Three common mistakes to avoid in creating user personas

Hint: You’ve been doing it wrong. And here’s why.

This article has been adapted from its original publication in October 2016, for startup bootcamp ALPHA Camp Academy in Singapore.

If you’re reading this, you’ve probably encountered personas before — arguably the UX industry’s most infamous design deliverable. When used wisely, personas can be highly effective at not only helping your team empathize with your users, but more importantly, prioritize which features to build and pivot on quickly.

Unfortunately, a lot of businesses waste time and resources on superficial personas because they aren’t aware of the common pitfalls in the persona development process.

Here are three mistakes to avoid while creating personas for your business.

MISTAKE #1: Treating UX personas the same as traditional marketing personas

Using personas to generate empathy with users is already a well-understood reason for using them. (You are not your users, nor does the term “users” itself help you understand how different people will use your product differently.) This mistake happens when teams use personas to segment their market but by only using demographic data, a.k.a. traditional marketing personas.

UX personas and marketing personas may only look slightly different on the surface, but the difference is a big one for users of your product. This is because UX personas bring in the elements of the behavioral economics and psychology of how individuals and groups behave with specific products, services, environments, and situational circumstances.

While UX personas may borrow demographic information from the marketing industry, the primary focus of segmenting users isn’t based on what i call traditional “fixed identifiers” (i.e. demographics of age, gender, geography, income, and household size — anything you can tell about a person either by looking at a picture of them or by looking at their credentials on a piece of paper). Instead, the focus is on behavioral attributes that influence how and why people use a product or service. These behavioral attributes or “variables” are centered on characteristics that can be informed by, but are mostly agnostic of, those traditional fixed identifiers.

For comparison and examples, see the table below:

TABLE: Fixed/Demographic Identifiers vs. Behavioral Identifiers

The other key difference is in how consumer insights are generated. In traditional market research, people are usually asked questions through self-report surveys or in a focus group setting. While these methods are helpful in quantifying how people identify and in evaluating their attitudes, they don’t typically focus on studying specific behaviors or why people respond in certain ways. On the other hand, user-centered research usually begins with conducting qualitative research (e.g. open-ended, one-on-one interviews or observational field research) to uncover new or unspoken influences that inhibit, or enable, certain behaviors — much like how a therapist may ask questions to probe at a patient’s deeper subconscious.

SIDENOTE: I’m oversimplifying these approaches as differences, but the truth is, dynamics of today’s consumer marketplace increasingly require marketers to incorporate behavioral research into their work — and vice versa, for UX practitioners to validate their user findings with survey metrics. In fact, this is the exact dynamic that has grown into the field of data science. Anyway, my main takeaway for you here is that it’s critical for marketers and UX researchers to work together on identifying both market and behavioral segments in a cohesive way that fulfills both the business’s market goals and the end-users’ behavioral goals. Which brings me to the next commonly made mistake.

MISTAKE #2: Using the wrong research and/or validation methodology

This mistake is likely the number one reason why personas get a bad rap in the UX industry. Though the end product of personas may look alike, they are not all created equally. The process of how user insights were generated and how segments are defined are more critical than how the final deliverable looks.

Generally speaking, there are three types of personas defined as follows:

TABLE: Three types of personas and how they differ

Proto-personas. These are the simplest, quickest way to segment users but with a very big caveat. Generating user insights without conducting research effectively means you are limiting yourself to the imagination and intuition of your past experiences with (and biases of) people who are different from you. If you are a pessimist, this may mean user insights are completely falsified and have no merit — which leads to the commonly misperceived notion that personas are not a meaningful tool. If you are an optimist, this is an opportunity to flex your creativity muscles and tap your individual and team’s collective intuition for insights. Regardless, the main purpose is to use this process as an exercise in developing empathy for users and as an introduction to segmentation. (You can read more about proto-personas in this UX Magazine article by Jeff Gothelf on using them for the specific purpose of gaining internal alignment.)

BEST FOR: Novice entrepreneurs (and intrapreneurs) with traditional engineering, hard science backgrounds, or less experience in consumer markets to flex your empathy and creativity muscles. Also for teams in any field struggling to gain internal alignment on business and user goals.

Qualitative personas. Taking the time to conduct user research is the next step towards creating more meaningful personas. One of the most common methods for gathering user insights is conducting user interviews. This may seem like a daunting task for teams that don’t yet have a large user base to study, but remember that user research comes in many shapes and sizes — you just have to be resourceful with where to look for user insights.

Here are a few of my recommended tactics, user interviews aside:

• Search engine keyword analyses (i.e. Google Trends). Look for what terms and vocabulary people are using in your domain of interest. Each market has its own terminology and societal “rules” that influence what people are looking for and why.
• Digital ethnography / online field observations. Go to where the search engine results are leading people to, and observe what people are sharing, commenting, posting questions, and writing reviews about.
• Competitive UI/UX assessments. Study the design choices your competitors are making with their products and see if you can look for clues on what is or isn’t working well for them.
• “In real life” field observations wherever possible. Throughout history, humans have been making — and leaving behind — artifacts of behaviors. Like an archaeologist from the future, go into the “real world” and look for physical evidence of behaviors. These clues will point to invaluable insights about what is or isn’t working in existing contexts of use.

