4 ms·

My biggest issue with OKRs comes from those who practice what could be called "Aimless Key Results". Instead of thinking about an intuitive objective and then

My biggest issue with OKRs comes from those who practice what could be called "Aimless Key Results". Instead of thinking about an intuitive objective and then try to come up with some form of measurable Key Result that can help assess whether the org is getting closer to—or further from—the objective, some skip the whole Objective part altogether. They become so stressed with having something measurable that they end up with KRs based on how easy they are to measure, regardless if they actually correlate to a supposed objective. This is Goodhart's Law[1] on steroids. Healthy OKRs do exist but they have a few preconditions: - the individuals devising them must fully internalize which are the objectives that do matter for the org, among an ocean of "priorities". - the KRs must be carefully thought out and evaluated against their vulnerability to the law of unintended consequences. - the team(s) implementing the KRs must understand why theses KRs exist, their relationship to the Objective they serve, and how well (or not) they measure the objective. - the leaders must be quick to change the KRs if there's any doubt as to whether they are actually helping, or causing more damage. The "why" should remain relatively constant over time, the "how" can change. - KRs that have thresholds should always be ambitious and very hard to actually meet, as they stretch the organization's goals. - Tying perforance bonuses to them is dangerous because it immediately results in gamification: Damn the Objective, I want my bonus! OKRs can be incredibly tricky for middle management because the top leadership may have a decent objective to be pursued, and the line managers may have a good grasp on what success looks like. But the middle part is caught in between the 2, and this can create all sorts of fun artifacts of measurement, aka. dysfunctions. Especially if that middle management layer doesn't have a rock solid understanding of the actual objectives and are only obsessed with the measuring part. [1] When a measure becomes a target, it ceases to be a good measure. — https://en.wikipedia.org/wiki/Goodhart%27s_law#cite_note-Strathern1997-1 https://en.wikipedia.org/wiki/Goodhart%27s_law#cite_note-Str...

andrewingram 4y ago
I have a slightly different take. I think there's a tendency to attach too much importance to the key results; though I very much agree with your point about the law of unintended consequences, this is why the concept of anchoring metrics/KRs comes up a lot. To me, the Objective is what you want, whilst the KRs are just a set of heuristics that tell you if the work you're doing is taking you in that direction. But I see it as a classic "the map is not the territory" thing. Too much emphasis on the KRs (like, as you say, tying bonuses to them) can lead to gaming the system and other failure states.

andrewingram 4y ago

I have a slightly different take. I think there's a tendency to attach too much importance to the key results; though I very much agree with your point about the law of unintended consequences, this is why the concept of anchoring metrics/KRs comes up a lot. To me, the Objective is what you want, whilst the KRs are just a set of heuristics that tell you if the work you're doing is taking you in that direction. But I see it as a classic "the map is not the territory" thing. Too much emphasis on the KRs (like, as you say, tying bonuses to them) can lead to gaming the system and other failure states.