Why a cluster of departures is worth reading before you run the searches.
The first resignation is a person. The second one, a few weeks later, is a question.
By the third, most founders are deep in logistics. Backfill the roles, brief the recruiters, reassure the board, keep the trains running. The work of replacing people is loud and urgent, and it crowds out the quieter question the exits are actually asking.
I’ve sat with a lot of founders in that stretch, and the pattern is consistent. When two or three strong people leave in a short window, they tend to be pointing at the same thing. Something in how the organization is built made those roles harder to win in than they should have been. The departures are the visible edge of a structural issue, and that issue sits upstream of any one seat.
“Structural” usually means something specific: decision rights that never got defined, a mandate the role couldn’t actually deliver on, a seat set up to absorb blame for a problem it didn’t own. Capable people tolerate that for a while, then leave. Backfilling each role one at a time treats the whole thing as a staffing problem. You run three searches, fill three seats, and inherit the exact conditions that emptied them. A few quarters later the pattern repeats, and now it reads as bad luck twice.
A useful read is more specific than exit-interview sentiment. Look at where decisions kept stalling in the months before each person left, what they escalated more than once and never saw resolved, and which problems the role kept getting blamed for without the authority to fix. Those three questions tend to converge on the same upstream point, and that convergence is the thing the searches should be built around.
A single departure carries the same signal
You don’t need a cluster to read this. Even one high-quality exit is information about the seat.
At Lieferheld, the CMO seat opened and I was brought in to lift bottom-line performance. The obvious brief was to run the marketing function harder. The real constraint sat upstream of marketing entirely, in the positioning and the messaging. The market wasn’t responding to how the business framed itself, and no amount of channel work was going to rescue a message that wasn’t landing.
So the work started upstream. We repositioned the brand around warmth and humor, then let the spend follow the sharper message. Q4 acquisition rose 54% year over year, CPA fell by nearly half (roughly €24 to €12.45), and SEO reversed into double-digit growth.
Here’s the part worth holding onto. The problem lived one layer above the role, in the brief the seat was given. Had we simply refilled the CMO role as written, a capable person would have inherited a mandate pointed at the wrong layer and spent a year fighting it. When it’s a cluster of exits rather than one, that same dynamic is just louder and more expensive.
The open seat is the cheapest time to change the role
There’s a short window most founders spend entirely on logistics that is actually the most valuable moment in the whole cycle.
For a few weeks after a resignation, you have permission to rethink the role. You can ask what the seat is really for, what it quietly became over the last two years, and whether the job you need now matches the job description you inherited from the person who left. The moment someone new is sitting in it, that permission closes. You’ve re-committed to the role exactly as it was.
That window is when a diagnosis pays for itself. A job description written around the last person hires for a role that no longer exists. A description written around the real constraint hires for the company you’re actually running now.
Sequence the structure, then run the search
The move is simple to say and genuinely hard to do under time pressure. Define the role around the real constraint first. Then run the search. In practice that means writing the mandate around the problem the exits exposed, giving the seat the decision rights it was missing, and hiring for the person who can own that rather than a copy of whoever just left.
Do that, and you hire once, into a seat built to hold the right person, and the function finally does what you needed it to do. A great hire into a well-built role compounds. A great hire into a role pointed at the wrong problem still fails, and you carry the cost of the search, the ramp, the lost quarter, and the second search that follows. The salary is the smallest number in that equation.
The order is the whole thing. Find the constraint. Fix it in the right order. When people are leaving and the board wants the seats filled yesterday, the pull toward speed is real and reasonable. The few weeks you spend making sure each seat points at the right problem are what keep you from running the same search twice.
If you’re watching people leave right now
A wave of exits is hard to sit with, and the instinct to move quickly is a good one. It’s worth aiming that speed at the right question first: what are these departures telling you about how the organization is built?
If that’s the moment you’re in, I’m happy to spend a short call helping point toward where the constraint likely sits. No pitch. Just an operator’s read on what your exits are actually telling you.