Your Persona Has a Spotify Playlist. Your Customer Doesn't.
Somewhere in your organisation, there is a slide deck. In that slide deck, there are between four and eight illustrated people. They have names. They have ages, household incomes, and deeply held values. One of them “seeks authentic experiences." Another is "time-poor but quality-conscious." And at least one of them, inexplicably, has a Spotify playlist or a favourite coffee order.
Her name might be Mindful Mia.
Mia is 34. She shops sustainably, but not so sustainably that she won't buy things. She values her wellbeing, her morning routine, and brands that "get her." She has strong opinions about oat milk and will always share them in Instagram. She does not, technically, exist.
Nobody in the room has questioned why Mia needs a Spotify playlist. Nobody has asked who decided she listens to Fleetwood Mac. And nobody has said what everyone is quietly thinking: I have absolutely no idea if any of this is true.
Welcome to customer personas. The tool that promised to make businesses more customer-focused and instead gave marketing and CX teams a very detailed way of talking to, and about, themselves.
What personas actually are
The official story: personas are research-based archetypes. Composite portraits built from real insight that help teams keep the customer at the centre of their thinking.
Here is what they are in practice: a slightly more sophisticated version of whoever ran the workshops.
Study any set of personas carefully and you will notice the pattern. The ones described as "savvy" or "engaged" or "values-driven" tend to share the professional background, approximate income bracket, and general worldview of the people in the agency or strategy team who created them. The ones described as needing "more support" or "not digitally confident" tend to be older, less affluent, or from somewhere outside a major city.
This is not a coincidence. It is confirmation bias doing what it always does: making the world look exactly the way you already expected it to.
Once Mia exists, you start seeing her everywhere. A customer calls in to complain and someone says "that sounds very Mia." A focus group participant mentions sustainability and a researcher notes "strong alignment with the Mindful Mia persona." The persona stops being a hypothesis about customers. It becomes the lens through which all customer behaviour is interpreted.
You built Mia from your assumptions. Mia now confirms your assumptions. You have created an expensive mirror with good taste in music.
Why Mia never dies
Most business tools get retired when they stop working. Mia becomes an institutional fixture. She gets printed on the wall. She gets laminated. In extreme cases, and these cases exist, she gets a birthday cake.
Partly this is practical. A proper persona project costs real money. Nobody wants to write off the budget that produced Mia, her playlist, and her complicated relationship with fast fashion.
Partly it is psychological. Mia gives the organisation something raw data cannot: someone to talk about. "What would Mia think?" is a much easier question to answer in a meeting than "what does the behavioural data suggest about our median customer's decision-making process at point of purchase?" Mia fits in a slide. The truth does not always fit in a slide.
And so Mia stays. Long after the research that created her is out of date. Long after the market has shifted. Long after several people have privately noticed that the customers they actually speak to bear very little resemblance to Mia at all. She has become, to use a term Mia herself would appreciate, a permanent fixture.
What behaviour tells you that Mia can't
Instead of asking what a fictional person might do, you look at what real people actually did. You look at patterns: the sequence of actions before a purchase, a cancellation, a complaint, a referral.
Behavioural data is less flattering than personas. It does not produce illustrated characters with strong brunch opinions. It produces things like: customers who describe themselves as loyal in post-purchase surveys then buy from a competitor within 60 days. Or: customers who contact support within 48 hours of purchase are three times more likely to churn, regardless of whether the issue was resolved. Or: the customers spending the most per transaction have the lowest satisfaction scores, which should probably worry you more than it currently does.
None of this makes a good wall print. All of it is true.
Research by Kahneman and Tversky established that the first number a person encounters shapes every subsequent numerical judgement they make, even when that first number was entirely arbitrary. This is called anchoring, and it means the same customer will evaluate your price completely differently depending on what they saw before they arrived at it. Mia has no opinion on anchoring. Your behavioural data does.
What to do this week
Pull up your persona deck. For each persona, write down the three most important claims it makes about customer behaviour. Then ask: what is the actual evidence, and how old is it?
Find one decision your team made in the last quarter that was justified with reference to a persona. Ask whether the actual customer outcome matched the prediction.
The next time someone in a meeting says "our customer wouldn't do that," ask how they know. Not to be difficult. To establish whether it is evidence or Mia.
Mia is not your customer. She never was. Your customer is whoever actually bought something from you last Tuesday, for reasons they probably couldn't fully articulate, in a context you didn't design for and couldn't have predicted.
That is harder to put on a wall. It is also the only version of customer understanding that has any chance of making you better at serving the people who actually give you money, rather than the ones who listen to Fleetwood Mac and care about your brand values.