Jason and Eiso share their love-hate relationship with OKRs and discuss the pedantic nature of setting objectives and how they ultimately help to align and empower organizations. They talk about setting and tracking goals, the often performative nature of OKRs, and the dangers of abstracting oneself away from the product you’re building.
The notion of setting objectives and empowering an organization bottom up is why I love OKRs. But the reality of it is that they seem to only work for the person who also shows up every day at 6:00 AM to the gym. And they don't really work for everyone else.
You can abstract yourself away from the actual thing that you're building and insulate it with a whole bunch of metrics and numbers that make everyone feel incredibly good, that you're doing the right sorts of things, and literally have no idea what is going on.
I don't think goal setting's natural to every person, in fact, I don't think goal setting's natural to most organizations, because of the people who tend to run them react as opposed to plan and think about this.
OKRs are a symptom of a higher level thing: at the exec level we don't know how to run companies. How many executives are looking at a spreadsheet that's saying the project is on track, it's green, as opposed to seeing the product, as opposed to seeing the features being built, as opposed to seeing the day-to-day progress or the week-to-week progress of that thing.
During the discussion about how some companies use tools without considering their practicality, Jason mentions the Radford Levelling as an example. He explains that some companies view these tools as a safety net, allowing them to do whatever they want without thinking about the consequences.
Radford's approach for determining salaries is based on aligning job levels with their value to the company and the market. The system groups employees into global grades based on what the market pays for similar positions, taking into account both individual contributor and managerial roles.
Here’s an example of how different job families are graded and grouped into two tracks, one for management, and one for individual contributors:
"High Output Management" is a book by Andrew S. Grove, the former CEO of Intel and father of OKRs. The book is an essentially a guide to OKRs, and how they can help managers boost their org’s productivity.
Jason mentions the book in the episode, and cautions about taking it off the shelf and adopting it in your org without considering your context.
Jason: It's the performative nature of it. It's, it's what is an O? What is a K, what is an R? What are we supposed to be doing? Like, oh, it's safe for me just to show up on the Monday morning or the Tuesday morning and just go over this. But it's not safe for me to say something doesn't feel right about this. In that meeting, you might do the behind the scenes, all that sort of stuff.
But like it should be there and it should have been like, Hey, everything's off track. Blow up this meeting. Who the F cares? This thing's not delivering and we're off sales says they're gonna hit their mark. They're not gonna hit their mark. This thing's not gonna be out there. Marketing says they got the quotas in place. They don't. We, we know that. Like, let's just stop doing this.
VO: Welcome to Developing Leadership, the podcast for engineering leaders. Were Eiso Kant and Jason Warner share their lessons on the ins and outs of managing software teams. On today's episode, Jason and Eiso share their love-hate relationship with OKRs and discuss the pedantic nature of setting objectives and how they ultimately help to align and empower organizations.
The guys talk about setting and tracking goals, the often performative nature of OKRs, and the dangers of abstracting oneself away from the product you're building. As always, this episode comes with accompanying show notes with a deep dive into the main topics, mental models, and key moments from this episode. Find them at developingleadership.co and linked in the episode description. Developing leadership.
Eiso: Hi everyone. Welcome back to another episode of Developing Leadership. Jason, today you and I are talking about a topic that we've been procrastinating on.
Jason: OKRs. . Yeah. Sounds like we need to get this outta the way.
And it's about time, right?
Eiso: Well, I guess OKRs are something people procrastinate on as well, so it, it's only perfectly fine that we procrastinate on the topic. ?
Jason: Yes. Uh, what, what are OKRs? Anyway, let's define that. No, no. I don't wanna, oh, thank
Eiso: God I was about to slap you over the back of the head. Uh, , if you don't know what OKRs are, don't listen to this episode, uh, or do listen to this episode.
Uh, you might not end up using them. I have a huge love-hate relationship with OKRs. I use OKRs in my own. But at the same time, I hate the pedantic nature of what's the O and what's the K? Yep. So I think that's, that's an obvious one. We can go on a little bit about that, but I wanna maybe even like go a level deeper, which is, you know, the notion of of, of setting objectives and empowering an organization bottom up is why I love them.
But the reality of it is, is that they seem to only work for the person who also shows up every day at 6:00 AM to the gym and they don't really work for everyone else. and I kind of wanted to start us off there, like the human psychology and why they work for some people and not for others.
