OKRs sound good but they don't work (Part 1)

Hello hello, Tom & Corissa here,

We used to be called ​Trigger Strategy​ but now we're Crown & Reach.

This is part 1 of my OKRs don't work series.

• In Part 1, I introduce some social and neurological fundamentals of why OKRs don’t work the way they say they do on the tin.
• In Part 2, I dug into what we mean by objectives and how they relate to the adjacent possible.
• In Part 3, I introduce a little twist to OKRs that helps them work, which I’ve called Adaptive OKRs (AOKRs).
• Later: what to do when you’re struggling to choose an objective; context and how it matters more than you’d like; what to do as a leader to get more of the promised benefits of OKRs; and what to do if you’ve been gifted terrible OKRs.

Originally published via Substack, Sep 28, 2023.

If you’ve been in the tech world for any length of time, you’ll have come across OKRs. (If you haven’t yet, Christina Wodtke explains them better than anyone else)

OKRs are a good example of an approach that tells a pretty story about how folks think management ought to work, but that doesn’t match how management really works.

(There are many of these kinds of ideas out there. They sound nice, neat and so so simple. But when you try to put them into practice, they turn out to be painful, messy and complicated. Once you know this pattern, you’ll see it everywhere. Sorry.)

The disconnect between the simple story and the painful reality of OKRs has generated two camps:

Camp 1: “ugh, OKRs always suck and are done so badly.” They end up accepting that they have to do just enough to play along with OKR theatre, whilst basically ignoring the OKRs.

Camp 2: "you're doing them wrong!" This is consultants who know OKRs make a mess and promise to unsuckify yours. They say things like, “hey, don't blame the OKR framework for your problems! I can help you make them work properly.”

Both camps miss the point.

Let’s dive into why OKRs can’t work in the way they promise — and why that never really mattered for their meteoric rise in popularity.

Your brain hates OKRs (and here's why)

Humans don't work like OKR frameworks expect.

We don't do this:

1. pick an explicit objective (O)
2. define success metrics (Key Results)
3. Then make plans.

We do this:

1. scan the situation
2. do a first-fit pattern match against our past experiences
3. grab the first idea that feels good
4. unconsciously justify the idea so it feels like it's rational

Ideas that feel good are the ones that feel like they manage the risks our ancestors cared about.

Ideas that feel good feel like they don’t increase our exposure to risks – especially implicitly fatal ones like being cast out of our tribe or losing status.

Among the ideas that feel good, the very best ideas feel like they could plausibly help with several desirable objectives, not only one. And most of our desirable objectives are subconscious and personal, not conscious, explicit or tied to business metrics.

Why smart people game the system

I saw this pattern of pragmatic risk avoidance play out when I coached various teams through setting OKRs.

Most people had an idea or two that they just knew in their guts was worth doing next, but they were reluctant or unable to be specific about one objective that would be the result. They were choosing an idea that felt right. And it was based on juggling dozens of mostly subconscious personal goals, not on working backwards from a business objective.

Ideas that “felt right”:

1. were “the kinds of things that successful people with my role do”
2. were personally interesting, with enough-but-not-too-much challenge
3. could plausibly-but-vaguely be justified as helping the team and company.

At the time, it frustrated me that they couldn’t zero in on an objective so we could create some crispy OKRs. But now I see the rationale. These team members resisted specifying an explicit objective because they subconsciously felt like they would be putting themselves at risk.

Status risk

Firstly, the team members felt the risk that they might lose face — or even their job — if they didn’t achieve the objective.

Career-wise, it’s safer not to promise results, but to get on with justifiable actions. Then you can’t really be blamed when some of these actions inevitably don’t go to plan.

The reality is that most initiatives don’t produce predictable linear results. However, they do frequently have surprising unintended consequences – both good and bad. When there are unexpected positive results, the canny operator wants to be able to post-hoc rationalise that those were the goal all along.

Derailment risk

Secondly, the team members perceived that OKRs put them at risk of being stopped from doing the idea they felt good about. Their real, subconscious reasons for feeling so good about the idea were often personal, private objectives – interest, career progression, learning, or comfort.

