I think the dictionary definition of “malpractice” has references to improper, illegal, or negligent professional behaviour.
The way I see it is that we wish to elevate the practice of management to be a profession (consider the Australian Institute of Management or its NZ equivalent). It is in this context that the word malpractice has specific meaning.
Oops - somehow posted this outside the thread. Trying again.
"Management Malpractice" - I assume it's being a negligent manager or leader based on the definition you shared. So by extension we are examining is ANY use of OKRs are a form of negligence by managers. Fair?
There's no debate that some use of OKRs is reflective of negligent behavior - such as using OKRs for managing teams rather than supporting alignment and awareness across teams, linking OKRs to performance and other incentives, disempowering by managers setting the OKRs for their teams, disempowering by defining goals as outputs or actions on behalf of teams, disconnecting from learning by tracking progress through activities without considering what causal relationships may exist, working without a hypothesis...
But what Noah has put out there and you have asked me to consider (and I have, long before anyone asked me to) is are all applications of OKRs negligence. Albeit Noah qualified his view in his post identifying that link to incentives, cascades, managers setting the OKRs to name a few are the specific concerns he has with OKRs. But you wouldn't find any experienced practitioner recommend any of these practices. So those stripped away, where exactly is your discomfort.
He also suggested he'd instead use goal trees which are all recommended amongst the same set of experienced practitioners (the list of the people I am thinking of when I use this term was linked to from my post).
You mentioned no discomfort with team goals so what of the remaining aspects of OKRs provide the discomfort you have? Is it that they are public? Is it that cultures that use them may expect them used universally? These are all areas worthy of debate. I suspect there will be shifts in these areas true.
Another area worthy of debate is OKRs as targets. A lot of practical applications of OKRs in organisations soften their effect as targets because failure is more than acceptable its seen as learning and that is valued more than the result over a short period. OKRs are most widely used in SaaS businesses currently where competition is fierce and these companies are trying to prevail over the long term, not just juice metrics. But is it enough to not have the destructive effects of targets. It's part of my affection with PuMP - Stacey Barr has designed a performance measurement approach entirely around not setting targets.
Is there some discomfort with expressing outcomes separate with the method as per John Willis reference to Deming 'by what method'? This seems to be based on what you imagine happens when the organisations I've worked with use OKRs.
In my experience where teams are defining their goals they also have a rich view on what their hypotheses are and how they will test them. OKRs being for alignment only don't typically define how they approach the activities so they use any of the wide range of options for engaging and theorising and testing their theories. It starts with blame-free cultures and a culture of being comfortable using evidence for decisions.
You don't have to take my word for it, plenty of my colleagues have written about their experiences working this way. Hence its hard to reconcile all use of OKRs with "Management Malpractice".
My initial comment was "I’d like you to consider, for 1 minute, that all the forms you outlined above are some form of management malpractice." You rejected that with your MECE argument. OK, I understand you see nuance that makes the history of OKRs quite different in each iteration. I am afraid I don't. I have read many of the references you use but do not see a fundamental deviation from the MBO or MBR concept.
However, I'd like to back down from my "management malpractice" argument. That got me into trouble with neglicence, etc. You'll note I dont refer to this term except in references to your and Noah's articles.
I think I will change to arguing OKRs are just bad management.
I guess it is the separation of results from means that is my core point. In all my experience with OKRs (limited) and my reading (wider but still limited compared to you) I have never been able to reconcile OKRs as part of understanding variation. I am sure there are examples where they are used in this way but I am of the opinion that OKRs assume a stable system and effectively ignore that system in OKR setting.
Thanks for returning to this discussion - it felt there were some open threads.
Why would you use OKRs for managing teams? It's purpose is for alignment and to a lesser degree, to aide coordination between teams. Those issues mostly present themselves when there are a lot of teams. There can be drivers for using OKRs among a single team or startup too - maybe a compare and contrast of those two scenarios might make a good post? My experiences using OKRs with Seek Asia vs Weirdly were quite different. But most come to OKRs in environments where there are many teams.
If you don't have the challenges that OKRs are a decent fix for, don't use them. You can read some of my posts tagged 'alignment' for real work examples that led us to use OKRs and we then adjusted them in the ways I described in the Improving OKRs post and since then, writing about experiences using OKRs I found others globally who had arrived at a similar mix of practices.
We had over 20 teams at Seek Asia all working on complicated systems and it was not always obvious when there might be good opportunity to collaborate for mutual benefit. We had various ceremonies for sharing what we learned, what each team was doing, how they were approaching problems etc. but we also found opportunities to collaborate by merely sharing what intended to try and achieve. Maybe that better illustrates what I mean by alignment. If you are imagining this is how work is managed I can understand why you might be confused.
