“I'm in someone else's house and can't find the bathroom”

Let me tell you what happens when you do all of this “understanding your customer” stuff (to use a technical term), and do it correctly, and it still doesn’t matter.

Back in 2012 I did some consulting work for a major book and music publishing company that was about to relaunch its website. They’d spent the better part of a year on it and a considerable amount of money. They were proud of it. Really proud. The thing was nearly ready to ship, and they brought me in for what everyone assumed would be a friendly final once-over.

I asked if I could run a few moderated user interviews before they launched. They said sure.

A moderated user interview is a deceptively simple thing. You sit a real customer down in front of the product, you give them a realistic task (”Find a book about floral arrangements for Mother’s Day and order a copy.”), and then you do the single hardest thing in the entire discipline. You shut up, and you watch.

What I watched was painful.

The company’s navigation (the menus, the categories, the whole information architecture) was beautiful. Logical. Clean. It made perfect, intuitive, obvious sense.

To them.

They had built it, so they could move through it in their sleep, and they had assumed, the way every company assumes, that what felt obvious to them would feel obvious to a customer.

It did not. The customers I interviewed got lost. Repeatedly.

They looked for things where the things were not. They used words the company did not use. They expected the catalog to be organized in ways the company had not organized it. One customer, halfway through a basic task, said a sentence I have never forgotten: “I feel like I’m in someone else’s house and I can’t find the bathroom.”

I brought the company the recordings. I showed them, in their own customers’ words and their own customers’ confused clicking, exactly where the new site was going to fail and exactly how to fix it. The fix was not subtle and it wasn’t even especially expensive. It was mostly a matter of reorganizing and relabeling the site around how customers actually thought.

(If any SEOs out there are reading, yes I know restructuring information architecture is not as simple as a snap of the fingers. I have done that job. It’s still worth it.)

They were shocked with the results of these user interviews. And then they thanked me, paid me…and launched the site exactly as they had already built it.

I want to be fair to them, because they were not foolish people. They had a year and a budget sunk into that site. The launch date was set. Vendors were lined up. To absorb what I had shown them would have meant going back to the leadership team and saying “the thing we have been proud of for a year needs to be partly torn up.” That is a sunk cost, a year of work and money already spent, and as we established in the last chapter, the brain treats a sunk cost as something to protect rather than a loss to accept. So they protected it. We parted ways on entirely friendly terms.

About eighteen months later, out of idle curiosity, I went back and looked at their website.

They had completely retooled the navigation. The new structure was, give or take, spot-on what I had shown them in those interviews a year and a half earlier.

The signal had been sitting right there. They’d had it on video. The cost of acting on it before launch would have been a few uncomfortable weeks. The cost of acting on it eighteen months later, after a confusing launch, after a year and a half of customers getting lost, after a year and a half of soft sales nobody could quite explain, was vastly higher. The market had simply charged them late fees on a bill they could have paid early.

Finding the signal is only half the work. The other half is having the stomach to act on it once you have it. The signal doesn’t care whether you act. It will go on being true either way. The only thing you control is whether you pay early, on your terms, or late, on the market’s.

* * *

The persona on your conference room wall almost certainly describes a demographic. The customer you actually serve, the human being on the other end of the transaction, has a job they’re hiring you to do, a felt need driving the purchase, a moment shaping the decision, and a set of objections nobody on your team has ever directly asked them about.

That gap is where your competitor is winning the deals you should have closed.

A real persona tells you who to serve.

There is one more thing a real persona requires in order to be useful, and almost no persona document I have ever seen includes it.

It needs a negative.

A real persona is defined as much by who it is not as by who it is. The buyer above has a shadow somewhere out there: a customer who looks similar on paper but is actually the wrong fit. Maybe their estate is complex enough to require a real attorney. Maybe their family dynamics include things our product is not built to handle. Maybe they simply prefer an in-person experience and are going to resent every minute of an online flow.

That person is not our customer. Marketing to them is expensive. Selling to them is harder than necessary. Serving them, when they do buy, eats disproportionate support resources and produces disproportionate negative reviews.

They are the Anti-Avatar, I call them. Their existence is half of what makes the real persona useful, because the persona without its inverse is still a guess. You know who you are for. You also have to know who you are not for. Without both, you have a target. You don’t have a persona.

Naming them is half the work. Having the discipline to stop chasing them is the other half. We will get to that in the next chapter, where we raise the question almost no company ever asks itself out loud: who are you serving right now that you shouldn’t be?

Next week, we start to find them.