From Skeptic to Customer: Mapping the On-Page Journey

When people talk about the customer journey, they usually mean something that spans weeks: an ad, a visit, an email, a retargeting campaign, a purchase. But there’s a smaller journey inside that one that almost nobody maps, even though it’s where most of the losses happen. It’s the journey a single visitor takes down a single page, from the moment they land as a stranger to the moment they either act or leave. It takes maybe ninety seconds. And it has stops, just like the big journey does.

I started thinking about this as a state machine, because I’m a programmer and that’s the hammer I own. A visitor arrives in a state you could call oriented-or-lost. Within a few seconds they either know roughly what the page is about or they don’t, and if they don’t, the machine halts. Nothing further on the page matters, because nobody reads paragraph three of a page they can’t categorize. This is the first stop on the journey, and it’s astonishing how many pages fail here. Not because their product is confusing, but because the headline was written by someone who already knew what the product was, for readers who already knew too. The curse of knowledge, rendered in 48-point type.

Suppose the visitor makes it past orientation. The next state is something like relevant-or-not. They know what you do; now they’re asking whether you do it for people like them. This is where specificity pays. “Software for teams” keeps everyone in a fog. “Invoicing for freelance translators” instantly tells most people to leave and a few people to lean in, and the leaning in is worth far more than the fog. A page that tries to be relevant to everyone keeps every visitor in a superposition of maybe, and maybes don’t convert. They drift.

The third state is the interesting one: believing-or-doubting. The visitor understands the offer and considers it relevant, and now the skepticism engine turns on. Every claim on your page gets silently cross-examined. “Fastest on the market,” says the page. “Says who?” says the visitor. This is the stop where evidence has to appear, and the order matters more than the quantity. A specific concrete detail early beats a wall of logos late. If your product genuinely does something measurable, show the measurement. If real people use it, let one of them say something a real person would say. The cross-examination isn’t hostile, exactly. The visitor wants to believe; believing would solve their problem. They just refuse to be the only one doing the work.

Then comes the state most pages fumble at the very end: ready-but-hesitant. The visitor believes you. They’re considering acting. And here a new set of doubts appears, different in kind from the earlier ones. Not “is this good” but “what happens to me if I click.” Will I have to enter a credit card? Will I get spam? Can I undo this? How long will it take? These are small questions, but at the moment of action small questions have outsized weight, the way a pebble in your shoe matters more at mile ten. The pages that convert well answer these questions right next to the button, in small honest text. No card required. Takes two minutes. Cancel anytime, if it’s true.

The reason to map your page this way is that it turns a vague problem, “conversion is low,” into a locatable one: at which state does the machine halt? You can often tell just by reading your own page while role-playing each state honestly. The exercise is humbling. I’ve done it on my own pages for GazeSite and found halts I’d walked past a hundred times, and in fact the tool itself grew partly out of wanting this critique done systematically, by something that hasn’t seen the page before and can’t fill gaps with insider knowledge. A machine reading your page is usefully stupid in the same way a stranger is.

The deepest mistake this map exposes is impatience. Most failing pages aren’t missing information; they have it in the wrong order. They ask for the click while the visitor is still in the doubting state, or pile up testimonials on a visitor who is still lost. It’s like a salesperson who answers objections you haven’t raised and ignores the one you have. Each state has to be exited in sequence, and the only exit is through, never around.

There’s something almost respectful about designing a page this way. You’re not trying to trick anyone or rush them. You’re acknowledging that a stranger owes you nothing, that their skepticism is rational, and that your job is to resolve it step by step. Do that, and the click at the end stops feeling like a conquest. It feels like the natural last step of a journey where every stop was handled. Which, on a good page, is exactly what it is.