Marketing Personas Are Almost Always a Boondoggle (but they don’t have to be)

Boss: “We need marketing personas for each of our customer targets.”
You: “Got it. What problems are these personas meant to solve?”
Boss: “You know, um… Strategy. Targeting. And, uh… give ’em fun names like Workin’ Wendy and Stay-at-Home Steve, or something.”
You: “Oh god, not again…”

A few years back, I worked on a 6-month project to collect survey, interview, and industry data that would build out a series of “high-fidelity personas,” for my previous company. A product research team of passionate, dedicated, smart folks met a few dozens times over those months, collected the research they needed, and produced a set of personas that were sent to product and marketing leaders. The personas were included in our company strategy and board of directors decks. Advertising copy was written for them, marketing campaigns were targeted at them, landing pages were built with them in mind, product initiatives were re-prioritized.

The whole thing was backwards.

The Process Should Have Been:

Which problems require better data about our customers? → What’s the best way to structure that? → If personas are the solution, what elements do we include that map to the problems we’re solving?

But, Instead It Was:

Build personas → Make sure they get used across the company → Create product and marketing initiatives to employ the personas → Get no benefit at all → Doh!

Customer Personas Meme

The customer archetypes generated in this process weren’t bad or wrong. Data-gathering was rigorous, intentions were superb, belief in the value of this work was near-universal. But, in retrospect, it was a solution searching for a problem.

That’s the first problem with personas—they’re often interesting, rarely actionable. Or worse, they suggest actions that could be taken, but don’t solve prioritized problems inside your organization.

And what about that second problem? Read on, friends.

Boss: “I want you to rewrite this marketing copy to better target our Stay-at-Home Steve persona.”
You: “OK… What do we know about Steve?
Boss: “He’s a 30-50 year old male, Bachelor’s degree, 2 kids under the age of 15, likes to feel that he’s part of a community, and doesn’t have a lot of time to weigh different options when shopping for household items. His representational quote is ‘I pride myself on my family’s health and happiness,’ got it?”
You: “Oh god, not again…”

If you search for “marketing personas” or “buyer personas” you’ll find loads of articles, illustrations, Powerpoint slides, and classic marketing “4Ps” type of stuff suggesting exactly what our fictional Boss described.

Here’s a few examples from top Google results to whet your appetite:

Corporate Chef Persona

Source: Smashing Magazine

Were the product designers for whom this persona was created planning to make hard-to-use software until the “Eureka!” moment they read this?

Persona Example from Neon Marketing


What’s the marketing course creator who reads this over going to do differently? Is knowing Maria’s imaginative, organized, focused on efficiency, and uses Facebook Groups going to help them in any way?

Sure, it’s possible these personas (and the many others you’ll find scrolling through Google’s search and image results) were indeed created to solve specific business problems. But, I doubt it.

My guess is they followed a few templates (the originals of which are lost to time), and as the classic telephone game played out over the past 50 years of persona development, we came to a place where marketers weren’t sure why they were doing the work, execs weren’t sure why they were assigning it, but someone, in some department somewhere, needed this year’s “updated personas” to complete a line item task, so they keep getting made.

But you’re reading SparkToro, so you’re savvier than that 😉

What Personas Often Contain:

  • Information with no business application
  • Demographics and psychographic traits that are hard-to-apply at best and stereotyping or biasing at worst
  • Data that’s aggregated based on informal interviews and guesswork alongside data from surveys or market research tools (with no indication of which ones came from which sources)
  • Idealized versions of singular, representative customer targets that narrows, rather than broadens, to whom a product or marketing campaign might appeal

What Personas Should Contain:

  • The information your team (or client) needs to effectively do their job
  • Details about that information’s source and accuracy
  • Enough breadth to represent the range of individuals covered by the persona
  • Segmentation of those personas only when there’s a relevant need to target, build-for, or market-to individuals in those groups differently

Beautifully, I’m talking about doing a smaller quantity of more meaningful work to build something more useful and less biasing.

Sadly, it’s not the same way marketing/buyer personas have been taught and done for the last half-century. And new ideas have a way of being tossed aside in favor of “that’s how it’s always been done.”

The upside is: when your competitors use archaic, hard-to-make-actionable, poorly-sourced personas and you don’t, you win.

What’s the Right Way to Build a Persona?

Aha! That’s the trick. There is no one right way, because there is no persona format that’s going to work well for more than a handful of organizations/applications. Almost every one should be custom-built to solve the problems you’re having. Not your competitor. Not the version you were taught in school or at that seminar. Not a replication of some template you found online.

In this respect, I like what Hubspot did with their online Make My Persona tool. It’s not perfect, by any means, and I’d have made plenty of different design decisions if it were my tool, but it avoids the biggest pitfall that every other persona tool and template I found fell into: one size fits all.

Hubspot’s tool is a step-by-step builder. It asks you some standard questions, and as you fill them in, it builds to a malleable visual that looks like this:

Hubspot's Persona Creator Tool

I tried to build a version that represented one of SparkToro’s personas, and ran into some bumps created by Hubspot’s concepts of what a persona should include. The defaults are more than a bit tech-software-centric, and left me presuming that Hubspot’s own, internal persona process is a big influence on the tool.

