I’ve tried OKRs long enough now that they are not resulting in a positive outcome from a quarterly perspective. The problem lies in a couple things:
- Changing OKRs from quarter to quarter seems too fast and doesn’t result in actual time to develop long-term strategy.
- Once a strategy has been developed for an OKR it is hard to not drop it with a new OKR the next quarter.
I may be doing this wrong, but the whiplash of constantly chasing new OKRs doesn’t create consistent throughput. New OKRs can create misalignment since changing directions, mid-route is fighting the inertia of the previous OKR.
So I’ve changed tactics. I am no longer doing OKRs. Instead, there is a singular company-wide mission with key metrics that don’t change from quarter to quarter. This mission will be the North Star for the foreseeable future.
Hopefully this results in some long term benefits:
- Not wasting a lot of time planning on what new OKRs to make up. From a quarter to quarter standpoint opsZero doesn’t change that much. Companies want much the same thing they want last quarter. My goal is to make it more efficient to deliver it.
- Build metrics dashboards to understand key results somewhat in realtime.
- Focus on tasks that continuously benefit the mission’s key results as opposed to changing tactics quarter to quarter.
Getting rid of OKRs does not negate the need to plan. Instead, planning will be in shorter iteration cycles of six weeks. Planning should be to figure out the high leverage tasks, scaffold new initiatives, and basically treat building as continuously fixing bugs.
The optimal time for an iteration cycle seems to be 6 weeks. Four weeks is long enough to get major tasks done without having to stop for planning, two weeks are spent down-cycle, relaxing, and coming up with new initiatives. After 6 years of running opsZero, going full steam all the time is no longer a real objective. Taking breaks more frequently, has been a godsend.