Jason: You know, I never thought about it that way, but that's interesting that you put it cause I think that that goes to maybe somebody who would've been successful no matter what was around them, whether they had OKRs or didn't have OKRs.
They would've been able to figure out how to. Something done otherwise, right? Like there's a drivenness to them.
Eiso: Is goal setting natural to everybody?
Jason: No. I don't think goal setting's natural to every person. In fact, I don't think goal setting's natural to most organizations because of the people who tend to run them react as opposed to plan and think about this.
So I think that in general, like if you think about how people approach their life, nobody's actually planning out their life and then take it to an extreme. There's people who talk about on Twitter how they have quarterly board meetings with their families, which I think is in different type of extreme that doesn't need to exist in the world, but, you know, I don't think goal planning's all that natural to people.
Eiso: It's funny by the way, you, uh, you put us on the tangent straight away with the, you know, the, the quarterly family board plannings. I really, really wonder, like, I'd love to sit in one of those, like, you know, like the fly on the wall in one of those rooms one day because it feels absolutely terrible. But at the same time, like deeply fascinating, which I guess, uh, is probably similar to how people come look at OKR processes.
Jason: So this to me, like OKRs and even that comment about the offhand comment, quarterly goal planning about the family and all that sort of stuff, like to me it, it's at its root. Like I've always said, love hate relationship with OKRs is something you and I are similar to that way.
I said I love alignment tools and that's when I put OKRs and V2MOMS and other things in the same bucket, which is an alignment tool. And I believe that there is an a minimal amount of effort that needs to go into aligning across the organization to actually get maximal alignment. But it's important that it's the minimal to get the maximal.
This has to do with like all the other things in life in general, which is how much, do I need to work out? What type of intensity do I need to do to maximize my return for investment? Now, I'm never gonna step on a stage again and do any sort of like competing in, you know, bodybuilding and power lifting or something like that.
So I'm after a different goal. So what is my goal? What is my input to, to my output that I actually want here? That sort of thing. Taking a random. workout off the shelf. That is not for my goal is going to lead to bad results. Taking OKR processes off the shelf from Google or high output management and just applying them to your organization is not gonna have the result you want because you've not thought about.
Ultimately what you're actually trying to achieve.
Eiso: And I think it's a zero to one problem as well, right? Like it's, uh, usually when somebody wants to bring in OKRs in organization, it's from zero to hero is the intentions trade away. And, and that's why we have these classic examples of it failing time and time again.
So I'm kind of curious. You know, I, I was talking about this recently with my new VP Operations and we were making a change in introducing individual level OKRs, uh, for the first time in the organization, and it created a natural, sense of, of, of some chaos and fear and breaking things.
And she said to me, you know, when, when bringing OKRs as a tool for alignment, it can't be smooth, it has to be painful for a little bit to actually drive that change home. And, and I was curious if, if you think about OKRs in, in the setting of engineering, is there a non-painful way of, of making them part of a tool of alignment? Or will there always be some kind of needed pain you have to push through?
Jason: So a couple things about play here that I think that contribute to this. So one is, I think it's a very, very common pattern. Perhaps not in your case, given what I know about the, the way you guys run internally and stuff. But a lot of times when someone actually wants to bring in OKRs, it's because it's really far going a different way.
And when it's very far going a different way, think about Newton's cradle. You need an equal and opposite reaction to get an output that changes. So in that regard, yes. But I, as a general rule, I would say if you're gen gently introducing these things over, you know, a kind of a consistent period of time with healthy, steady growth. No, I would not say that. But I would say that if you've, if you're in Chaos Land, you've been in chaos land, you know, you know, chaos overlay the Lord of the Flies overlay with lost type of scenario. . Well, yeah, you're, you're actually trying to break a whole bunch of things at once and OKRs or the pain of an OKR process or an alignment process could be useful.
That said you in, like, I've been the one to introduce OKRs to several organizations, including GitHub when I was there mostly. That is what we were, it was Lord of the Flies overlaid, with LOST, overlaid with, massive amount of just anarchy and chaos. And, you're trying to just rebalance something. You're trying to do almost anything to get this thing on some sort of a better trajectory.
