Building a User Persona That Actually Changes Your Marketing

Quick question: which one of these two messages feels more personal?

OPTION #1:

Hello Sir/Madam, We hope you’re pleased with your recent purchase from our brand. Below are a selection of other products you may enjoy.

OPTION #2:

Hey Cathy, it’s 11pm on a Sunday and you’re standing in front of your closet again, trying to figure out what to wear tomorrow that doesn’t feel three years old. We get it. Here’s what just landed in the collection you were eyeing.

Obviously the second one wins. But notice why it wins. It’s not because we used her first name. Plenty of “personalized” emails do that and still feel like a form letter. It’s because it names a specific moment: 11pm, Sunday, standing in front of the closet, that particular flavor of “nothing feels right.” That’s the difference between writing to a demographic and writing to a person.

That distinction is the whole story of this post. We wrote about user personas back in 2023, and the bones of it still hold: you do need a clear picture of who you’re talking to. But the way good creative strategists actually build that picture has moved on, and honestly, so has the risk of getting it wrong now that AI can spit out a “persona” in ten seconds flat. So we’re updating it.

What a User Persona Actually Is

A user persona is a representation of the person who uses your product or service: built from real research, not guesswork, and specific enough that you could write an ad speaking directly to them.

The word “representation” is doing a lot of work there. A persona isn’t a real person. It’s a composite, built from patterns in real customer behavior, language, and reviews. The goal is a stand-in specific enough that when you sit down to write copy, you know exactly who you’re writing to and what’s happening in their life at the exact moment they need what you sell.

Most companies land somewhere between one and three personas. If you serve one tight niche, one might be all you need. If your products and customers vary widely, you may need several, each one earning its keep by describing a genuinely different buyer, not just a different age bracket of the same buyer.

User Persona vs. Buyer Persona

Worth a quick distinction, because it still trips people up: buyer personas represent the person who makes the purchase decision. User personas represent the person who actually uses the product day to day. Often they’re the same person. Sometimes they’re not. The classic case is B2B, where you might sell to a VP of HR but the people using your software every day are the HR coordinators on that team. Get this wrong and you end up designing an onboarding experience for an executive who will never open the app, while the actual daily user is left frustrated with something nobody built for them.

The Old Way: Fun Facts Dressed Up as Personas

Here’s where we’re being honest about our own past work. Our original post included two example personas, “Chatty Cathy” and “Techy Tom,” built the way most agencies have built personas for the last decade: age, income, education, a bulleted list of “frustrations,” a tidy little bio paragraph.

That approach isn’t wrong, exactly. It’s just incomplete in a way that quietly costs you money. Here’s the test: does the persona tell you anything you could actually write from?

• Fun fact: “She’s stressed about money.”
• Insight: “She just saw her rent renewal notice go up $200, three days before she has to decide whether to renew her gym membership.”

The fun fact is true of almost everyone. The insight tells you the exact moment to show up in her feed, the exact tension to name, and the tone to use, because “we know money’s tight right now, so here’s a plan that doesn’t ask you to choose” lands completely differently than a generic “get fit this summer!” ad.

A persona built entirely from demographics (age, income, “loves Taylor Swift”) gives you a costume. It doesn’t give you a decision. And every piece of research you do should change a decision: what you build, how you say it, who you target, or why. If it doesn’t move one of those, it’s not research yet. It’s just a fun fact with good formatting.

The Modern Approach: Build the Persona in Three Layers

Instead of starting with demographics, we start with customer reality: how this person actually experiences the problem your product solves, in their own words, before your product ever entered the picture. That reality gets built in three layers, done in order.

Layer 1: Generational context. What shaped this person before they ever encountered your brand? Someone who came of age during a recession relates to money differently than someone who didn’t. Someone who grew up scrolling has a different trust radar than someone who grew up with three TV channels. This layer isn’t about the product at all. It’s about understanding the water your customer swims in, so the language you use actually sounds like theirs.

Layer 2: The micro-moment. This is the layer most personas skip entirely, and it’s the one that matters most. Not “she wants to lose weight,” but “she’s getting dressed for lunch with a friend who just lost 40 pounds, and her favorite jeans don’t fit the same way anymore.” Not “he wants to grow his business,” but “he just told his best hire ‘no’ to a raise he knew she deserved, because the spreadsheet said he couldn’t afford it.” Specific, real, uncomfortable moments give you something to actually write toward: a scene, a line of dialogue, a feeling.

