OKRs
Two jobs in one tool: write a set of OKRs from what the quarter is actually for, or paste the OKRs you already have and get a score out of 100 with a specific fix for every line that loses points. The score is counted in code from what each key result says, not judged by a model in prose.
- Asks for
- For writing: what the quarter is for, and what you would see if it went well
- For scoring: the OKRs you have, pasted as they are written
- Your current numbers, where you know them, so a target can have a baseline
- Gives back
- Objectives worth wanting, with key results that can be measured and can fail
- A score out of 100, counted from what each line says
- A named fix for every line that lost points, rewritten
- Which key results are tasks in disguise, and what the outcome behind each one is
- Plan
- Free
- Use next
- Vision board, Roadmap, Prioritisation, One-pager
What it gives you
Objective — Onboarding stops losing people in the first week. Score 68/100 KR1 "Ship the new onboarding flow" −12 task, not outcome → "Day-7 retention of new signups from 31% to 45%" KR2 "Improve activation significantly" −9 no number → "Share of signups reaching first export from 22% to 40%" KR3 "Day-7 retention from 31% to 45%" ok baseline, target, fails
How it works
Almost every bad OKR set fails on the same thing: the key results are a list of work rather than a list of outcomes. "Ship the new onboarding" is a task that is either done or not; it tells you nothing about whether onboarding got better, and a quarter of those turns OKRs into a project plan with worse formatting.
Writing mode starts from what the quarter is for and what you would see if it went well, and produces objectives that are worth wanting with key results that are measurable, have a baseline and a target, and could in principle come out badly. A key result you cannot fail is not a key result.
Scoring mode is the one people come back for. Paste your existing set and each key result is checked for the things that actually go wrong — no number, no baseline, a target nobody can measure, a task in disguise, an objective that is three objectives. The deductions are arithmetic over those checks, so the score is reproducible and the fix is attached to the line that lost the point rather than offered as general advice.
Who it is for
A PM writing a quarter’s OKRs on a deadline, or one who has inherited a set from somebody else and suspects half of them are tasks.
Where it stops
It judges whether an OKR is well formed, not whether it is the right thing to want: a perfect 100 is entirely compatible with a quarter spent on the wrong objective. It also cannot invent your baseline — give it a number you do not have and the target it writes will be as made up as the number.
Questions people ask
What is the difference between an objective and a key result?
The objective is what you want to be true and it is allowed to be qualitative — onboarding stops losing people in the first week. The key results are how you would know, and they have to be numbers with a starting point and a target, so that at the end of the quarter nobody has to argue about whether it happened. The usual mistake is writing key results that are deliverables, which can be completed without the objective moving at all.
How is the score out of 100 worked out?
By running each key result through a fixed set of checks — is there a number, is there a baseline, is it measurable with data you actually have, is it an outcome rather than a task, is the objective single — and deducting for each failure in code. That makes the score reproducible: the same set scores the same twice, and every deduction points at the line that caused it with a rewritten version attached.
Can I score OKRs that were not written in PikoPM?
Yes, and that is the more common use. Paste them in whatever shape they are in — a doc, a spreadsheet column, a list from your OKR tool — and each line is read as written rather than as intended. Scoring an inherited set is usually how a team finds out that four of its six key results are project milestones, which is a cheaper discovery in week one than in week eleven.
Run it on your own work
OKRs needs an account, because it writes against what you give it and remembers what you have told it before. It runs on the Free plan. Or see what the plans cost first.