Your team already knows which customer you're going to lose

She can feel it three weeks before it happens.

The account manager on this file has watched enough relationships end to recognise the early signs: replies that used to come same-day now take three, a call moved twice, a contact who used to loop in the whole team now emailing just her. Nobody has said anything. She knows anyway.

So she does what she has always done when an account gets hard: she flags it to the founder.

He gets on the call. He negotiated the original deal himself and still remembers the pricing history without checking a document. He can speak to scope, to what was promised, to what can flex, in a way that makes the client feel heard immediately. The immediate problem eases. The account manager watches how it happened, the same way she has every other time, without being the one who does it.

This is the exact moment that should be building her judgment, and instead it replaces the need for it. The accounts worth saving are the hard ones, and the hard ones are where a team member needs practice reading the signal, deciding what to offer, and holding the relationship without a rescue call. Every time the founder takes it instead, the team gets better at escalating and no better at resolving. The pattern doesn't fail loudly enough for anyone to name it. It just repeats.

Eventually one of these accounts leaves anyway, not necessarily this one, but a later one, on a week when the founder is travelling, or in back-to-back meetings, or a call behind on the day it mattered. The founder's explanation afterward is usually some version of "they just didn't get what we were building." He's describing the customer. The account manager knows the more accurate description is the system: an escalation path that only had one destination, and the destination was busy.

The real cause is structural: the judgment needed to save a hard account was never transferred out of the founder's head and into something the team could run on their own. This is the part of the work that has nothing to do with the founder trying harder. It's building the playbook, the pricing guardrails, and the escalation criteria that let an account manager act on what she can already see three weeks out, instead of waiting for someone else's calendar to open up.