Sarah wants to play with transformer models … Girish wants to experiment with RUST … Alex wants to fix a usability snafu (guilty as charged) … what if they pin down an OKR that sort of contains their idea and then an executive decides that OKR isn’t worth going after? Or makes them do something boring that achieves the same objective?

Why execs craft vague OKRs

I saw a slightly different flavour of risk avoidance playing out when I worked with executive teams to help them set company-level OKRs.

In those cases, it was even more obvious that they wanted to hide their real objectives.

Often, a company’s real objectives have to be kept secret (i.e. not plastered all over the walls and shared in every meeting like OKRs are).

For instance, many exec teams urgently need their business to “become investable”. But you can’t say that out loud! Not only does it sound uninspiring to the troops, it also makes you sound desperate — and therefore uninvestable — to potential investors.

What’s more, often the execs’ real objectives are unpalatable. Stating them clearly might expose the fact that three quarters of the workforce won’t be doing anything particularly valuable this quarter. Or the objectives might involve preparing to shut down a department … or worse.

So instead, the execs craft nice-sounding but deliberately vague OKRs that give everyone something to keep them busy. Then they micromanage a handful of teams to work on the real objectives (often even contradicting the official OKRs).

Naïve?

At this point you’re probably thinking, “Christ on a bendy bus, Tom, are you really this wet behind the ears? Of course this is how businesses use OKRs!”

Let’s just say I’m writing this post to my younger self, hoping he’ll stop believing that OKRs are really supposed to do what they promise.

The real reason companies love OKRs

In the end, none of the mismatch between how things are supposed to work and how they actually work matters.

Business leaders don’t impose OKRs because they’re effective at generating business results. They impose them because they’re a safe choice. They look tidy to investors, they’re logical for continued employability in the C-suite, and they make it look like you’ve done strategy when you stuff them into presentations.

In short, they “feel right”! They:

1. are “the kinds of things that successful execs with my role do”
2. can be personally interesting, with enough-but-not-too-much challenge
3. could plausibly-but-vaguely be justified as helping the business.

Sound familiar?!

So whether or not OKRs generate results is by-the-by. That’s not what they’re really for.

But maybe I’m just a hater? Maybe I never liked OKRs because I never saw them done properly?

Not at all! 10 years ago, I was all-in on OKRs. I read all the books, spent energy coaching teams and execs, and even used personal OKRs outside of work.

OKRs are effective … in a certain sweet spot

I recall one 3-month period when a team I was leading had an OKR and it was great.

We had:

• a specific, tractable problem evidenced by lots of user research
• within an otherwise stable and well-understood process
• that was also easy to measure
• with a constrained range of options that we could try.

Because this happened during a reorg, we fell between the cracks and lucked out. We had 3 months off the critical path to focus on the OKR, and also total freedom to set our own OKR.

The result? One of our A/B tests fixed the problem in a surprising way, making customers happier and the company millions.

The reward? Our experimentation program was put on hold indefinitely. Classic!

So yes, one time out of literally dozens, an OKR worked.

Every single other time, the OKRs have been theatre, gamed, a wish-list, or a task-list.

OKRs work in exactly one context: when you already know what to do and just need to measure it.

Everywhere else, you’ll find those 2 camps of people arguing: “OKRs are bullshit!”, "Um actually, you're doing them wrong!"

OK, so what if you want to stop using OKRs?

Then you’re going to hit the ultimate hurdle.

It’s hard to stop doing something — even if it’s ineffective, or even harmful — if you can’t provide an easy, comfortable alternative.

So even if you can persuade someone that OKRs are ineffective and harmful, you’ll also need to give them an alternative that feels easy and comfortable. That “feels right”.

I have an alternative that works with human psychology instead of against it. More in part 2.

Tom x

​

£5.00 Buy us a coffee We publish loads of articles, podcasts and videos for free. If you've found what we've shared helpful and you'd like to... Read more Send a tip

​