Who Is Your Cafe Customer? Build Simple Personas
Ask an owner who their customer is and the answer is almost always everyone. It is understandable, since a cafe genuinely does serve a wide range of people, but that answer leaves nothing to actually build a menu, an opening hour, a post, or a photo around. It is also, if you watch a shift closely, rarely true in the way it is meant.
Short answer: a persona for a cafe should be small, drawn from what an owner already sees at the counter, and it should end with a decision. Three short personas beat a long marketing template every time, provided each one closes with what the shop will do differently because of it. A persona that never changes a single decision is just decoration, and this post is a method for avoiding that, not a worksheet to fill out and forget.
I have sat through my share of marketing exercises that produced a beautifully detailed customer profile, a name, an age, a stock photo, a list of hobbies, none of which had anything to do with what happens behind a counter on a Tuesday morning. That kind of persona looks impressive and does nothing. The version worth building looks almost boring by comparison, three short entries built from things an owner already knows, each one ending in a specific, small decision the shop is going to make.
Method
Start with the last two weeks of counter memory and whatever the point of sale system shows, and list the visit types that show up. Most cafes see some version of the same handful, the commuter grabbing something fast on the way somewhere else, the remote worker settling in for a longer stretch, the parent navigating a stroller through the door, the student studying between classes, and the visitor who is not local and is trying the shop once.
Two weeks is a reasonable starting window because it is long enough to smooth out one unusually slow or unusually busy day, but short enough that the observation still reflects how the shop runs right now rather than some older version of itself from a year ago. A shop that just went through a menu change, a nearby construction project, or a seasonal shift might want to lean on a slightly longer or more recent window to make sure the pattern you are seeing is still accurate.
That list is usually longer than it needs to be for the purpose of building a persona, so the next step is picking the three that matter most to the shop, the three who show up most often, spend the most, or shape the busiest parts of the day. This is not about ranking every visit type the shop has ever seen. It is about narrowing down to the handful worth writing something specific about, since a list of eight barely used personas is really just a longer version of the same vague everyone answer this whole method is trying to move past.
For each of the three, write five things: when they come, what they order, what they might search for before or after visiting, and one change the shop will make because of them. That last piece is the one almost every persona exercise skips, and it is the only piece that matters. A persona description with no attached decision is trivia about your own customers, not a tool.
The search behavior piece deserves a small note of its own, since it is the one field on the list that involves a genuine guess rather than a direct observation. Nobody at the counter can see what a customer typed into their phone before walking in, so this field is necessarily an informed estimate based on how that visit type would plausibly find the shop, a quick near me search for someone in a hurry, a more specific phrase like quiet cafe with wifi for someone planning to stay a while. Treating this field as a reasonable guess rather than a measured fact keeps the whole persona honest about what is known and what is inferred.
None of this requires a survey, a focus group, or any tool beyond attention already being paid at the register. The counter memory an owner and staff carry around, who tends to come in when, what they tend to order, what they mention wanting, is first party observation, and it is more honest and more specific than any invented demographic profile could ever be.
It also helps to involve more than one person in this step if the shop has any staff at all. The person running the register at seven in the morning sees a completely different slice of the day than whoever closes, and each of them notices different things without realizing the other side of the day looks different at all. A short conversation, even just five minutes at a shift change, comparing notes on who tends to show up when, often surfaces a pattern no single person would have written down alone.
Writing the actual persona down does not need to take long, and it should not. A few sentences per persona, not a page, keeps the exercise from turning into the same bloated worksheet this method is trying to avoid. The goal is something an owner could glance at before making a decision, not something that gets written once and filed away.
Examples
Picture a downtown shop with a steady weekday commuter crowd. One persona might be built around someone who comes in between seven and eight in the morning, orders the same drink most days, is usually on their phone or in a hurry, and likely searched for something like coffee near me on their way to work the first time they found the shop. The decision attached to this persona might be as simple as making sure the fastest order path, a mobile order option or a clearly marked quick pickup spot, is genuinely fast, since speed is the actual thing this customer is optimizing for, not discovery of something new. Notice that nothing about this persona required knowing this person's job, age, or income, only their timing and their actual behavior at the counter.
A second persona at the same downtown shop might be the remote worker who arrives midmorning and stays for hours, ordering one drink and maybe a pastry, working from a laptop, likely having searched for something like coffee shop with wifi or quiet cafe to work from before ever walking in. The decision here might be about seating, making sure enough tables exist away from the espresso machine noise, or about outlet access, since this customer's actual need is a workable space, not speed. It is worth noticing how different this decision is from the commuter's, even though both customers are buying roughly the same drink. The commuter and the remote worker are not competing personas, they are simply asking for different things from the same shop at different times of day, and the physical space usually has to accommodate both without either one crowding out the other.
Now picture a different shop entirely, a neighborhood cafe that sees far more parents with strollers and far fewer commuters. One persona there might be built around a parent who visits midmorning between errands, often with a child in tow, ordering a drink and something to share, and who likely found the shop through a neighborhood recommendation rather than a search at all. The decision attached to this persona might be about physical space, whether a stroller can navigate the entrance and seating comfortably, something an owner who does not push a stroller through the door every day might never notice on their own, and something a five minute walkthrough of the space with that specific customer in mind would surface immediately.
A third persona at that same neighborhood shop might be the weekend visitor, someone not local, trying the shop once while exploring the area, who likely searched for something like best coffee near me or a specific neighborhood name paired with coffee. The decision attached to this one might be about signage or a single sentence on the website that tells a first time visitor what makes the shop worth stopping at, since this customer has no prior relationship with the shop to draw on the way a regular does, and has only a few seconds to decide whether to walk in at all.