I had some other comments to make but lost them when the page refreshed so I will post this and come back in another comment tomorrow 😁
Hi Daniel, I’m not sure where the team aspect has come from. I don’t think I specifically mentioned OKRs for teams or team outcomes. I scrolled back up but didn’t see anything.
A few reasons - OKRs are often set both organisationally and by teams. You also mentioned bad management - managers manage both their area of responsibility and the people involved (manage is a loaded word).
I am usually going to respond by describing how OKRs have been used in the actual environments I've seen them used. To describe what was the stimuli at the time and what led to using OKRs. You said you had less experience with them so short of trying it yourself when a relevant problem comes up (and you shouldn't if it doesn't) you figure the next best thing is you asked someone who used it in multiple organisations and observed it used in others. If that's why you are asking I am happy to keep sharing.
I've mentioned organisations align these constellations of objectives that exist with teams and company wide through causal trees which map not just the current focus but all relevant theorised causal links (that's helps with theorising where interventions might be taken in a way that's accessible whether you are familiar with the technical systems involved or not). Its remarkable how many organisations I've observed arriving through simultaneous discovery this way of working. So much so I keep writing about it :)
Its also interesting that similar approaches have been described dating back many decades and such approaches feature in the work of Goldratt as one example. Even though there is prior art I say simulataneous discovery because I believe most of the orgs I am referencing arrived through necessity and experimentation rather than necessarily being aware similar approaches had been used before.
You also mentioned OKRs being set that were not considerate of the whole of the system so I think providing some visibility on how teams use them vs how you might imagine teams use them may be helpful.
Oh and one question of yours that i did not answer was the John Willis one. I think the separation of means (or method) from objectives, outcomes and key results is my biggest concern.
Thanks Martin - yes you've made variation of this statement "the separation of means (or method) from objectives, outcomes and key results is my biggest concern." numerous times dating all the way back to the first LinkedIn thread where we discussed this. I asked my question hoping you will elaborate. Your post helps in that it illuminates the theory well and does a little to describe how you think OKRs are applied.
People working with OKRs in the manner I describe focus most of their energy on the means - they don't use OKRs in a vacuum. They use all the tools we've mentioned (goal trees, XmR, architecture diagrams, team interaction diagrams, other models) to try and understand how the system is working and to collaborate on creating shared causal theories of what changes will yield the desired behaviour of the system. OKRs is another set of information that teams define to help share intents about adjustments to the systems they will experiment with and find connection with what may be desirable for uses of the software they are producing. And when multiple teams have interrelated, interdependent changes they often articulate the intent in a shared goal in addition to all the other useful artifacts for collaboration on a shared understanding.
They don't set them to be judged by management they set them aligning on what they desired to achieve as a team and for mutual discovery amongst other teams - helpful when there's a lot of teams. Will bad managers judge teams based on lack of achievement of goals? Without doubt. In a system which incentivises them that success is fairy dust in the form of manipulated numbers they will do that - so you focus on addressing the incentives that lead to that.
Can OKRs help reinforce that poor response? Yes - where there is low literacy of how causality and systems work then any use of numbers can be interpretted to be how they may have been used elsewhere. I've written a lot about how we approached learning and raising the understanding of these concepts across the organisation. Its one of the reasons they can appear deceptively simple but in practice can be hard to do well. Sometimes the relief from other issues and the rewards justify the investment.
If OKRs are being used to manage outcomes I can see how separation from means leads to an issue - e.g. if all leaders did was to set a goal and measures and then come back three months later they'd have very comfortable jobs and the company would fail with declining quality and overwhelming side effects. I've even seen clients I've been consulting with do this (I don't consult on OKRs - I do consult on strategy, how to draw connections between team interactions, system interactions, linkage with strategy etc.) and its clear abdication of responsibility. They do this with or without OKRs.
I can empathise that its tough to imagine without experiencing the conditions that led us to hire OKRs. That's why I am happy to answer any questions about the experiences using them in different contexts. Why we used OKRs at Seek differed to how we arrived at using them when they were used in a startup and how the implementation differed in each context. I've had a long career working at many companies and consulted and coached at even more and OKRs were not an answer to a question we had at many of them - either because the need wasn't there or something else was hired to solve a similar problem.
I hope that provides some specific examples that are relevant for you - I am afraid I feel like I am guessing because you've ended a few times with your statement about separating from means without translating that into what you think actually happens.
Thanks Daniel. I’ll focus my answers on your specific concerns. Apologies that I have not done so in the past. I’m also up for a call to discuss in person if that’s easier for you.
Maybe we should start with what is management malpractice? Situational, isn’t it?
I took it to mean more detrimental than not doing it :)
I think the dictionary definition of “malpractice” has references to improper, illegal, or negligent professional behaviour.
