02 / Methodology
Seven quality criteria, six anti-pattern checks. Each criterion targets one specific failure mode. The scoring is automatic; the logic behind each score is visible, so you can argue with it, teach from it, and improve against it.
The rubric is opinionated. It will flag some genuinely decent OKRs as incomplete because a baseline is missing or an alignment reference is not stated. That is a feature. The goal is not perfect OKRs by some abstract standard. It is OKRs that actually change outcomes rather than producing well-formatted planning theatre.
Try it now →Does the Objective name a specific customer and a specific scope?
Vague beneficiaries produce vague KRs downstream. If the Objective does not say who benefits or from what, the team cannot prioritise between the many ways to reach the stated direction.
Is there an explicit date or quarter?
An Objective without a timebox cannot be tracked. Teams defer the hard conversation about whether they are on track because there is no date to be on track against.
Is the Objective problem-framed, with no solution prescribed in the text?
A team that writes solution-first objectives has usually skipped the problem definition step. If the solution changes mid-quarter, the Objective becomes false. Problem-framed Objectives survive pivots.
Does the Key Result follow the structure "who does what by how much"?
Output-verbs (launch, migrate, deliver, create, build, implement) score 0. A metric with a vague actor scores 1. The full "who + does what + by how much" structure scores 2. This criterion applies per Key Result.
Does the KR include both a baseline and a target?
One present, one missing scores 1. Neither scores 0. Both, plus an implied or named data source, scores 2. If the baseline is unknown, the correct OKR is to instrument the metric first, not to improve it.
Does the OKR set reference its parent objective or the strategy it contributes to?
Alignment is not just governance overhead; it is the mechanism that connects team effort to organisational outcomes. The work may be well-intentioned and still be optimising the wrong thing.
Are there placeholders in the OKR set?
Anything marked X%, TBD, (owner), (tbc), or "numbers tbd" scores 0. A placeholder is a deferred decision. Submitting an OKR with placeholders is submitting a draft as a commitment.
A KR that describes work your team does rather than a change that happens in the world. The verb is the tell: migrate, launch, deliver, build, implement.
A KR so high-level and lagging that no single team can control it. A team that writes this kind of KR cannot tell at week 6 whether they are contributing or bystanders.
A plausible-sounding number that does not connect to a specific actor or behaviour. Vanity metrics are easy to move without moving the thing that matters. The test: can you imagine a scenario where this metric goes up and the business gets worse?
A KR with unknown numbers committed as though they were known. If the baseline is unknown, the KR is a wish. Instrument the metric first, then revisit the improvement goal next cycle with real numbers.
A pass/fail milestone that tells you whether something happened, not whether it worked. Usually an Output-as-KR in disguise. Ask what the milestone was supposed to change, then measure that.
Three or more KRs that are really one project plan. These are inputs, not results. A set with seven KRs where two do the heavy lifting and five are there for coverage is a set with five hidden tasks.
For every KR, ask three questions before committing. Any "no" means the KR needs rewriting.
The test surfaces the gap between activity and outcome. Most OKR problems are visible the moment you ask these three questions. Teams that skip the test usually discover the gap in the retrospective, which is too late to act.
Each of the 7 criteria scores 0, 1, or 2. The KR-level criteria (Outcome Form and Measurability) apply per Key Result, so a set with three KRs has more KR-level points in play than a set with one. The total raw score is normalised to a 0-100 percentage.
| Score range | Tier | What it means |
|---|---|---|
| 0-33 | Critical issues | Core structural failures. The OKR cannot be tracked or evaluated as written. |
| 34-55 | Weak | Some criteria pass, but the set has gaps that will cause problems mid-quarter. |
| 56-77 | Strong | Solid foundation. A few criteria need sharpening before commitment. |
| 78-100 | Excellent | All criteria met or nearly met. This is a committable OKR. |
The tiers are diagnostic signals, not grades. A score of 42 means specific criteria are dragging the set down. The per-criterion breakdown shows exactly which ones and why.
Try it now →