Your persona is lying to you

Somewhere in a shared drive, most businesses have a slide with a stock photo on it. Her name is Sarah. She’s 42, always gets three quotes before deciding, and “values honest, reliable service.” Nobody in the company has ever met her, because she’s a composite: a useful shorthand built from research, designed to give a scattered team one person to design around instead of an abstract idea like “customers.” That’s not a bad idea. It’s why most agencies, including us, still build them.

The problem isn’t that Sarah exists, or that you built her. The lie creeps in the moment everyone in the room starts treating her as fact instead of a starting hypothesis. A persona tells you who someone is in general. It doesn’t tell you what they were worried about the moment they actually landed on your site and closed the tab, and that gap doesn’t close itself just because nobody checked.

Real customers don’t behave in neat demographic boxes. Two people who match Sarah’s profile exactly, same age, same job title, same income, can behave in completely different ways on the same page, because behaviour isn’t driven by who someone is. It’s driven by what they’re anxious about right now. A persona tells you Sarah is 42. It tells you nothing about why she just closed the tab on your pricing page.

Empathy is not imagination. It’s attention.

Good UX people don’t get better at empathy by imagining harder, even with a well-built persona in front of them. They get better by paying closer attention to what’s actually happening: where people stall, what they type into search boxes, which line keeps showing up in enquiry emails, what question gets asked on every sales call before anyone will commit.

That’s the piece worth adding, not swapping in. The persona still earns its place. It gives a team a shared shorthand, a name to design around, a way to keep “the customer” from turning into an abstraction in a meeting. What it can’t do on its own is tell you where trust actually breaks down on a real page. For that you need evidence sitting alongside it.

Where that evidence actually comes from

None of this needs a new methodology. It’s a handful of things most businesses already have and rarely look at properly:

Analytics that show exactly where people drop off, not just that they dropped off. The words people actually type into your on-site search, which are often not the words in your own navigation. Support tickets and enquiry emails, which are basically a free transcript of every objection a customer has ever raised, in their own language. Sales call notes, especially the questions that come up right before someone says yes or right before they go quiet. Session recordings and heatmaps on your busiest pages, which show the scroll that stalls, the rage click, the hover that never turns into a click.

Put together and checked against the persona, this is what tells you whether Sarah’s profile is actually holding up, or whether it’s quietly gone stale.

What changes when you do this

Say your persona told you Sarah “values honest, reliable service.” That’s a fair, useful starting point. It sets a tone. What it doesn’t do on its own is tell a copywriter what sentence to write.

Now line it up against the evidence: 40% of people who land on your pricing page leave within eight seconds, and your enquiry emails are full of some version of “is this a one-off cost or ongoing.” That’s not a personality trait, and it doesn’t contradict Sarah either. It sharpens her. It turns “values honest, reliable service” into a specific, fixable gap: your pricing page doesn’t answer the one question she actually has before she’s finished reading it.

That’s the difference between a persona working alone and a persona working with evidence behind it. One gives you a tone of voice. The other tells you exactly which sentence is missing, and where it needs to go.

Try it alongside your next persona review

Pull the last twenty enquiry emails or contact form messages you received. Not just the good ones you’d use as a testimonial. All of them, including the confused ones and the ones that never converted. Read them back to back and write down every question that comes up more than twice.

Now hold that list up against your current persona. Does it still sound like Sarah? Then go and look at your website. Find the page where that recurring question should be answered. If it isn’t answered there, or isn’t answered until three scrolls too late, you’ve just found more useful, specific direction in twenty minutes than most persona workshops produce on their own.

That’s what empathy actually looks like on a website. Not choosing between a persona and the evidence. Using both: one to keep the team aligned on who they’re designing for, the other to catch it the moment Sarah starts lying to you and nobody’s noticed.