The way I see it is that we wish to elevate the practice of management to be a profession (consider the Australian Institute of Management or its NZ equivalent). It is in this context that the word malpractice has specific meaning.
So, now what is “management malpractice“?
Oops - somehow posted this outside the thread. Trying again.
"Management Malpractice" - I assume it's being a negligent manager or leader based on the definition you shared. So by extension we are examining is ANY use of OKRs are a form of negligence by managers. Fair?
There's no debate that some use of OKRs is reflective of negligent behavior - such as using OKRs for managing teams rather than supporting alignment and awareness across teams, linking OKRs to performance and other incentives, disempowering by managers setting the OKRs for their teams, disempowering by defining goals as outputs or actions on behalf of teams, disconnecting from learning by tracking progress through activities without considering what causal relationships may exist, working without a hypothesis...
But what Noah has put out there and you have asked me to consider (and I have, long before anyone asked me to) is are all applications of OKRs negligence. Albeit Noah qualified his view in his post identifying that link to incentives, cascades, managers setting the OKRs to name a few are the specific concerns he has with OKRs. But you wouldn't find any experienced practitioner recommend any of these practices. So those stripped away, where exactly is your discomfort.
He also suggested he'd instead use goal trees which are all recommended amongst the same set of experienced practitioners (the list of the people I am thinking of when I use this term was linked to from my post).
You mentioned no discomfort with team goals so what of the remaining aspects of OKRs provide the discomfort you have? Is it that they are public? Is it that cultures that use them may expect them used universally? These are all areas worthy of debate. I suspect there will be shifts in these areas true.
Another area worthy of debate is OKRs as targets. A lot of practical applications of OKRs in organisations soften their effect as targets because failure is more than acceptable its seen as learning and that is valued more than the result over a short period. OKRs are most widely used in SaaS businesses currently where competition is fierce and these companies are trying to prevail over the long term, not just juice metrics. But is it enough to not have the destructive effects of targets. It's part of my affection with PuMP - Stacey Barr has designed a performance measurement approach entirely around not setting targets.
Is there some discomfort with expressing outcomes separate with the method as per John Willis reference to Deming 'by what method'? This seems to be based on what you imagine happens when the organisations I've worked with use OKRs.
In my experience where teams are defining their goals they also have a rich view on what their hypotheses are and how they will test them. OKRs being for alignment only don't typically define how they approach the activities so they use any of the wide range of options for engaging and theorising and testing their theories. It starts with blame-free cultures and a culture of being comfortable using evidence for decisions.
You don't have to take my word for it, plenty of my colleagues have written about their experiences working this way. Hence its hard to reconcile all use of OKRs with "Management Malpractice".
Apologies for dropping the conversation. You wrote such a lot and I know it is a topic you feel strongly about. I just placed most of my arguments in this Medium post: https://medium.com/thinking-about-digital-transformation/by-what-means-management-without-objectives-5d1d10643b3c
My initial comment was "I’d like you to consider, for 1 minute, that all the forms you outlined above are some form of management malpractice." You rejected that with your MECE argument. OK, I understand you see nuance that makes the history of OKRs quite different in each iteration. I am afraid I don't. I have read many of the references you use but do not see a fundamental deviation from the MBO or MBR concept.
However, I'd like to back down from my "management malpractice" argument. That got me into trouble with neglicence, etc. You'll note I dont refer to this term except in references to your and Noah's articles.
I think I will change to arguing OKRs are just bad management.
I guess it is the separation of results from means that is my core point. In all my experience with OKRs (limited) and my reading (wider but still limited compared to you) I have never been able to reconcile OKRs as part of understanding variation. I am sure there are examples where they are used in this way but I am of the opinion that OKRs assume a stable system and effectively ignore that system in OKR setting.
Am I wrong?
Thanks for returning to this discussion - it felt there were some open threads.
Why would you use OKRs for managing teams? It's purpose is for alignment and to a lesser degree, to aide coordination between teams. Those issues mostly present themselves when there are a lot of teams. There can be drivers for using OKRs among a single team or startup too - maybe a compare and contrast of those two scenarios might make a good post? My experiences using OKRs with Seek Asia vs Weirdly were quite different. But most come to OKRs in environments where there are many teams.
If you don't have the challenges that OKRs are a decent fix for, don't use them. You can read some of my posts tagged 'alignment' for real work examples that led us to use OKRs and we then adjusted them in the ways I described in the Improving OKRs post and since then, writing about experiences using OKRs I found others globally who had arrived at a similar mix of practices.
We had over 20 teams at Seek Asia all working on complicated systems and it was not always obvious when there might be good opportunity to collaborate for mutual benefit. We had various ceremonies for sharing what we learned, what each team was doing, how they were approaching problems etc. but we also found opportunities to collaborate by merely sharing what intended to try and achieve. Maybe that better illustrates what I mean by alignment. If you are imagining this is how work is managed I can understand why you might be confused.
