At 20 People, Everyone Talks to the Customer

In a small company, nobody has to be told to stay close to the customer. It happens on its own. The founder is answering support tickets at night. The person who built the feature is on the call when a customer complains about it. Sales, product, and the CEO are often the same three people, and they all heard the same frustrated voicemail. Customer understanding isn't a program. It's just in the air, because the company is small enough that the customer's voice reaches everyone without anyone trying.

That closeness is usually the reason the company worked in the first place. You built something people wanted because you could feel, in real time and without a filter, what they wanted and where you were letting them down. The loop between "a customer said something" and "we changed something" was one conversation long.

Then you grow, and you do the sensible thing. You professionalize. You hire a support team so the engineers can build instead of firefighting. You add customer success to manage relationships, product managers to "own" the customer, analysts to make sense of it all. Every one of those hires is reasonable. Every one of them is also a layer of insulation between the company and the person it serves. And here is the quiet cruelty of how it stacks up: the more senior you get, the further you sit from the customer. The people making the biggest, most expensive decisions about customers end up being the people who talk to them least.

What fills the gap is abstraction. You stop hearing customers and start reading about them. The raw, specific, sometimes infuriating voice of an actual person becomes a line on a dashboard, an NPS score, a persona, a ticket volume, a quarterly survey rollup. These things are useful. They also present a false sense of closeness. Everything surprising, inconvenient, and texture-rich about a real customer gets averaged out on its way up the stack, and what lands on the leadership team's slide is the smooth, confident summary of something that was neither. You begin to manage the representation of the customer instead of the customer.

From there the decisions change in a way nobody notices. You start making calls about customers in rooms that contain no customers, informed by summaries of summaries, and you mistake the confidence of good-looking data for the knowledge you used to have. Decisions that used to be made with customers get made about them. And the company slowly becomes certain it understands people it has not actually listened to in a long time.

That certainty is the dangerous part, and there's a well-known number that shows how dangerous. When Bain surveyed 362 companies, 80% of them believed they were delivering a superior experience to their customers. When they asked those companies' customers, 8% agreed (Bain & Company). The gap between 80 and 8 is not a rounding error. It is the sound of a lot of confident leadership teams who have lost touch with the people they serve and have no idea it happened, because every internal signal kept telling them they were fine.

The cost isn't only that you misjudge what to build. It's that you lose the early-warning system that closeness used to give you for free. When you were small, you could feel a shift in the market before it showed up anywhere, because you were in constant unfiltered contact with the people living it. At scale, you find out from the dashboard, and the dashboard is a lagging indicator. By the time a problem is big enough to move a quarterly metric, it has been real for months, and some smaller competitor who never stopped talking to customers has already moved on it.

None of this means you shouldn't professionalize or specialize. You have to. It means distance from the customer is the default, and like most defaults at scale, it grows unless someone fights it on purpose. The fight isn't complicated, but it has to be deliberate, and it has to include the senior people, not just the front line.

Keep the people who make decisions in direct, regular, unmediated contact with customers. Not a quarterly readout. Actual contact. Put executives on support rotations. Have product leaders sit in on sales calls and watch real people struggle with the real product. Read the raw verbatims, not just the sentiment score. Refuse to let the customer relationship be fully outsourced to the one team whose job it technically is, because the moment the customer becomes "their" responsibility, everyone else quietly stops listening. Use the dashboards the way a pilot uses instruments, to supplement what you can see, not as an excuse to stop looking out the window. And treat direct customer exposure as a requirement of senior roles, not a chore you graduated out of on the way up.

Companies rarely lose touch with their customers in a single bad decision. It happens the way these things always happen as you scale: one reasonable layer at a time, each one defensible, until the people deciding what to build next haven't spoken to a customer in a year and don't feel the absence. The work is to notice the distance before the customers do.

So it's worth asking plainly: how many layers now sit between the people making your most important decisions and the people those decisions are actually for? And when did any of them last hear a customer say it in their own words?