OKRs (Objectives and Key Results) are one of the most widely adopted goal-setting frameworks in modern organizations — and also one of the most frequently misimplemented. Teams that do OKRs well use them to focus collective energy on the highest-leverage outcomes, make progress visible, and create a shared language for prioritization. Teams that do OKRs poorly end up with a spreadsheet that nobody looks at and a process that feels like bureaucracy for its own sake.

This guide walks through the full setup process — from understanding what makes a good OKR to running an effective end-of-quarter retrospective. It is designed for team leads and managers who are implementing OKRs for the first time, though the same principles apply to teams that have tried OKRs before and not gotten the results they expected.

Step 1: Understand the OKR formula

Before you write a single OKR, make sure the people setting them understand the structure and the underlying philosophy. An OKR consists of two elements:

Objective. A qualitative, inspiring statement of the direction you want to move. It should be memorable, motivating, and clearly connected to what matters. "Become the most trusted vendor in our category" is an objective. "Increase revenue by 20%" is not — that is a key result.
Key Results. Two to five specific, measurable outcomes that define what "achieving the objective" means in quantifiable terms. They should be outcome-oriented (changes in the world) rather than output-oriented (things you did). "User activation rate increases from 42% to 60%" is a key result. "Launch three new features" is not.

The relationship between objectives and key results is where most teams stumble. A useful test: if you achieved all of your key results but the objective still does not feel done, you have the wrong key results. If you cannot draw a direct logical line from each key result to the objective, they are not aligned.

OKRs are also explicitly designed to be ambitious. A score of 0.6–0.7 out of 1.0 at the end of the quarter is considered healthy at companies like Google that pioneered this approach. Consistent 1.0 scores mean your goals were not stretching the team. Building psychological safety around ambitious goals that partially miss is essential — otherwise teams will sandbag.

Step 2: Set company-level OKRs first

OKRs work best when they cascade from company to department to team. But cascading only works if the upper layers are visible first. If teams write their OKRs before company-level OKRs are published, they will optimize for what they think the company cares about — which may be quite different from what leadership actually prioritizes this quarter.

Company-level OKRs should be set and published by leadership at least one week before teams begin writing theirs. Leadership should also share the strategic reasoning behind each objective — not just the goals themselves, but the context that makes them the highest-priority outcomes for this period.

This context matters because it allows teams to make good tradeoffs. When a team understands why a company objective is the priority, they can make better decisions about which team-level work directly serves it versus which work should be deprioritized.

Step 3: Write team-level objectives

Team objectives should directly support one or more company-level objectives. Each team objective should be traceable back to its company-level parent — if you cannot draw that line, the objective probably should not exist this quarter.

Practical guidelines for team objectives:

Aim for 3–5 objectives per team per quarter. More than that dilutes focus to the point where nothing is actually prioritized.
Objectives should reflect choices — not just ambitions. If everything is a priority, nothing is. The best OKR sessions are the ones where teams have honest conversations about what they are not going to focus on.
Involve the team in writing the objectives, not just the manager. Bottom-up input surfaces the highest-leverage opportunities that leadership may not see, and ownership of the goals is higher when people helped write them.

If your team is also tracking Individual Development Plans, consider writing 1–2 development-focused key results for each team member alongside the business-focused OKRs. This integrates personal growth into the same tracking framework as business outcomes.

Step 4: Define measurable key results

This is where most OKR implementations go wrong. Key results that are actually outputs (things you will ship or do) rather than outcomes (changes in the world) undermine the entire framework. They create a false sense of progress — you can complete every action item and still not achieve the objective.

Test each proposed key result against the following criteria:

Is it outcome-based? "User activation rate from 42% to 60%" — yes. "Ship the new onboarding flow" — no.
Is it independently measurable? Could someone other than the author verify whether it was achieved, without interpretation or judgment?
Does it have a clear baseline and target? "Increase NPS" is not a key result. "Increase NPS from 32 to 45 by Q3 close" is.
Is it time-bound? Key results should have a clear end date, typically aligned with the end of the quarter.

Also consider whether achieving all of the key results would naturally constitute success on the objective. If there is a gap — a dimension of the objective that none of the key results would address — add another key result or revise the existing ones.

Step 5: Assign owners and set baselines

Every key result needs a single owner. Not a team, not a committee — one person who is accountable for making it happen, even if the work is distributed. Shared accountability tends to produce the same result as no accountability when things get difficult.

Before the quarter begins, owners should:

Confirm and document the baseline metric. You cannot measure improvement without knowing where you started.
Identify the primary data source and confirm they have access to it. Discovering mid-quarter that the data you need is not tracked is a common and avoidable problem.
Understand what they can actually influence. If the key result depends primarily on work done by another team, the owner needs a plan for coordination — or the key result needs to be reframed.

Step 6: Check in weekly or bi-weekly

OKRs that are reviewed only at the end of the quarter are not OKRs — they are a grading system. The value of the framework comes from using the goals to guide decisions and prioritization throughout the quarter, which requires regular check-ins.

A lightweight weekly check-in format that works well:

Each key result owner updates their confidence score (0.0–1.0) and notes any change since the last check-in.
Anything below 0.5 confidence triggers a brief discussion: what is blocking progress, and what adjustments are needed?
The team identifies any cross-team dependencies that need attention.

This does not need to be a long meeting. Many teams do OKR check-ins async in a shared doc or tool, with a brief synchronous discussion only when a goal is significantly off track. The important thing is that goals are reviewed and updated frequently enough to remain decision-relevant — not just reported on at the end. This connects naturally to the 1:1 cadence described in the How to Run Effective 1:1 Meetings guide.

Step 7: Score and reflect at end of quarter

At the end of each quarter, key result owners score their results on a 0.0–1.0 scale. A score of 0.0 means no progress was made. A score of 1.0 means the full target was hit. Most key results should land between 0.5 and 0.9 for the OKR system to be working as designed.

After scoring, run a brief retrospective for each objective:

What made progress happen where it did?
What blocked progress where we underperformed?
Were the key results the right ones — did hitting them actually advance the objective?
Were the targets the right level of ambition — too easy, too hard, or about right?

Feed the learnings from this retrospective directly into Q+1 OKR writing. The teams that consistently improve their OKR quality are the ones that treat end-of-quarter retrospectives as a design input for the next cycle, not just a backward-looking accounting exercise.

OKR scores from the cycle also become useful inputs for performance reviews — not as the primary driver of ratings, but as one data source among several that managers can reference when documenting performance narratives.