I had some other comments to make but lost them when the page refreshed so I will post this and come back in another comment tomorrow 😁
Hi Daniel, I’m not sure where the team aspect has come from. I don’t think I specifically mentioned OKRs for teams or team outcomes. I scrolled back up but didn’t see anything.
A few reasons - OKRs are often set both organisationally and by teams. You also mentioned bad management - managers manage both their area of responsibility and the people involved (manage is a loaded word).
I am usually going to respond by describing how OKRs have been used in the actual environments I've seen them used. To describe what was the stimuli at the time and what led to using OKRs. You said you had less experience with them so short of trying it yourself when a relevant problem comes up (and you shouldn't if it doesn't) you figure the next best thing is you asked someone who used it in multiple organisations and observed it used in others. If that's why you are asking I am happy to keep sharing.
I've mentioned organisations align these constellations of objectives that exist with teams and company wide through causal trees which map not just the current focus but all relevant theorised causal links (that's helps with theorising where interventions might be taken in a way that's accessible whether you are familiar with the technical systems involved or not). Its remarkable how many organisations I've observed arriving through simultaneous discovery this way of working. So much so I keep writing about it :)
Its also interesting that similar approaches have been described dating back many decades and such approaches feature in the work of Goldratt as one example. Even though there is prior art I say simulataneous discovery because I believe most of the orgs I am referencing arrived through necessity and experimentation rather than necessarily being aware similar approaches had been used before.
You also mentioned OKRs being set that were not considerate of the whole of the system so I think providing some visibility on how teams use them vs how you might imagine teams use them may be helpful.
Oh and one question of yours that i did not answer was the John Willis one. I think the separation of means (or method) from objectives, outcomes and key results is my biggest concern.
Thanks Martin - yes you've made variation of this statement "the separation of means (or method) from objectives, outcomes and key results is my biggest concern." numerous times dating all the way back to the first LinkedIn thread where we discussed this. I asked my question hoping you will elaborate. Your post helps in that it illuminates the theory well and does a little to describe how you think OKRs are applied.
People working with OKRs in the manner I describe focus most of their energy on the means - they don't use OKRs in a vacuum. They use all the tools we've mentioned (goal trees, XmR, architecture diagrams, team interaction diagrams, other models) to try and understand how the system is working and to collaborate on creating shared causal theories of what changes will yield the desired behaviour of the system. OKRs is another set of information that teams define to help share intents about adjustments to the systems they will experiment with and find connection with what may be desirable for uses of the software they are producing. And when multiple teams have interrelated, interdependent changes they often articulate the intent in a shared goal in addition to all the other useful artifacts for collaboration on a shared understanding.
They don't set them to be judged by management they set them aligning on what they desired to achieve as a team and for mutual discovery amongst other teams - helpful when there's a lot of teams. Will bad managers judge teams based on lack of achievement of goals? Without doubt. In a system which incentivises them that success is fairy dust in the form of manipulated numbers they will do that - so you focus on addressing the incentives that lead to that.
Can OKRs help reinforce that poor response? Yes - where there is low literacy of how causality and systems work then any use of numbers can be interpretted to be how they may have been used elsewhere. I've written a lot about how we approached learning and raising the understanding of these concepts across the organisation. Its one of the reasons they can appear deceptively simple but in practice can be hard to do well. Sometimes the relief from other issues and the rewards justify the investment.
If OKRs are being used to manage outcomes I can see how separation from means leads to an issue - e.g. if all leaders did was to set a goal and measures and then come back three months later they'd have very comfortable jobs and the company would fail with declining quality and overwhelming side effects. I've even seen clients I've been consulting with do this (I don't consult on OKRs - I do consult on strategy, how to draw connections between team interactions, system interactions, linkage with strategy etc.) and its clear abdication of responsibility. They do this with or without OKRs.
I can empathise that its tough to imagine without experiencing the conditions that led us to hire OKRs. That's why I am happy to answer any questions about the experiences using them in different contexts. Why we used OKRs at Seek differed to how we arrived at using them when they were used in a startup and how the implementation differed in each context. I've had a long career working at many companies and consulted and coached at even more and OKRs were not an answer to a question we had at many of them - either because the need wasn't there or something else was hired to solve a similar problem.
I hope that provides some specific examples that are relevant for you - I am afraid I feel like I am guessing because you've ended a few times with your statement about separating from means without translating that into what you think actually happens.
Thanks Daniel. I’ll focus my answers on your specific concerns. Apologies that I have not done so in the past. I’m also up for a call to discuss in person if that’s easier for you.