Notice what these four examples share, even though they describe two very different shops. None of them include an age, an income bracket, or any other demographic guess. Each one is built entirely from something observable, when someone shows up, what they order, roughly what they might have searched for, and each one closes with a specific, physical, or practical change the shop can make. That is the whole shape of the method, repeated across whichever three visit types matter most at a given shop.
It is worth being honest that these two shops could easily overlap in real life. A downtown shop might see a version of the stroller parent during a weekend brunch rush even if weekdays skew heavily commuter. A neighborhood shop might pick up a small remote worker crowd if a coworking space opens nearby. The three personas an owner builds are not a permanent, unchangeable map of the shop's customers. They are a current best guess, built from recent observation, worth revisiting whenever the pattern at the counter seems to be shifting rather than just having one unusual week.
One more thing worth saying plainly: a persona is not a way of deciding who the shop wants to serve at the expense of everyone else. Building a persona around the weekday commuter does not mean turning away the occasional evening visitor who wanders in for the first time. It means having a clear enough picture of who shows up most, and why, that the shop's actual decisions, its hours, its layout, its menu, its photos, reflect reality instead of a vague guess. Everyone is still welcome, every single day. The shop just stops pretending it does not already know who is walking through the door most often.
A useful way to test whether a persona is doing its job is to ask, after writing it, whether it would change anything if it turned out to be slightly wrong. A persona built purely from a stock image and an invented backstory would not change anything either way, since nothing about it was ever tied to a real decision in the first place. A persona built from real counter observation, with one specific change attached, either gets acted on and tested against reality, or gets revised once it turns out the observation was off. Both outcomes are useful. Neither happens with a persona that only exists on paper.
This is the whole reason the method insists on ending each persona with a decision rather than stopping at a description. A description can be interesting to read and still change nothing about how the shop runs. A decision, even a small one, a mobile order option, a wider path for strollers, a sentence added to a homepage, is the thing that turns an hour of writing into an hour that affects the shop going forward.
Related Reading
More from Hillcane on this topic, plus the pages on the site that sit next to the work.
Frequently Asked Questions
Why build customer personas for a cafe instead of just saying everyone?
Because everyone gives an owner nothing to act on. A menu decision, an opening hour, a photo choice, a social post, all of these get easier once there is a specific, small set of real visit types in mind rather than a vague sense of the general public. A persona is not about excluding anyone who walks in, it is about having something concrete to design around.
How many personas should a cafe build?
Three is a reasonable target. Most cafes see more than three distinct visit types over any given stretch, but narrowing down to the three that matter most, by frequency, by spend, or by how much they shape the busiest parts of the day, keeps the exercise useful rather than sprawling. A list of eight or ten personas tends to get built once and never looked at again.
What should each persona include?
Five things: when this type of customer tends to come in, what they typically order, what they might search for before or after visiting, and one specific change the shop will make because of understanding them. That last piece is the one that matters. A persona with no attached decision is just a description, not a tool an owner can use.
Where does the information for a persona come from if there is no survey data?
From counter memory and the point of sale system, both of which already exist without needing anything new to be built or purchased. An owner and staff already carry a real sense of who tends to come in when and what they tend to order, and that first party observation is more honest and more specific than any invented demographic profile would be. No age, income, or spend statistic needs to be assumed or made up.
Is it okay if a persona is really about a neighborhood or market type rather than an individual?
Yes, and this is often more useful than an individual sketch. Depending on the shop's location, a persona might reflect a broader pattern, a campus crowd, a tourist season, a commuter corridor, rather than one imagined individual customer. What matters is that the pattern is real, observed at the shop, and specific enough to attach an actual decision to it.
What is the difference between a persona and positioning?
Positioning is about what the shop wants to be known for overall, its identity in the market. A persona is about who specifically walks through the door and what that means for a specific decision, a menu choice, a seating layout, a signage detail. They work together but answer different questions, and this post covers personas specifically, not the broader positioning question.
Should a persona include age, income, or other demographic details?
No, not as invented or assumed figures. This post does not recommend guessing at age ranges or income levels, since a cafe generally has no reliable way to measure those things about its actual customers without real data behind it. A persona built on what is observed, timing, order, likely search behavior, and the resulting decision, is more useful and more honest than one built on assumed demographics.
How often should a cafe revisit its personas?
There is no fixed schedule, but revisiting them whenever something in the shop's actual customer mix noticeably shifts, a new office building nearby changing the commuter pattern, a seasonal shift in visitor traffic, makes more sense than a rigid calendar. The point of building personas from real observation is that they should change when the observed reality changes, not stay fixed as a one time exercise.
What if a cafe's counter memory and point of sale data seem to disagree?
That disagreement is itself useful information worth paying attention to rather than a problem to resolve immediately. Counter memory can be shaped by whoever happens to be paying attention on a given shift, while point of sale data reflects actual transactions across all shifts. Looking at both together, rather than trusting one blindly, tends to produce a more accurate picture than either alone.
Can a cafe with very little foot traffic still build useful personas?
Yes, and it may take a bit longer to notice clear patterns simply because there are fewer visits to observe. Two weeks of counter memory might need to stretch to four or six weeks for a newer or slower shop to see a real pattern rather than noise from a handful of visits. The method itself does not change, just the amount of time needed to gather enough observation to trust the pattern.
Work with Hillcane
Ask an owner who their customer is and the answer is almost always everyone, which leaves nothing to actually build a menu, an hour, or a photo around.
If your cafe wants help turning counter observations into personas that actually change something, reach Hillcane at hillcane.co/contact or call (256) 384-2449.
Comments