But, I respect that their model is designed for flexibility. You can delete sections. You can add custom-created new ones. You can include as much or as little detail as you like.

Is it as flexible as making your own in a Microsoft or Google Doc? No. Does it nudge you to make good decisions about backing up your persona details with data and methodology? Also no.

That said, it might help you build a representational, visual version of what you need. We all need starting points.

Remember that chef persona from the Smashing Magazine article? It was ostensibly made to help product designers craft better software for people like the alliterative corporate chef, Fred Fish. My guess is those designers had real questions a series of personas might have helped uncover. Questions like:

  • What specific software are our targets regularly using (so we can follow design conventions they’re familiar with)?
  • What tasks do they need to accomplish over and over, and which ones are rare, very rare, and almost never done (but still need to be possible)?
  • Why do they choose their current workflow or providers, and what would be the “Aha!” moment that would make them go through the painful switching process?
  • What media, social, and online sources do these folks regularly consume and pay attention-to (so we can do some product marketing and competitive market research in those places)?

Infuriatingly, the “Fred Fish” persona example solved none of these, at least not in direct, obvious ways. But a solution-oriented, question-addressing set of personas, crafted from smartly segmented data, probably could. What might that process look like?

How to Build An Actionable, Data-Credible Persona

If I’m tasked with customer segmentation work that requires breakdowns of various targets, their behaviors, demographics, opinions, and needs, personas (or something a lot like them) might be something I’d build. In that case, here’s how I’d do it (and how I suggest you might upgrade your own):

Step 1: Create a list of all the applications your personas are solving for. I.E. who’s doing what with the work you’re going to produce? What information do they need to get that done effectively?

Step 2: Audit your list for elements that are realistic, reasonable to product, which require data-gathering (and what kind), and which need to be provably correct and sourced vs. aggregated from interviews or opinion data.

Step 3: Collect the data you need. Usually that’s a combination of four things:

  1. Internal customer data from places like your email subscribers/buyers, web analytics, social accounts, etc.
  2. Statistical data about your market, either from your own surveys or a market research firm’s (in either case, make sure you’re closely auditing methodology and bias—it always exists, just account for as much as you can)
  3. Interviews with customers, hopefully with a representative panel and the same set of questions asked each time. From these, you likely need additional work to group responses in useful ways.
  4. Audience Intelligence data, usually from products that gather information passively, at scale. SparkToro obviously does this, but so too do tools like Brandwatch, SimilarWeb, and (each in different ways, with different outputs).

Step 4: Segment your personas based on the data that most accurately and clearly delineates one group from another. When there’s mostly-overlapping characteristics and behavior on the actionable to-do items among a set of your market, group them together. When those don’t overlap, that’s when you separate them into different personas.

Whatever you do, don’t segment based on demographics used to stereotype, pander, or bias. That’s how you end up with advertising like this:

Step 5: Determine a presentation format for your personas. That could be stylized cards with human imagery (like most of the personas you see online), or it could just be a simple text document with bullet points. You only need to follow tradition if it’s useful to the people who need your data. Or if it’s super-fun for you. Fun, in my book, is a totally valid reason to do something (so don’t let my derision of alliterative names or card-style layouts stop you from using them if you like it).

Once you’re finished, you should have something that looks like this:



I have a sneaking suspicion that maybe the data SparkToro produces about things like what an audience talks about, how they describe themselves, what they follow, listen-to, read, watch, etc. could be useful depending on the problems you’re trying to solve.

SparkToro Persona Data

SparkToro’s audience intel data: employ only when valuable.

But, if I created a standard persona template that included a bunch of our data points, and suggested those were better than every other persona template out there, I expect you’d cancel your email subscription to this blog. Rightfully so.

So instead, I’m recommending you roll your own, based on your needs, and nothing else. You got this 🙂

The short version of this post would go something like:

The standard “persona” is a representational model of the human beings we want to reach with our messages, marketing, or products. It has an alliterative name (VP Violet, Admin Adam, Engineer Emma), some broad demographic traits (gender, ethnicity, age, education, geography, etc), and a few inferred psychological attitudes or opinions (on-the-go professional, prefers high-convenience options, seeks out affordable luxury brands).

The standard persona should go suck on a lemon.

  • There’s no point in doing work unless it’s either really fun, or it makes progress toward your organization’s goals.
  • Personas have a long history, tradition, and all-too-often, a serious lack of usefulness . That’s why lots of people make ’em but very few get value from them.
  • You don’t have to be those people.
  • Instead, you can start with why you/your company/client needs segmented, representational models of your market.
  • Then, only build the elements you need into each of them.
  • Back each data element with its source and relative accuracy, so you don’t mislead and can support/refine/update conclusions.
  • Publish in whatever format works for you. If cute pictures and alliterative names are your catnip, go for it. And if you don’t like ’em, bullet points in a well-structured outline work just as well.

A bad persona can detract from the effectiveness of your product and marketing efforts. It can nudge you, your team, or your client to make bad decisions. Given the choice between no personas and bad personas, go with none. But, if you absolutely need to build them, now you’re armed with the tools to do it right.

Oh, and feel free to send your stuck-in-the-past colleagues to this page. I’m happy to take the flack.