The Customer(s) Are Always Right

Survey and research your customers

The Customer(s) Are Always Right

I won’t argue that the customer is always right. Note the singular in that statement. I have had more than enough experience with unrealistic customers, and with egotistical employees. Blame often sits on both sides.

But I have long held a slightly adjusted belief. The customer base, or in the plural sense, the customers, are most definitely always right.

It’s a debate that never seems to go away. I saw a post on this phrase yesterday. I’ve written about it myself, once on putting customers first and again for founders, as well as in my best-selling book and framework, Customer Transformation. I think the concept has recently come back into focus because of the impact of AI and the removal of the customer.

What the phrase was for

The phrase started as a store policy, and it applied to customers as a whole. Its earliest known appearance in print dates to 1905 and is credited to Marshall Field and his Chicago department store. Harry Gordon Selfridge and John Wanamaker get credit for it too.

These were the first big retailers, serving thousands of shoppers a day. The rule told staff to take complaints seriously so shoppers never left feeling cheated. It set a default for how the store treated everyone who walked in.

People often quote a longer version, “the customer is always right in matters of taste.” It sounds good, but nobody has found a source for it.

So the original meaning was already closer to the plural. It described how a business treats its customers. Over a century, we turned it into an argument about whether one person at the counter is correct.

One customer can be wrong. Your customers can’t.

A single customer can be unrealistic. They ask for something you don’t sell, or want a refund on something they broke. I have dealt with plenty of them. Most recently, I heard about a hotel guest who asked to change rooms five times. Each for a different reason, until management finally suggested an alternative hotel. There was simply nothing the hotel could do or offer to satisfy that guest.

A single employee can be wrong too. It is easier to decide the customer is the problem than to fix what they are complaining about. This happens substantially more than it should, but that’s also part of the debate. Employees who don’t agree that the customer is always right often take on a negative attitude and create a poor experience for the customer.

Customers as a group are different. When hundreds of them ask for the same thing, complain about the same step, or leave at the same point, your market is talking. You can disagree with one person. You can’t disagree with what your customers keep doing.

They decide what they buy, what they renew, and what they tell other people. If they stop coming, they were right, whether you agreed with them or not. And this isn’t exclusive to B2C. B2B is equally responsible for ensuring it listens to and supports its customers. If you shove AI down their throat because it benefits you, they will leave.

How many customers can you afford to lose?

Let’s use a scenario. Say you make a decision that alienates 50% of your customer base.

Is 50% the right number? Maybe for you it’s more, maybe less. If you poll your customers and only 10% agree with your idea, do you move forward? What about 20%? Are you comfortable losing 30% of your market?

Two brands found out. Cracker Barrel changed its logo in August 2025 and at first described the feedback as positive. A week later, it reversed the decision, saying, “We said we would listen, and we have.” They just didn’t listen the first time.

In 2023, a Bud Light promotion set off a boycott. Weekly sales fell as much as 26%, and the brand lost the top-selling spot it had held for two decades. Its parent company’s North American revenue dropped $1.4 billion that year.

And yet most organizations poll customers with questions written to get the answers they want, to justify decisions they have already made. Listening to your customers means that, on some level, you also have to accept that they are right.

Why AI brought the debate back

It’s hard to agree or disagree with the customer when they are forced down a pipeline. The funnel looks unbiased, and they are stuck dealing with it either way. Granted, there are fewer egotistical employees. But no one has wiggle room to be “always right” or “unrealistic.”

The old argument assumed two people. A customer made their case, and an employee decided what to do about it. The employee could give in, hold the line, or bend a rule because the situation called for it.

An automated flow takes that decision away. Customers with unreasonable demands and customers with real exceptions get the same answer. Nobody is there to say “you’re right, let me fix that.”

The funnel only looks unbiased. The company built it, and the company decided in advance what every customer is allowed to be right about. Most of those decisions were made to lower cost.

In my experience, customers always remember the moments when an employee stepped outside the script to help them. A pipeline can’t do that.

What the customers are saying right now

If the customers are always right, then you have to listen to what they are saying about AI.

  • 39% of Americans say AI does more harm than good, up from 31% a year earlier (Bentley-Gallup, May 2026).
  • 27% trust businesses to use AI responsibly, down from 31%.
  • 61% of customers prefer a human agent (Verint, survey of 5,000 U.S. consumers).

Any one of those customers might be unrealistic. Together they are telling businesses what they want, and most businesses are building the pipeline anyway.

That is the part of the debate nobody is having. We argue about whether one difficult customer is right, while the whole customer base is telling us something and being ignored.

Five questions about your own customers

  1. When many customers complain about the same thing, who in your company hears it, and what happens next?
  2. Can a front-line employee make an exception without asking permission?
  3. In your automated flows, where can a customer disagree and reach a person?
  4. What have your customers, as a group, told you about dealing with AI? Did you ask them?
  5. When did your company last change something because the customers were right and you were wrong?