Tuesday. 08:47. I arrive early. Coffee in hand. Silence in the office. That rare moment when you can think without anyone interrupting you with an urgent matter that has been urgent for three weeks.
I open the laptop. I review the roadmap, I ask BorjIA to tell me what value the teams delivered yesterday. And then I realize something I've been ignoring for weeks: the team is rowing, but nobody knows where to.
It's not their fault. It's mine. There are too many open fronts, too many priorities competing with each other, and the strategic context has been dissolving into Slacks, dailies and "let's talk about it later".
So I do something I normally avoid: I go looking for Manolo.
—Manolo, I have an idea.
He looks at me like you'd look at a vegan ordering a T-bone steak.
—You? A management idea? For me?
—Yes. OKRs.
—OKRs?
—Objectives and Key Results. It's an alignment framework. John Doerr popularized it when he took it from Intel to Google. He wrote Measure What Matters. The idea is simple: you define an ambitious, qualitative objective, and then you define two or three key results that tell you whether you're on the right track.
Manolo pulls out the notebook. The expensive pen. The "Jose is finally speaking my language" look.
—I like it. Structure. Measurement. Focus.
—Yes, but listen carefully: the objective is NOT a task. And the key result is NOT a deliverable. It's a measurable change in the real world.
—Sure, sure.
He hasn't understood a thing. I know because he nods too fast. Manolo nods fast when he's already thinking about his own version, not yours.
—Look, here's an example. Objective: "Become the reference for user experience in our vertical." Key Result 1: "Cut average onboarding time from 12 minutes to 4." Key Result 2: "Reach an NPS of 55 in the enterprise segment."
—Fine, fine. And the tasks?
—The team decides the tasks. That's the whole point. You define the destination, not the path.
Pause. Manolo blinks. As if I'd just told him the earth is round and he'd spent years selling flat maps.
—Jose, but if we don't define the tasks… how do we know they're working?
—Because you'll see whether the Key Results move.
—Right, but I need to see activity.
—No, Manolo. You need to see progress. And they're not the same thing.
He leaves. Happy. Too happy. The kind of happy that precedes an organizational disaster.
The following Monday. Manolo sends me a document. Subject: "OKRs Q2 — Draft v1 FINAL".
I open it.
And there it is. The masterpiece. The crime.
Objective 1: "Improve the platform."
Key Results:
Close 45 Jira tickets before April 30. Deploy every Thursday. * Reduce open bugs to fewer than 20.
Objective 2: "Be more efficient."
Key Results:
Have everyone log hours in Harvest. Hold retrospectives every two weeks. * Cut meetings by 15%.
I breathe. I count to ten. I don't make it to five.
I call him.
—Manolo.
—Did you see? They're ready. I worked the weekend.
—I can see that. And it's exactly what I told you NOT to do.
—What?
—These aren't OKRs. This is a task list dressed up with a pretty title.
Silence.
—"Improve the platform" isn't an objective, Manolo. It's a wish. Like "be happy" or "eat healthy". It has no direction, no ambition, no soul.
—But it's clear.
—It's empty. A good objective has to hurt a little. It has to force you to choose. Christina Wodtke explains it very well in Radical Focus: if your objective doesn't give you a bit of vertigo, it's probably not an objective, it's maintenance.
—Fine, and the Key Results?
—Closing 45 tickets isn't a result. It's output. It's motion. A team can close 200 tickets without having moved the needle a single millimeter. Remember Goodhart's law?
—Don't bring up Goodhart again.
—Then stop summoning him.
I explain what he'd know if he'd read past the table of contents of Doerr's book:
Key Results have to be outcomes, not outputs. You don't measure what you do, you measure what changes because of what you do. "Closing tickets" is output. "Cutting the average resolution time of critical incidents from 48 to 12 hours" is an outcome. "Deploying every Thursday" is a practice. "Going from a 14-day cycle time to 5" is a result.
—But that's harder to measure.
—Exactly, Manolo. And that's why it's worth it.
Andy Grove, who basically invented OKRs at Intel before anyone called them that, had a very clear idea: the system exists to generate focus and alignment, not to control people. If your OKRs look like a timesheet, you've failed before you started.
—But the team needs to know what to do.
—No, Manolo. The team needs to know what to achieve. The "what to do" they figure out themselves. That's autonomy. That's what separates a senior team from an assembly line.
—That's all very pretty, Jose, but people here need to be told things plainly.
—Plainly is: "We want 80% of users to complete the purchase flow without contacting support." That's plain. "Close 45 tickets" is noise with formatting.
Manolo stirs. He looks at the document the way you look at a sandcastle right before the wave.
—And what about logging hours in Harvest as a Key Result?
—Manolo, that's not a key result. That's an internal policy. It's like putting "arrive on time" as the company's strategic objective.
—Well, there are people who don't arrive on time.
—And there are companies that confuse discipline with strategy. Don't be one of them.
I tell him about Marty Cagan, about Empowered. Cagan distinguishes between "feature teams" and "empowered teams". The former get lists of what they have to build. The latter get problems they have to solve. OKRs, done well, are the tool of the latter. Done badly — like Manolo's — they're the favorite costume of the former.
—So you're saying my OKRs are feature-team OKRs.
—I'm saying your OKRs are a Microsoft Project with delusions of grandeur.
Low blow. I know. But low blows are the only thing Manolo reacts to.
He stays quiet for a while.
—Jose… every time I try to bring order, you tell me I'm doing it wrong.
—No. I'm telling you that you confuse order with control. Order is giving clarity about what matters. Control is deciding how each person has to do it. The first generates alignment. The second generates obedience. And obedience, in knowledge work, is the prelude to mediocrity.
Peter Drucker said it decades ago: "Knowledge-worker productivity is the biggest challenge of the 21st century." And then he added something Manolo should get tattooed: the key isn't supervising more, it's better defining what constitutes a result.
Manolo closes the laptop. Slowly. With that "you've ruined my Monday but I won't give you the satisfaction of admitting it" face.
—So I redo the OKRs.
—Yes. But this time do them WITH the team. Not FOR the team. OKRs imposed from above without context or participation turn into what you already have: a checklist with corporate ambitions.
—And if the team proposes things I'm not convinced by?
—Then negotiate. That's leading. It's not handing out homework.
He gets up. Gathers the notebook. The expensive pen stays on the table, like an involuntary white flag.
—I'll give it some thought.
—Manolo.
He turns around.
—An OKR is not a contract. It's a compass. If at the end of the quarter you don't hit it but you've learned why, it has worked. If you hit it but nothing has changed… it has failed.
He leaves.
And I'm left thinking that maybe, just maybe, this time Manolo won't wreck the idea completely.
Thanks for reading.
Onward.