3 ms·

Writing OKRs is useful. Strict grading of OKRs is worse than useless. During the year (or quarter), your picture of what's useful and feasible becomes more and

Writing OKRs is useful. Strict grading of OKRs is worse than useless. During the year (or quarter), your picture of what's useful and feasible becomes more and more precise. So your OKRs that you wrote at the beginning are worse than your current knowledge of the true goals. It's really bad to cling to your old misconceptions of what was important when you wrote those OKRs. Finally, the desire to quantitative-ize every OKR is beyond harmful. Does your organization really not care about anything qualitative? Well, it shows in your products. And no, you can't measure quality by surveying your users and averaging the results.

skeeter2020 5y ago
I think you can even back it up further; the real value is in defining the problem statement that the OKR addresses. The resulting OKR is just one of the many paths, and measuring results against it useful for indicating your efforts but not really progress. I agree with you on the quantitative mania. I'm not that clever but even I can frame almost any bias so it's "backed up by data". If "listen to the science" is today's new religion, our "evidence-based decision" is an opinion in disguise.
sokoloff 5y ago
The converse problem is a variant of the Texas Sharpshooter where, as you progress through the time period, you refine your goals to be closer to what is happening in the business, much like aligning the shadow of a batted ball to the ball. Eventually, those things inevitably converge and the updated goal is always achieved. According to the green OKR report anyway. You shouldn’t stick overly stubbornly to bad goals, but any updating of goals in the time period should be carefully considered as well.
rixed 5y ago
> Well, it shows in your products. If it does show in the product, shouldn't there be a way to quantify it? It seems to me that the problem with OKR is that it's hard to formalise goals, the same way it's hard to formalise requirements when designing a product. Does not imply it's not desirable to do so.
- lostdog 5y ago
  For example, Facebook is dying because they focus on engagement and ignore enjoyment. You can measure engagement, so it fits neatly into a key result, but enjoyment is hard to measure, so the entire company is blind to it.
  - rixed 5y ago
    Sure, no need to give me examples because I've seen plenty on my own. Still, I don't think we should abandon the idea of basing our decisions on data just yet. Bad metrics are efficiently worsening products and work culture, but the solution is better metrics, not going back to HiPPO.
  - pnt12 5y ago
    Completely agree. Measurable results go hand in hand with quick reports and pretty graphs, but they fail to capture so many things. I think this is replaced by vision, trust and transparency: if competent people (or at least competent leaders) share a vision, they will make it happen. The next iPod or iPhone won't happen due to an OKR.

skeeter2020 5y ago

I think you can even back it up further; the real value is in defining the problem statement that the OKR addresses. The resulting OKR is just one of the many paths, and measuring results against it useful for indicating your efforts but not really progress. I agree with you on the quantitative mania. I'm not that clever but even I can frame almost any bias so it's "backed up by data". If "listen to the science" is today's new religion, our "evidence-based decision" is an opinion in disguise.

sokoloff 5y ago

The converse problem is a variant of the Texas Sharpshooter where, as you progress through the time period, you refine your goals to be closer to what is happening in the business, much like aligning the shadow of a batted ball to the ball. Eventually, those things inevitably converge and the updated goal is always achieved. According to the green OKR report anyway. You shouldn’t stick overly stubbornly to bad goals, but any updating of goals in the time period should be carefully considered as well.

rixed 5y ago

> Well, it shows in your products. If it does show in the product, shouldn't there be a way to quantify it? It seems to me that the problem with OKR is that it's hard to formalise goals, the same way it's hard to formalise requirements when designing a product. Does not imply it's not desirable to do so.

lostdog 5y ago
For example, Facebook is dying because they focus on engagement and ignore enjoyment. You can measure engagement, so it fits neatly into a key result, but enjoyment is hard to measure, so the entire company is blind to it.
- rixed 5y ago
  Sure, no need to give me examples because I've seen plenty on my own. Still, I don't think we should abandon the idea of basing our decisions on data just yet. Bad metrics are efficiently worsening products and work culture, but the solution is better metrics, not going back to HiPPO.
- pnt12 5y ago
  Completely agree. Measurable results go hand in hand with quick reports and pretty graphs, but they fail to capture so many things. I think this is replaced by vision, trust and transparency: if competent people (or at least competent leaders) share a vision, they will make it happen. The next iPod or iPhone won't happen due to an OKR.