At the end of a quarter, teams that score by feeling end up in long debates. Was 95.5 percent on-time delivery a success or a failure against a target of 97? A shared formula ends the debate and leaves time for the more useful question: what did we learn?
One formula
The score of a key result is the progress made divided by the progress planned, limited to the range from 0.0 to 1.0:
Score = (end value minus start value) divided by (target value minus start value)
The formula works in both directions. On-time shipment from 91 to a target of 98 percent, end value 96.5: (96.5 minus 91) divided by (98 minus 91) = 0.79. Expedited freight per week from EUR 3,540 to a target of EUR 1,150, end value EUR 1,850: (1,850 minus 3,540) divided by (1,150 minus 3,540) = 0.71.
The objective score is the average of its key results.
The plant in Q1 2027
| Team | Key result | Start | Target | End | Score |
|---|---|---|---|---|---|
| Plant | On-time delivery, percent | 91 | 97 | 95.5 | 0.75 |
| Plant | Customer complaints per quarter | 14 | 6 | 7 | 0.88 |
| Plant | Lead time, days | 12 | 8 | 10 | 0.50 |
| Engineering | Changeover time line 3, minutes | 74 | 30 | 41 | 0.75 |
| Purchasing | Line stops from missing parts | 11 | 3 | 5 | 0.75 |
The plant objective scored 0.71. Logistics reached 0.78 and 0.74, line 3 0.80, engineering 0.75 and purchasing 0.88.
What a good score is
For aspirational key results, 0.6 to 0.7 is a good result. A team that scores 1.0 everywhere set its targets too low. A team that scores 0.3 everywhere either overestimated what is possible or did not have the capacity.
For committed key results, 1.0 is expected. A committed key result that is at risk is escalated during the quarter, not reported at the end. In the example, logistics committed to trained standard work for all six core processes before a customer audit, and reached it.
The retrospective
Two hours in week 13, four questions:
- What did we reach, and which initiatives really moved the numbers?
- What did not move, and why? Target, approach, capacity or dependency?
- What do we change in how we run OKR next quarter?
- What do we carry over on purpose, and what do we drop on purpose?
In the example, lead time moved least. Changeovers got faster, but batch sizes on line 3 were not reduced yet, so orders still waited. Batch size reduction became a key result for the next quarter.
Standard work in the warehouse scored 1.0. It was dropped from the OKR and returned to the health metrics: done and audited, now it must hold.
Never silently repeat
The worst outcome of a quarter is an unchanged OKR list for the next one. Every key result that was not reached is either carried over with a reason and possibly a new approach, or dropped with a reason. Both are fine. Repeating it by habit is not.