But you gotta stop somewhere in my opinion too. And at some point it becomes an exercise in OKRs for OKR sake, in my opinion, most organizations end up in that land. And that to me is when it becomes playing company, paper, pushing, whatever you want to call it, all the way around. But it's all this performative stuff. And when it's performative stuff, then you're not reevaluating why you're doing something in the first place.
Eiso: So, so let me break down the OKRs really into two parts, right? It's the setting of the OKRs, right? I would say the magic of OKRs is when you really use it in a way where everyone feels empowered for really aligning around how do we move the business forward. How do I see that what I do moves the business forward?
But then you have the, the tracking and updating and progress tracking part of OKRs. And I'm curious, I already see your face. I'm just gonna stop here. I'm curious to get your thoughts on that part and when does that look good? When is it successful? When is it not?
Jason: This is tough because I don't think I've ever seen what I consider to be a successful tracking of the OKR. I've never done it well personally. When it was no longer mine, we didn't do it well. It became this, this weird thing where, and I joke about this a lot, where projects were green, green, green, green, green, two weeks before they go live, red .
And it was, it became a joke with me, which was, you know, okay, so everything is green on this page. It'll be red two weeks before we go live. Just Mark, everything as red and let's start talking about it all. And so you're basically throwing out the entire process no matter what. And in reality, that's what kind of happened.
And it happened more as we got larger. And GitHub's context says we got bigger and we brought more quote unquote professional executives in to run things. It became more performative. And then you needed tpms, then you needed teams of people to do this spreadsheet, to this spreadsheet, to this spreadsheet alignment.
And then you say, well, you're looking at the wrong one. That one's outdated. And oh my God, let me tell you how many times I flipped my lid over that, uh, in the two or three years post-acquisition. But-
Eiso: I believe that, why does the green, green, green, red, red happen? Because before we go down the rest of the part, like is it, like you said, there is a, there's a performative side that makes sense, but there's also a non performative, genuine reason the green, green, green happens.
Jason: So I don't even know what territory we're gonna end up in, in this answer. So listeners ahead of time, forgive me for taking whatever this is supposed. This episode is supposed to be off track, I believe, mostly because OKRs are a symptom of a higher level, thing, which is at the exec level, we don't know how to run companies. And what I mean by that is, again, it gets to performative. It's safe. It's safe to introduce OKRs because they have been put into a bucket of you are doing the right things. It is safe to do performance management a certain way, or its safe to do HR systems, or it's safe to call Radford for your leveling and all that sort of stuff.
Like we've entered OKRs, have entered a safe territory. And safe does not mean valuable or useful or anything other than you won't get fired for introducing them and, and operating within the principles by which they have been constructed to adhere. But how many executives are looking at a spreadsheet that's saying the project is on track, it's green, as opposed to seeing the product, as opposed to seeing the features being built as opposed to seeing the day-to-day progress or the week-to-week progress of that thing.
And at the end of the day, if you're looking at a spreadsheet, like, you know, you've got that on your tab, you open it up, you're like, oh, that project's going well, great, okay. That sort of thing. When is the next time you're gonna see the screens of that? Or you're gonna be, you're gonna feel it yourself, or you're gonna log into the product and go use it or have it demo to you, or any of those sorts of things.
And like, and I, I know I'm. I'm like all over the place, but the point being like you can abstract yourself away from the actual thing that you're building and insulate it with a whole bunch of metrics and numbers that make everyone feel incredibly good, that you're doing the right sorts of things and literally have no idea what is going on. Literally have no idea what has happened.
Eiso: The feature might not even exist as long as it exists on the spreadsheet, right? Exactly. Right. To an extreme.
Jason: It's the same thing that happens with Jira. It's the same thing that happens with any of these, you know, planning tools. It's all that sort of stuff.
This, to me, is, again, it gets into the safe, performative nature of, but, but what is the exec supposed to do? What is the high level leaders supposed to do? And, and you know, that's why I fundamentally believe in building companies differently for the long haul and there's a mechanism by which you can do it in the, in the early days versus the midday versus the later stage versus the long-term hall. But they're not fundamentally all that different. They just have to scale.
Eiso: So it's interesting cause it's bringing us in the, in even more tangent territory, something I spent a lot of my time thinking about hell, that's why I'm building my company is that that problem that you're describing is, you know why as execs, do we get to the point where we look at a spreadsheet?
