What Your Loudest Customer Costs

6 tickets a week. 3 feature requests a month. A monthly executive call with a Director of Operations who escalated every dropped-character bug to a P1. At one company I saw, that account was 4% of total ARR and its share of CS and engineering attention was closer to 40%.

The math rarely gets more lopsided than that, and it gets that lopsided more often than customer success teams admit.

CS optimizes for the squeaky wheel by default. The squeaky wheel files tickets, requests calls, escalates issues to executives. The systems built to respond (SLA dashboards, escalation playbooks, executive review cycles) all fire when the customer makes noise. The quiet customer, the one using the product and paying on time, triggers none of them. Attention flows where the alerts flow.

The result is a structural mis-allocation. The customer who pays the most is rarely the one who escalates the most, and the one who escalates the most is rarely the one worth saving. A team spends its highest-value hours on the accounts that contribute the least. It is the mirror image of the concentration nobody wants to name: the accounts that matter most often get watched least.

The diagnostic is one query against the support system. Tickets per customer per quarter, plus a rough estimate of CS hours, divided by ARR. A handful of accounts show up with attention shares two to three times their revenue share. Most are not strategic. Most have learned only that the company responds to noise, the same lesson that ends with an escalation landing on the CEO's desk.

Two moves correct it. The first is structural: cap engineering's customer-specific feature work at a share of capacity that scales with the customer's revenue share. The account worth a twentieth of revenue does not get a third of the roadmap.

The second is uncomfortable: fire the customer. Not all of them. The one whose attention share runs five times their revenue share. The conversation is short. We're not the right vendor for you anymore. We'll support you through the contract term and help you transition. The customer pushes back. The CS team breathes for the first time in two years.

At one company I watched do this, the arithmetic afterward ran in one direction. At that account, the $80K of ARR walked out the door. So did 80 hours a month of CS time and two engineering quarters previously committed to custom requests. All of it got pointed at the quiet majority, the customers who had been subsidizing the noise. That is the same trade behind building the wrong product for your biggest customer: the loudest voice in the room is not the one paying the bills.

Track attention next to ARR. Put the two columns side by side and the accounts that cost more than they pay stop hiding. The loudest customer is a line item, not a verdict on your product, and once you can read the line item you can decide what it is worth keeping.