A collective hallucination
a rant about OKRs
In the latest episode of “Developing Leadership”, Jason Warner’s rant against the fact that we all read the same book and listen to the same “thought leaders”. Reading this amazing story about Microsoft Office finally convinced me that managers, even the most senior ones, live in a state of collective hallucination fueled by the hype cycle of management practices.
The hype cycle
Gartner’s hype cycle is one of the ever present graphs of our modern times, you can map basically everything on it and it serves Gartner’s well to counterbalance competitiors’ like ThoughtWorks’ Technology Radar
Disclaimer from the author: I don’t and will ever understand how any of those companies add any value or why anyone would pay them for their work. I acknoledge that this is my personal limitation and not a proof that consultancy companies are useless.
Going back to the hype cycle, it’s form is pretty simple: it tracks the evolution of some kind of practice or technology, from inception to going mainstream with the inevitable bumps in the road of adoption. The hype cycle is also a self-fulfilling prophecy, as soon as something gets mapped its adoption grows, it’s what marketeers call “earned media”. But that doesn’t mean that is deserved or useful in any ways.
And company management continously fuel the cycle, some C-level reads a book, they start buying books for all other executives and “nudge them” to read the book. And they all convince themselves that this book is going to turn things around and tell their friends. Especially if this book is used by some FAANG company.
Nobody knows anything
A statement that I always make to younger managers is that nobody knows anything when it comes to people. Thousands of years of history show very clearly that humans aren’t rational and that they are most definitely non deterministic: You can do something a hundred times with a hundred people and always get a different result, this is what makes management difficult.
And if anyone tries to convince you that they figure it out all of it, well they are lying to you because someone told them that overconfidence and “fake till you make it” are the same thing.
And in this jungle we get lost: how can I lead my team knowing that I know nothing and that everyone expects me to fix whatever is broken at all time?
The seductive power of the theory of everything
In this context it’s easy to see why some books become overnight successes and practices spread like wildfire. Wouldn’t you like to fix all your problem with this one simple trick?
And in this way a every few years a number books take our companies by force with the promise of fixing all our problems: Good to great, the Innovator’s Dilemma, Measure what matters (OKR), Radical Candor. Honorable mentions to the Spotify method and the Netflix HR culture.
Not that the people writing those books are at fault, I am sure those practices worked for them and in fact I am a dude writing his own ideas on substack right now and I am generally against censorship1.
In this group especially two practices had a deep disturbin effect on our industry: OKRs and Radical Candor.
Radical Candor or how to destroy years of pedagogical work in one book
I admit it, when radical candor came out I had to read it to understand what was the fuss about. I was horrified by that book, but what horrified me the most was what I saw around me: a lot of people, generally rational and reasonable telling me that the book was “enlightening” and that they will definitely apply radical candor to their lives. And of course plenty of candidates in interviews telling me that they are absolutely apply radical candor on the workplace.
Radical candor is problematic on many levels but the main one is that it assumes that the only reason why people are not candid with each other is because they are extremely polite with each other or just outright lying.
But in my experience that is almost never the case, in fact most of the time the real reason is psychological safety: being candid assumes a level of trust in good faith that cannot be fabricated quickly, it’s not a quick fix and most definitely radical candor does not increase trust among individuals that don’t trust each other.
This requires a lot of work on your team to create trust, and a company culture that supports that. You can’t have a dysfunctional clusterfuck of a company and slap some extra candor on it to make it better. It makes it just more toxic.
There is also another subtle reason, not everything that you are thinking needs to be broadcasted to the rest of the world. I know we live in the age of social media and we have been told multiple times that feedback is a gift. Don’t get me wrong, thoughtful feedback is very valuable but many times even the smartest people in the world don’t have something smart to say.
So if you have been infected with the virus of radical candor please throw the book in the recycling and read “The 5 dysfunctions of a team” instead. Longer term, harder work but at least it actually works.
OKRs or how to make waterfall cool again
John Doerr’s influential framework is maybe the most misunderstood concept in the history of modern company management frameworks. Every company has its own interpretation and often the same executives that are praising the method have also no idea what they are talking about.
The problem with OKRs
Objective Key Results is a very intuitive framework, the basic idea is that to succeed your company should be aligned toward the same goals but this is hard when you have many layers of management and teams so you need something smart and easy to do that.
Not only that, since your company is clearly made of smart and talented people, you shouldn’t tell them too much what to do but rather give them aspiring objectives and the space to experiment and stretch their talent.
It has all the characteristics of the quick and easy fix that every manager wants!
Except that, in order for OKRs to work, a number of things need to be true:
The amount of uncertainty in your business is fairly small (we already know this to be mostly false for engineering and any innovative company)
Your teams are extremely mature and independent (maybe?)
All your layers of management know the framework, have been trained on it and are extremely efficient at running the process heavy practice of OKRs (not even gonna comment this one)
And even if all of the above are true (and please I would like to hear your story!) it is fundamentally a method rifled with wrong incentives (how can you as a manager be accountable for the objectives of your teams if you don’t have any agency in them?) and heavy process.
Fixing OKRs
In my personal experience most companies slowly transform OKRs into targets, they start with the whole framework only to discover that they just can’t keep it with the process and that the teams are just as confused as their managers are. Once I asked a Director of Engineering from Google how they were using OKRs and he told me that they reserved only 50% for OKR-related work while the other 50% was business as usual, I then asked what % of OKRs they should achieve and he said “if they commit to something they better make it”.
I found this formula to be quite reasonable, except it’s wasn’t OKRs!
And this is what I think is the best way to align your organization:
Assume that 50% time your people aren’t gonna work on the objectives you set, life is complicated, stuff breaks, people get sick etc etc. So instead of setting objectives for what I would call “Business as Usual” set some KPIs to make sure your teams stay on track (for example DORA metrics) and review them regularly.
For the other 50% set objectives that are very clear explanations of what you want to achieve, don’t try to be shakespeare just make sure the WHY is clear. Then associate to those objectives to SMART goals, measurable targets that are reasonable to achieve in the time frame you chose (one year maybe?). Those goals won’t be “stretch” goals, that concept is really confusing, your teams need clarity therefore those goals will be the minimum you expect them to achieve.
Do this exercise top/down-bottom/up (set a draft set of goals and get feedback from your teams) but at the end of they day you must be the one signing them off, being accountable for their success.
Don’t overcomplicate, don’t do too much process, no wordsmithing
Some final words
Not all management books are bad, in fact there are some great books out there that have been very useful to me (the 7 habits, Drive, the Five dysfunctions of a team) but none of those give you quick fixes, they all provide suggestions on how to fix yourself and your company culture.
We must resist the hype and the next “thought leader” that is going to present us with the next quick and easy fix.
None of them work, none of them ever worked.
This note to say that I am deeply against those movements that ask Spotify or Substack to censor people. I am horrified by Joe Rogan and the likes but you can’t force private entities to censor people without having the state doing the same. In fact I’d argue that is deeply unsettling to see people being happy when donald trump is censored by a tech company but horrified if that becomes a law. You can’t have both, either we have free speech or we don’t, making up your own rules and then pressure private companies to enforce them is going to end really bad (see abortion laws)