It's because we're, we're trying to get the synthesis or the summary because we're too busy for anything else, right? There's so much information at some point in a system that we say, give us the abstraction, give us the abstraction, and we kind of Chinese whisper it all the way to the top until it looks like something that's, that's green or red on the spreadsheet, which is I think a symptom of two parts.
One is we're we're focused on too many things we're doing and we're doing too much performative play as, as leaders, right? There's, if you cut older, performative play, we can spend more time on the things that are, are necessary.
But I also think it's, it's a lack of technology and tooling, right? We haven't gotten to a place yet where the world is good enough to be able to say, hey, I want to really understand like a genuine. Summary from the where the work happens in the organization, right? Can be the engineers all the way up to the top. And I think you and I have discussed this in, in, in an episode a while back, right?
Of status updates, right? From engineer to to to team lead to em and to the point where it become, you know, if you would show it back to the engineer at the bottom, be like, what the hell is this? And I feel that we're starting to get into a realm of technology in the coming years where we don't have to rely on the human meat bags to do that synthesis anymore. And I mean, let me pause there, before I go fully off the road.
Jason: I think we all hope or think or maybe dream of a day when things like AI can come in and help us know more about our own systems and things of that nature. I do think that there's a lack of tooling, uh, again to go back to experience with these lack of experience and some of these things, so people can't tell.
So, I mean, let's go back to sports land. And again, real quickly for a couple of different things, it's difficult to evaluate players when they're making a transition from college to the pros. So it becomes safe to do, like combine metrics. How fast are the 40 speeds? How high are their vertical jumps? What are their bench press things?
But none of that stuff directly translates to performance on the field. Yeah. So, but it's become safe to say you drafted the best athlete and all that sort of stuff and all. How, how often do they not work out? We've never actually gone back and evaluated and like, let's evaluate some. Some recruiter or some GMs best who, who's the best in the game and what do they do differently? Sports even just keeps doing the same thing year after year, after year, after year, after year, after year. Because it becomes a safe way to do that.
And so I go back to like what we do, lack of tooling. We don't measure, we don't have the right insights and that sort of thing. And then also, who's got the hair on the back of the neck? Who's got the experience, who's who, when they make judgment calls. make the best judgment calls. And if you think about where we're at the trajectory of our entire future with AI type of stuff, with people still involved in some of these cases, it's gonna be about the interpretation of a bunch of these things to make judgment calls at some point.
And you know, we got attune to that. Even going back to OKRs like, like specifically, since we're supposed to be talking about that here too. Again, it's the performative nature of it. It's what is an O? What is a K, what is an R? What are we supposed to be doing? Like, oh, it's safe for me just to show up on the Monday morning or the Tuesday morning and just go over this.
But it's not safe for me to say something doesn't feel right about this in that meeting, you might do that behind the scenes, all that sort of stuff. But like, it should be there and it should've been like, "Hey, everything's off track. Blow up this meeting. Who the F cares? This thing's not delivering and we're off sales says they're gonna hit their mark, they're not gonna hit their mark. This thing's not gonna be out there. Marketing says they got the quotas in place, we're they, don't we?" We know that. Like, let's just stop doing this.
Eiso: I couldn't agree more. You mentioned in relation to the hair's on the back of your neck about different people and, and their decision capabilities.
I, I had a leadership offsite last week, uh, including a new leader who had just joined. And I shared something with the team, which I realized I'd never shared. And gladly, I, I feel we have the psychological safety environment to do so. I said about every single one of you in this room, I keep a batting average in my head.
I keep a batting average on these decisions you've been right more than I have, or right. More than someone else. And you have to, of course have the culture in which that that's considered a, you know, a good thing to, to be able to say. But I think it's, it's one of the things that we implicitly do or should be doing as leaders, right?
And as engineering leaders in particular, we do this with engineers all the time, right, like this feature matter, if I'm a team lead, like this specific thing, I know it's critical to the business. I'm going to give it to person X. And of course we have to create the environments in which we allow people to actually stand at the bat. And while you see me trying sports analogies, right? But I think it's, it's something in within the notion of OKRs is that they're not linked to the batting average of the team, of the department. Well, there's no history of them. There's no, there's no history. Yeah. Yeah.
