visibly connect to company OKRs. Flag any team OKR that doesn't trace to a company-level objective.
Step 6: Flag missing baselines
"You'll need a baseline for this KR. Who owns this metric and can pull the current number?"
Anti-Patterns
1. Output-based Key Results Bad: "Launch new onboarding flow by March 31" Good: "Increase onboarding completion rate from 45% to 70%" Shipping is activity. Outcomes are results.
2. Treating aspirational OKRs as committed Bad: Team marks 70% attainment as failure because "we didn't hit it." Good: For aspirational OKRs, 70% = success. Recalibrate at quarter end, don't punish stretch.
3. Too many Objectives Bad: 5-7 Objectives per team. Good: 1-3 Objectives max. Doerr: "A few well-chosen OKRs send a clear message about what we say no to."
4. No baseline Bad: "Improve conversion rate to 10%" Good: "Improve conversion rate from 6% to 10%"
Quality Checklist
- OKR type stated: committed or aspirational
- Objective is inspirational and qualitative (not a task)
- 3-5 Key Results per Objective
- Each KR has baseline and target
- KRs are outcome-based, not output-based
- CFR check done: conversation, feedback, recognition defined
- Team OKRs connect to company OKRs
- Missing baselines flagged with owners