Two instincts pull at you, and both feel like virtue.
One says: this isn't mine to solve. Hand it down. Hold the line. An organization that needs you in every fire is one you have failed to build, and the discipline is in not touching it.
The other says: there is a customer on the other end of this, right now, and a promise we made is coming apart in their hands. Go. Lead from the front. The org can learn later.
Most writing on this hands you a rule. Delegate more. Or: get your hands dirty. The rule is useless, because the situation decides — and you already know that, which is why you're standing still, holding both instincts, watching the clock.
The trap underneath both instincts is the same one: a story that lets you avoid discomfort.
"Drive accountability down" is a real principle. It is also the sentence you reach for when you don't want to sit inside a hard problem — when delegation is just a respectable word for leaving. You call it empowerment. The owner calls it abandonment. The customer calls it nothing, because no one picked up.
"Roll up your sleeves" is real leadership. It is also the thing you do because solving is the only place you feel fully competent, and watching someone solve it slower than you would is unbearable. You call it urgency. The org learns to wait for you. Next time, they don't even try.
Both are avoidance wearing the costume of principle. The only question is which discomfort you're running from.
The customer cuts through the costume.
When the harm is compounding and won't reverse — a trust that won't survive the week, a relationship that breaks for good, a line you don't get to redraw — you go in. Not because going in is noble, but because the cost of the developmental moment now exceeds its value. You do not teach someone to swim by watching them drown on a customer's time.
When the harm is recoverable — slower, messier, embarrassing, but survivable — you hold. The owner owns it. Your jumping in buys a cleaner outcome today and a weaker org for every day after. That trade only looks good if you never count the second number.
The variable is not how bad it feels. It's whether the damage compounds, and whether it reverses. Urgency is loud. Irreversibility is quiet. Learn to hear the quiet one.
And even when you go in, going in is not one thing.
The instinct collapses straight to the most satisfying version: take the wheel. Hands on the problem, eyes on you, the room exhaling because the adult arrived. It feels like leadership. Usually it's the heaviest possible intervention applied to a problem that needed the lightest.
So before you take the wheel, ask what the owner actually lacks. Authority they don't have — lend it. A decision above their grade — make it, then step back. A blocker they can't move — move it. Hands, for one night — add yours, and leave them holding the work. Taking the wheel is the last option, not the first, because it is the only one that removes ownership instead of returning it.
Choose the lightest intervention that changes the outcome. Anything heavier is for you, not for them.
The frame was balance. Balance implies a trade — protect the customer or protect the ownership, pick one.
You don't pick. You sequence. Stop the bleeding first; the customer does not care about your org design. Then return the wound to the owner to close — the post-mortem, the fix, the rebuild, the accountability for how it got here in the first place. You went in to stop the harm. You leave so the lesson stays with the person who needs it.
Collapse the sequence and you lose both. Rescue the customer and keep the wheel, and you've taught the org that fires belong to you. Protect the lesson and let the customer burn, and you've taught them that a principle outranks a person on the other end of your promise. Neither is wisdom. Wisdom is doing both, in order.
So yes — pause. But not the pause that looks like hesitation, because you cannot hesitate over someone who is bleeding.
The pause is a single fast read: am I the only one who can stop this right now? If yes, move, and move all the way. If no, then who owns it, and what is the lightest thing I can hand them that changes how this ends — without taking the thing that was theirs to carry?
That read takes seconds. The instinct skips it entirely. The whole craft is putting the read back in.
When the volume is not the problem, but in fact the diagnosis.
There's a number where this rule quietly breaks. We learnt to ask whether the damage was permanent or merely expensive. We learnt to go in when a customer would feel it and carry it forward. That worked when the permanent ones were rare enough to chase.
Then one day the count is in the hundreds. Every one of them clears the bar. And you're still chasing. That's the moment to stop and read what the number is telling you. A hundred permanent issues is not a hundred fires. It's one system, manufacturing permanence, reported a hundred times. The instinct that served you when fires were rare — go in, fix this one — is now the most expensive comfort you can buy.
While you hand-fix case forty-seven, the factory ships case forty-eight. You are not behind on incidents. You are watching a defect and calling it bad luck, one case at a time. When this happens, Stop counting the cases and go fix what's producing them. Redesign upstream, where the defect is born, not downstream, where it lands.
Step in when the damage is permanent. Step back when it's just expensive.
When permanence becomes the norm, stop saving cases and go find the source of the problem.