Implementing OKRs
February 8, 2026
How to implement OKRs correctly: writing measurable key results, designing quarterly cycles, weekly check-ins, and avoiding the most common OKR mistakes.
Objective vs Key Result: The Difference
An Objective is qualitative and directional: "Build a product that our customers love enough to recommend." A Key Result is a specific, measurable outcome with a deadline: "Increase NPS from 32 to 50 by the end of Q2" or "Reduce monthly churn from 5% to 2.8% by June 30." The Objective answers "where are we going?" The Key Result answers "how will we know we arrived?" Neither is useful without the other — an Objective without Key Results is a vision statement; Key Results without an Objective are a task list.
Andy Grove developed OKRs at Intel in the 1970s as a lightweight goal-management system for fast-moving engineering teams. John Doerr brought them to Google in 1999 when the company had five employees, and Google has used them at every scale since. The system's durability comes from a deliberately simple structure: one to five Objectives per quarter, two to four Key Results per Objective. More than that, and the system stops being a focus mechanism and becomes an administrative exercise.
Quarterly Cycle Design
A quarterly OKR cycle has four phases, each taking roughly one week: draft (leadership proposes company-level OKRs), cascade (teams propose their own OKRs aligned to company OKRs), align (cross-functional review to catch conflicts and gaps), and lock (OKRs are finalised and published). The cascade step is where most early-stage teams fail — they skip it and produce OKRs that are disconnected from each other, so the sales team is optimising for new accounts while the product team is optimising for retention.
Week one of each quarter should include an explicit retrospective on the previous quarter's OKRs before new ones are written. For each Key Result, ask three questions: What was the score? What drove the result? What would we do differently? This prevents the pattern of copying last quarter's OKRs verbatim with new numbers — which is the most common sign that a team is going through OKR motions rather than using them to learn.
How to Write Key Results
The most common mistake is writing tasks as Key Results. "Update the landing page" is a task — it describes an activity, not an outcome, and you can complete it without moving the business forward at all. The Key Result version is "Increase landing page conversion rate from 2.1% to 3.5% by end of Q3." The task may be part of achieving the Key Result, but the Key Result itself describes a change in the world, not a change in your to-do list.
Good Key Results are scoreable on a 0.0 to 1.0 scale, where 0.7 is considered success. The 0.7 target is not a consolation prize — it's the design intent. John Doerr's framing is that a Key Result scored at 1.0 was probably not ambitious enough; if you always hit 1.0, you're sandbagging. A Key Result that reliably scores 0.6–0.8 represents meaningful progress on a genuinely challenging goal. If a Key Result scores 0.0, the team should ask whether the goal was relevant, whether it was blocked by something outside their control, or whether it was simply unrealistic.
Weekly Check-in Structure
The weekly OKR check-in is not a status report — it's a confidence assessment. For each Key Result, every owner assigns a confidence rating from 1 to 10: how confident are you that this KR will hit the target by end of quarter? A rating that moves from 7 to 4 between weeks is an early warning signal that something has changed, and it gives the team three to five weeks to respond rather than discovering the problem at the end of the quarter.
Two additional elements make the check-in useful: a named blocker and a stated change. The named blocker is whatever is most constraining progress right now — not a vague "we've been busy" but a specific "the integration with Stripe webhooks is broken and we're waiting on their support team." The stated change is what the team will do differently this week as a result of this information. Check-ins that lack both of these elements tend to turn into reporting rituals rather than decision-making forums. Keep the whole check-in under 20 minutes; anything longer signals that the team needs a separate problem-solving session, not a longer check-in.
Frequently Asked Questions
How many OKRs should a ten-person startup have? Two to three company-level Objectives with two to four Key Results each is the right scale for a team of ten. Functional teams — product, sales, engineering — can each have one to two Objectives directly aligned to company-level OKRs. More than 12 Key Results across the company creates too many competing priorities for a small team to track meaningfully.
Should OKRs be connected to performance reviews? Most practitioners recommend keeping OKRs separate from individual performance evaluations. When Key Results are tied to compensation or promotion, people set conservative targets to guarantee a high score. The system works best when it's a planning and alignment tool rather than a performance measurement tool. Use separate criteria for individual performance reviews.
What is the difference between OKRs and KPIs? KPIs are ongoing health metrics — churn rate, monthly revenue, support response time — that you track continuously regardless of what quarter it is. OKRs are temporary goal-setting vehicles that expire at the end of the quarter. A KPI might appear inside an OKR as a Key Result if improving it is a current priority, but most KPIs are monitored without a formal OKR attached to them.
What happens when priorities change mid-quarter? Priorities change, and OKRs should respond. If a Key Result becomes irrelevant due to a market shift or product decision, mark it as cancelled with an explanation rather than quietly ignoring it. Forcing a team to continue working toward an objective that leadership has already abandoned is one of the fastest ways to kill trust in the planning process.
Can OKRs work for a two-person startup? Yes, though the process should be proportionally lighter. A two-person company needs the focus that OKRs provide even more than a large company. Write three Key Results on a shared document, review them every Monday morning for ten minutes, and adjust confidence scores. The quarterly retrospective can be a 30-minute coffee meeting. The ritual matters; the formality doesn't.
