When Stepping Back Is the Intervention
Knowing when your silence helps a team learn beats knowing when to jump in.
When you see your team heading toward a mistake, your instinct to step in is usually right. Acting on that instinct right away, however, usually isn’t the most useful thing you can do.
Watching a team walk toward a problem you can already see, and saying nothing about it, feels a lot like negligence. As leaders, silence reads as inaction, so we fill it. I’ve filled it more times than I’d like to admit, catching a team right before a mistake landed and, in the same motion, taking away the opportunity that would have taught them something.
When Silence Feels Like Negligence
Most of the urge to intervene comes from our own discomfort at watching a mistake unfold in real time. That discomfort is honest, but it doesn’t tell you anything about what the mistake will actually cost.
Teams read a leader’s silence as either trust or absence, depending on whether the leader is paying attention and choosing not to interrupt, or has simply checked out. From the outside, though, the two can look an awful lot alike.
It’s easy to confuse “I could stop this” with “I should stop this,” since stopping it “feels” like leadership. Letting a fixable problem run its course, on purpose, doesn’t feel like much of anything in the moment. The payoff comes later, in a retrospective or a demo, when the team says out loud what went wrong, why, what they did about it, and, most importantly, what they learned from it. An early correction would have taken that away from them.
Two Doors, One Test
Jeff Bezos framed a test for a different problem (how much process a decision deserves), but it translates almost directly to this one. In Amazon’s 2015 letter to shareholders, he split decisions into two kinds. Irreversible calls are “one-way doors”: “If you walk through and don’t like what you see on the other side, you can’t get back to where you were before.” Most decisions, he wrote, are “two-way doors,” changeable enough that a suboptimal call just gets reopened and corrected.
Apply that same split to the mistake unfolding in front of you. A two-way-door mistake costs a sprint or a demo to fix once it’s visible. A one-way-door mistake doesn’t get that grace period.
Which Door Is This?
A wrong estimate, or a design nobody stress-tested. Recoverable inside a sprint. Let it play out.
Customer data leaving the system, or a safety check skipped. Step in before it closes.
Most of what surfaces mid-sprint is a two-way door. We still jump in, because up close, even a reversible mistake looks alarming.
Let the Retrospective Do the Work
If you hold back, the correction still gets delivered, usually in the team’s retrospective. Someone on the team names what went wrong, often pretty close to where you would have landed anyway, and the team hears it from a peer instead of from above. In my experience, that resonates a lot better than the same words coming from the boss.
I first wrote about this back in 2015, around Ben Linders’ story of a program manager who stayed quiet at a stand-up and let an unfinished story surface in the demo and retrospective. What I didn’t have then was a good test for when staying quiet is the wrong call.
Of course, all of this only works if the team already trusts that naming a mistake out loud won’t cost them anything. Psychological safety has to come first, before a team can catch its own two-way-door mistakes without someone senior stepping in. Without it, stepping back just looks like abandonment, and honestly, that’s a fair read if nobody on the team feels safe saying what actually happened.
It’s the same instinct behind not front-loading your own answer when coaching someone. I’d rather be right second, after the team gets there on its own, than right first and have them just take my word for it.
I still get the calibration wrong in both directions. Sometimes I jump in on something that was clearly a two-way door (old habits). And once, I stayed quiet on something a lot closer to a one-way door than I’d realized at the time. The test doesn’t make the call for me, but it does change the question I’m asking. Instead of how uncomfortable I am right now, it’s how much it would actually cost to reopen that door.
So, how much of what’s crossing your desk this week is actually a two-way door? For the ones that are, try letting the team find the mistake on their own.