Layer 3: The emotion underneath it. For each moment, ask three quick questions: What’s the strongest feeling here? Is it something they’d only admit privately, or would they talk about it with friends? Did it hit all at once, or has it been building for years? The answers change your tone more than anything else. Someone who’s struggled with something privately for years doesn’t want “Join thousands just like you!” They want “You don’t have to figure this out alone.” Nearly the same offer, completely different message, and only one of them actually lands.

Where This Research Actually Lives

You don’t need a research department to do this well. Small businesses usually already have more of this material than they realize:

• Your own customer emails, support tickets, and reviews, especially the gap between your 5-star and 3-star reviews. That gap is where objections live.
• Your sales team or front-line staff: the people talking to customers every week already know the language, the hesitations, and the “why now” behind a purchase. Ask them.
• Reddit, niche Facebook groups, and industry forums: unfiltered, first-person “here’s what I’m dealing with” language, in your customer’s own words rather than a marketer’s paraphrase of it.
• Your own website and email analytics: where people actually spend time, what they click, and where they drop off tells you what’s on their mind, even if they never say it directly.

Building Personas With AI: Don’t Trust the Output, Trust the Receipts

Here’s the part that’s genuinely new since our last post. AI tools can now take a pile of reviews, forum threads, or support tickets and hand you back a fully-formed persona in minutes. That’s a real time-saver, but it comes with a catch worth taking seriously: AI will generate a confident-sounding persona whether or not your data actually supports it.

The fix isn’t to avoid AI. It’s to treat every AI-generated persona the way you’d treat a claim from a stranger: ask for proof before you build a strategy on top of it. Here are five checks worth running every time:

1. Ask for receipts. Have the AI pull three direct quotes from your actual source material for each persona it names, not a paraphrase, the real words. If it can’t produce real quotes, the persona was invented, not discovered.
2. Check the sample size. A persona built from four reviews or a $200 ad test isn’t a pattern, it’s a coincidence. Look for at least a couple dozen data points before you treat something as a real trend.
3. Get a second opinion. Paste the same raw material into a different AI tool and ask for the same analysis. If they land on the same core persona, that’s a good sign. If they disagree, your data is probably too thin or too mixed to build on yet.
4. Look for an actual trigger. “Cares about her health” is a demographic. “Just got a blood pressure reading that scared her at her annual physical” is a trigger. And only a trigger gives you something to write an ad around.
5. Watch for blending. If you sell more than one product, make sure the persona isn’t quietly averaging buyers of different things into one profile that describes nobody accurately. Your entry-level customer and your premium customer are often two different people with two different reasons for buying.

If a persona can’t survive those five checks, it’s not ready to hand to your creative team yet. It’s a draft. Think of AI as the distillation step, not the whole experiment: it can boil down a pile of raw customer language fast, but it’s still on you to make sure what came out the other end is real before you build a data-backed strategy on top of it.

Personas Are the Starting Point, Not the Whole Strategy

One more update worth naming honestly: a stack of well-built personas still won’t save a campaign on its own. Even a sharp, trigger-based persona can miss if the format doesn’t match where that person actually is in their decision. Someone who just became aware of a problem needs a very different kind of message than someone who’s already comparing options with a credit card in hand. Personas tell you who and what moment. Getting the format and the offer right for that moment is its own piece of the puzzle, and one we’re glad to walk through with you separately.

LeadFlask’s Collaborative Approach

We’ve helped a number of small businesses build personas that hold up: not just a document that sits in a folder, but something that actually changes what gets written and produced. We don’t know your business the way you do, so the process is collaborative: we bring the framework and the research discipline, you bring the real knowledge of your customers, and together we build something with actual receipts behind it. Real people, real work, on both sides of the table.

Chances are you already have more of the raw material than you think: a customer list, a review history, a sales team full of half-remembered conversations that are actually gold. Our job is to help you turn that into a persona specific enough to write from, verified enough to trust, and built to change real decisions instead of collecting dust.

Ready to build a persona backed by data instead of good guesses? Let’s set up a call and see what your customer research is already telling you.