Tuesday. 09:12. My office door opens with the delicacy of an auditor at quarter close. Manolo walks in slowly, like a man bearing bad news wrapped in methodology.
New notebook. Expensive pen. And wearing a suit!!!
That “this month we're fixing this for real” look.
—Jose —he says without sitting down—, I've got it. We're going to professionalize how we evaluate the technical team.
—What did you break this time?
—Nothing, nothing. This is a good thing. KPIs for everyone.
There it is. The slow poison.
—KPIs for what.
—Productivity. Impact. Focus.
Estimated hours vs. actual hours, tickets closed, velocity per sprint, reopened incidents, commits, meeting participation…
He pauses. Smiles. He's proud. As if he'd just discovered fire… in Excel.
—This way we'll finally know who performs and who doesn't.
I breathe. A long one. Because otherwise I might pour my coffee on him on purpose.
—Manolo… do you know what happens when you turn intellectual work into a KPI?
—It improves.
—No. It deforms.
Silence. That awkward silence in which Manolo starts to suspect he might not be the misunderstood genius he thinks he is.
—Look —he continues—, this isn't micromanagement. It's tracking. Transparency. Accountability.
Accountability. The favorite word of people who can't measure value but sure can measure motion.
—Manolo —I tell him—, a senior engineer won't optimize for delivering value. They'll optimize for not looking bad in your spreadsheet.
—That's just your assumption.
—No. It's Goodhart.
He gives me a weird look.
—“When a measure becomes a target, it ceases to be a good measure.”
—That's economics, Jose.
—And you're managing people as if they were inflation.
I explain slowly, because Manolo needs time for ideas to pierce his armor:
If you measure commits, there will be useless commits. If you measure closed tickets, there will be tiny, mediocre tickets. If you measure velocity, nobody will take risks. If you measure absence of errors, nobody will innovate.
—But I need visibility.
—No, Manolo. You need trust. And that doesn't show up on a dashboard.
He shifts in his chair.
—But then things happen.
—Of course they happen. Because building product means deciding without certainty.
—But I need to know who's failing.
—And what if the one who's “failing” is the one keeping us from making a historic screwup?
That bothers him. A lot.
—Jose, this works in big corporations.
—Yes, Manolo. It works for optimizing mediocrity at scale, but here we have people who escaped the prison of indifference to survive as soldiers of fortune and….
They are not the A-Team.
I tell him what Daniel Pink already explained in Drive: motivation in complex work depends on autonomy, mastery and purpose. Badly placed KPIs destroy all three.
Even so, well-placed KPIs can be neutral or even compatible with Pink's model, but that's very unlikely to happen, because the KPIs running through Manolo's head are not those.
I remind him of Ed Catmull, the Creativity Inc guy: when people start working for the evaluation, they stop working for the result.
Preserving spaces of trust for the people who create is very important.
And Andy Grove (High Output Management) I understood as: if you need to watch people constantly, the problem isn't the worker, it's the system… or the boss.
—So we don't measure anything?
—We measure, Manolo. But to learn, not to punish.
—And how do I know if the team is doing well?
—Look at the product, look at process metrics, but careful, that's the HOW. Don't measure outputs.
If you measure the HOW MUCH, we're already in trouble. Some of the smartest people in your company will game the KPIs without a moment's hesitation.
—That's very abstract.
—No. Abstract is your Excel.
I explain the difference he never wants to hear:
The business side can and should have KPIs. Product and technology don't work for the KPI; they work to reduce uncertainty and create options.
When you turn everything into monthly numbers, you push people to look productive, not to be useful.
—But I need to justify decisions.
—Then hire robots.
—What?
—Oh, no. Wait. If you keep this up, they'll leave on their own; little Elon will unplug them himself…
He goes quiet. Looks at the notebook. Closes it slowly.
—Jose… I just want to get the best out of the team.
—Then stop measuring them like they're an assembly line.
One last silence.
—So what do I do?
—Define clear objectives. Give context. Remove blockers. Look at results over the medium term. And accept something that's going to hurt: not everything valuable can be measured every month.
He gets up.
—I'll think about it.
—Perfect, Manolo. But think about this: talent doesn't run from hard work. It runs from feeling watched by someone who doesn't understand the work.
He closes the door.
And for the first time in a long while… I feel Manolo has left not quite convinced…
Thanks for reading.
And while I'm at it, I wish you health, love and family.
The rest doesn't matter that much… trust me.