Read on Twitter
@niallm's thread is as thoughtful, judicious, and balanced as I'd expect. Allow me to be a bit more negative about Objectives and Key Results. 
https://twitter.com/niallm/status/1355988157108465664
https://twitter.com/niallm/status/1355988157108465664
First, the planning horizon.
At $megacorp and anywhere that follows that playbook, OKRs are a quarterly process.
"Be careful what you measure": if your planning cadence is quarterly, there's a strong chance it'll become your delivery cadence too. That is too slow.
At $megacorp and anywhere that follows that playbook, OKRs are a quarterly process.
"Be careful what you measure": if your planning cadence is quarterly, there's a strong chance it'll become your delivery cadence too. That is too slow.
I love having a plan. Honestly, I enjoy the chaos as we watch it crash into reality, particularly in teams that are constrained by operational concerns (read: basically everyone with real customers).
Plans go out the window. That's fine, they're still useful.
Plans go out the window. That's fine, they're still useful.
But a quarter - for most teams and orgs I've worked with, including at $megacorp - is too big a chunk to let us respond well to inevitable change.
So OKRs become a big, expensive activity that swallow a bunch of team and management bandwidth without allowing room to react.
So OKRs become a big, expensive activity that swallow a bunch of team and management bandwidth without allowing room to react.
Second, "KRs considered harmful" (with caveats).
Making good KPIs is a known hard problem, right? As an engineering example, consider the difficulty of producing meaningful SLOs.
Coming up with a brand new set of good KPIs each quarter is high overhead and prone to failure.
Making good KPIs is a known hard problem, right? As an engineering example, consider the difficulty of producing meaningful SLOs.
Coming up with a brand new set of good KPIs each quarter is high overhead and prone to failure.
"KR 1: produce the metrics we're going to use to track this objective." 
It can feel really hard to tie quantitative measurement of KRs - often proxy metrics - to the real outcomes we're trying to drive.
It can feel really hard to tie quantitative measurement of KRs - often proxy metrics - to the real outcomes we're trying to drive.
Finally though, it's not all gloom.
Objectives are a good tool for driving alignment. They can build a coherent narrative around strategy and tactics, from execs through to delivery teams.
I like to de-emphasize the measurement part though.
Objectives are a good tool for driving alignment. They can build a coherent narrative around strategy and tactics, from execs through to delivery teams.
I like to de-emphasize the measurement part though.
Where we do have KRs, ideally they're tied to longer-term KPIs that we can iterate on and come to understand better over time, on the order of months or years.
SLOs are a good example. Or from a team perspective, the DORA metrics. Or "north star" product / business metrics.
SLOs are a good example. Or from a team perspective, the DORA metrics. Or "north star" product / business metrics.
As for planning horizons: I am strongly attracted to ~6 week "bets" per https://basecamp.com/shapeup or https://www.intercom.com/blog/6-week-cycle-for-product-teams/
Where I've worked like this, it's felt lighter-weight and more agile in the best senses.
Where I've worked like this, it's felt lighter-weight and more agile in the best senses.
All that said, whatever your particular team, org and/or company are doing to plan their time: get involved, and try to make it work as well as possible.
Channel your reservations into creativity. Use and improve the system. Try experiments. Share them.
Planning is good!
Channel your reservations into creativity. Use and improve the system. Try experiments. Share them.
Planning is good!
