Personas have a bad rap. When they come up in conversation, usually someone gives a mocking example of some version of the above. The underlying message is: How ridiculous. This persona has nothing to do with our actual users or our business.  

And it’s true—when poorly constructed, personas deserve the mockery they receive. Those fans of personas find this a shame because they genuinely have the potential to be a great tool for communicating user/customer insights. 

In the past few weeks, personas seem to be experiencing a renewed surge in interest, at least based on my LinkedIn feed. Let’s look at the current state of personas: Why did they fall out of vogue? Why are they back?

Personas are meant to be archetypal characters. These characters, called “personas,” should represent different types of users/customers who interact with a product or a service.

As with every research artifact, their usefulness is directly tied to their actionability: Can we use these personas to make business decisions? Design decisions? 

Sounds great! So what went wrong? The nuance disappeared: Perhaps cursed by their own relatability to actual individuals, personas sometimes end up being treated more as caricatures, with exaggerations or minimizations of their original appeal: 

I saw this happen at a company where I worked. People would refer to the personas by name in conversation:

Or 

The good news was that the personas had integrated themselves into company culture to the point that they were often referenced in discussion. The problem was the references were often incorrect. 

Further complicating matters, these statements could have been relevant to the discussion had they drawn from actual data used to construct the personas. But instead, the personas had been created mainly with anecdotal evidence from customers and were used to extrapolate non-existent opinions on situations or features. 

Even personas created with robust data might not make the cut—not because of how they were constructed, but because they weren’t adequately evangelized or updated over time as the business evolved.  

Maybe this is the wrong question. Maybe instead, we should ask: Why did they ever leave?

After all, personas are useful, if not in their literal form, at least as an example of how we can create empathy in understanding our users and customers. 

Here are some examples of recent LinkedIn mentions:

Efi Chatzopoulou pointed out that the debate shouldn’t be about whether or not we use personas as a tool at our companies. The discussion should rather be situated around: What are we doing to make sure we have a company-wide understanding of our users? 

If we reject personas as a synthesis and advocacy tool, what are we using instead? We might have found a better solution for our particular company culture. Whatever the case, let’s keep an eye on the outcome of this artifact, not the creation of a particular artifact over another. 

Selda Koydemir said something similar: “A persona is one of the many ways you communicate your research insights.”

Selda continues: “Unless your personas teach you and your stakeholders how your users make decisions or solve their problems, what motivates them, or what they are currently missing to achieve their tasks, they’re useless.”

We know the drill, researchers: Ensure your personas accomplish their intended goal.

I wasn’t convinced that personas had made a comeback until I started to look beyond the research community. Then I noticed a few interesting posts: Other professionals, including growth advisors and founders, are using ChatGPT to speed up the work of understanding users by asking the software to create personas based on existing data. 

There’s another reason personas are top of mind: Our new fascination with the potential applications of ChatGPT, combined with the endless quest to make doing research “easier” and “faster.”

For example, Katya Sivkova recently posted about “Using AI for your personas and Jobs to Be Done Research.” She provides a step-by-step guide to streamline the process with the help of ChatGPT. She suggested using a prompt to improve a framework she’s already considered, asking it to provide her client with clear instructions on research set-up, questions, and analysis, including grouping interview questions per topic to ease the interviewing process. Then she used Notably.ai to synthesize the resulting interview data. Finally, she returned to ChatGPT to request a “vivid” persona description. 

Another recent example was posted by Kwadwo “Kocho” Swiatecki Adu, who used ChatGPT to create a buyer persona and ads for a fictive headphone brand. His process is fascinating to read about: Using specific prompts, he had the AI do market research into popular Generation Alpha lifestyle brands, create a table that was populated in steps by the brands, their description, their level of popularity, their annual revenue, their website, Instagram URL, and the location of their headquarters. Then he asked for a sentiment analysis of specific Instagram posts to reverse engineer the persona the company is trying to reach. This information was also generated in a table. Finally, he asked ChatGPT to create personas with descriptions and create an Instagram ad for each persona. 

These are fascinating examples of how to use AI tools to speed up our work. But let’s keep in mind: the artifacts AI tools produce for us are only as good as the data that goes in. These examples show the tool’s power—but the power of critical thinking and sense-making remains in the researcher’s mind. This is especially true for artifacts like personas, which can have a powerful influence on how we make user and customer experience decisions at our companies.

Let’s bring back research-guided personas, absolutely! But researchers: this is your time to shine. Whether you’re creating personas from scratch or attempting to improve existing ones lacking in evidence-based creation or trying to produce this research artifact quicker with AI tools like ChatGPT, you have a vital role to play here. 

The challenge with any research artifact is dedicating time to create, update, and advocate for it. 

With creation, ensure the rigor of your research work—AI tools may speed the summary of relevant text, but it needs to be checked for accuracy. Ensure the insights used to create personas are robust and the personas reflect the reality of your users. If not, they can do more harm than good. Not only could they hurt your business, but they could also damage your reputation as a researcher. 

In terms of updating, make sure to revisit them—your stakeholders will know they aren’t seeing an accurate representation of what’s happening today. NN/g has a great article on viewing personas as living documents and how to design them to evolve

Finally, personas won’t work if they sit on a shelf after they are created. You will need to advocate for their inclusion in decision-making processes—better yet, to speed up inclusion, get your stakeholders involved from the beginning. Check out this article from NN/g about common pitfalls that cause personas to fail and how to avoid them.

Ready to get started? Dr. Nick Fine recorded an excellent talk at UX Live 2018 called Creating Data-Drive Personas that Perform, where he walks you through the process. This is an excellent place to begin if you’re ready to take on this challenge. Good luck!