Jason: And, and so going back to that, that is, I love the batting average analogy. And this is something I've used in the past, particularly in the office of the CTO which we talked about in a couple of previous episodes. And I talked about what type of percentage I want to have. And I said, if you're in horizon one, and for those not. Listening, like I took the structure of Horizon one, two, and three and said, I want us to live in horizon three.
Horizon one from a product perspective is a hundred percent. You should know exactly what you're building for, if the adoption should be there, you should have done all the work already. Horizon two, like in my parlance, you should be hitting, you know, is an 80 percent-ish approach. Like 20% of it might be slightly off, but you're tweaking.
But 80% of it's dead on, we're right, all that sort of stuff. And by right, I mean on track it hits the numbers, you know, you hit your delivery dates. Like this is the work I'm talking. And the office of the CTO at the time, which is horizon three, I said 20%. And I said, it's, it's, it's so remarkably different.
And so what I was trying to reinforce here was think big, think long fail, go fast. Really have to push this. So go back to your analogy here. It's like, I think that this is something that we really need to push inside organizations, and this goes to the, again, a broader picture about building companies.
You know. How often do you want someone to be right and you know, micro decisions, you want them to be right quite a bit, but do you want them to be fail fast? Correct. At the macro level, if you've got somebody who's like Andy Jassy, when Bezos was there and you said, Hey, I want you to be right nearly a hundred percent of the time, if you're gonna bet over a billion dollars of our money, you know, that's one thing. Or you can say, I want, I want you to be right 80% of the time or I want you to be right 50% of the time, like, look at Alexa, Alexa's famous because they said it's probably a failed business line, losing 10 billion in this basically a black hole. But it was an experiment, like, is is somebody gonna be held accountable for that?
Or is it the type of place that they wanted to make that sort of a big strategic bet? And if it failed, you know, in, in my view, celebrate it, right? Like, so what type of place are you building?
OKRs don't help with any of that, they don't help with understanding the psychology of your organization and what it's supposed to be.
OKR is again the performative nature of the hard work. Like that, that conversation we, we just had at the strategic level, Bezos had his offsite and said like, F you all, I want you to be able to fail and spend 10 billion because we need to become this. That sort of thing. Like, I need you to think big and do that. OKRs never play a role in that. They never, they never have come up once that conversation. And Bezos just did a world-class job as a CEO.
Eiso: Yeah, and I think this is, I think this may be our final word on OKRs. Yes. They're an incredible, powerful tool for alignment. Probably you should be either doing them as a prophylac or or as a way to try to shake things up.
But the moment it becomes performative and the moment you see it as a solution for all of the other problems that you have. And the moment, I guess, and I think you said it well today, you don't dare to say in the middle of the quarter, fuck the OKRs. This is more important right now. Then they've taken over your company and they're holding it hostage.
Jason: The last comment on this would be like, going back to that Bezos example there. Whenever I see a CEO saying like, yeah, and I, I introduced OKRs. I'm like, okay, we're stepping back, we're recalibrating what your job is, that's not what your, your job is not to tell me or celebrate the fact that you introduce these things, that sort of stuff.
You gotta sit in a room and be that Bezos. And the other is, I really, really want people to understand that the moment that you're not thinking about the health of the business long term, or you're not building all that sort of stuff, but you're doing the performative nature, it, it's become a very different job.
And you know, you're in, anybody inside of an organization who's using OKRs knows exactly what I'm talking about. Anyone inside a very large organization who using OKRs, understands the incentive structure of hitting your OKR so you get your bonus and all that. The plot has already started to be lost.
Eiso: Absolutely, and it's directly correlated to happiness.
The moment you're in a leadership role and you feel that you're in, the more performative your nature gets or your job gets, I don't believe that leads to more happiness, I think all the times when you've been happy, you've been proud, you've been excited about what you did, is when you completely felt that none of your role was performative. No matter if that meant to fail or to succeed. I think that's all we have to say on OKRs people. It's gonna be a while before you hear us mention ' em again.
VO: Thank you for listening to Developing Leadership. Make sure you subscribe to follow us on our journey to more meaningful engineering leadership. If you have any challenges or topics you would like us to explore in an upcoming. Tweet or the at dev leadership underscore. To learn more about data enabled engineering and how metrics can help your teams and improve your processes, go to Athenian dot com.
See you in two weeks for another episode of Developing Leadership.