Talking to users one-on-one is still a necessary step, but I generally don’t recommend conducting interviews without first doing some of the above homework. The other benefit is you are also likely to uncover new places to find potential candidates for your user interviews. And when it’s time for you to talk to users, you can have a more productive conversation with them which further deepens your understanding of users’ realities.

BEST FOR: Early-stage teams in immature or new markets where marketplace viability is volatile or largely unproven and where you are still experimenting with potentially disruptive solutions — and you need to clearly communicate your user-focused strategy both internally and externally to investors.

Quantitative personas. In a nutshell, you start with the same user research and interview methods as with qualitative personas, but the game changer is being able to numerically prove (to the best of your ability) that the segments you’ve delineated are more than just arbitrary judgment calls. This is the ideal approach, though the preferred method of quantitative validation varies widely in practice.

Without getting overly mired in the details, here are a few examples spanning a range of quantitatively validation methods:

• Manually mapping out relationships between variables. There are many ways you can assign numerical values to measure or track your insights across users, with varying statistical significance depending on your sample size. Utilizing any logic framework is at least (perceived to be) better than none and can be done quickly by experienced UX practitioners. For a good example, see this article from Smashing Magazine by Mo Goltz.
• Deploying online surveys to validate each proposed segment. Before deciding which segments to use for your personas, you can test how much each resonates with actual people by deploying online surveys. The intent is that online surveys potentially provide larger sample sizes. It takes time and experience to craft well-designed surveys that minimize biases and the chances of faulty data being captured.
• Automatically generating statistically significant “clusters” to identify segments. Some teams may have the resources to deploy large surveys and potentially tie into large databases of high-quality customer data, such as in the enterprise and B2B industries. (It does require more technical expertise in statistical analysis, and you can read all about why and how in my former colleague Steve Mulder’s book, The User is Always Right.)

BEST FOR: Teams with some initial traction and looking to scale their business across multiple segments — and need to quickly prioritize features to deliver as well as potentially pivot their focus with different target segments. But note that YMMV depending on the resources you have to invest in good data science work.

Ultimately, the right type of persona for your team will depend on your need for market validation and available resources. Start with one of the methods above and evolve your personas as needed. Remember to set your team’s expectations of which category of personas you are creating and the pros and cons of each. Do this early and often to minimize any grumbling afterwards about the purpose and effectiveness — and to maximize the chances that the personas will be put to good use.

MISTAKE #3: Not correctly applying the personas after creating them

After all is said and done with the process of creating personas, don’t forget to put them to work for you. Quite frankly, if you do this correctly, it will save you and your team lots of headache and debate over which features to build first. Here are a few ideas to get started:

• Assess the business value of each persona segment. Find out how much users in each segment are willing to spend on your product, cost of acquisition, and any other metrics critical to the success of your business. This helps you make informed business decisions on how and for whom to pivot your product offering. (This may also be included in the validation process of quantitative personas as survey questions or conversion metrics.)
• Plan and prioritize features with the personas in mind. Assign each persona a priority level for each feature. For example, you can use a “high/medium/low” scale (see above image) or affix priority values based on the Kano Model of user expectations. This will help you quickly assess which features matter (or don’t matter) to the most valued personas.
• Revise and update segments as needed. If certain personas you thought were different are actually becoming too similar, you may want to combine them. Likewise, if certain personas start to exhibit more differentiated patterns, you may want to split them out. Go with what your gut (and the numbers) are telling you, and as always, try to validate any new insights you encounter along the way.
• Make your personas easily accessible to everyone. Print and post the personas at your desk, in meeting rooms, and wherever else decisions are being made. Your personas also shouldn’t be an eyesore since you will want to refer to them early and often. Give them names that are easy to remember and refer to in discussions.

And there you have it! Now it’s up to you to go forth in creating meaningful personas that will help your team make well-informed and user-centered decisions. There will undoubtedly be bumps along the way, but investing in and persisting with user-centered methods now will save you invaluable time and resources later. Good luck, and feel free to leave any questions or comments you have about personas below!

BACKSTORY: In 2016, I wrote this piece on personas both as an homage to my former UX colleagues who informed my understanding of user research — but more importantly, I wanted to distill and share what I’d learned through my own UX practice in working with countless (and tireless) early stage founders who held a lot of misconceptions about how to talk to users. By this time, I’d also applied and interviewed for enough UX researcher-designer positions to know that even hiring managers didn’t fully grasp the pitfalls and potential of user research not just as a grab-bag of individual methods but as a coordinated process. I suspect even now this is still largely the case, and I hope sharing these insights on the “UX of UX” will stand to benefit more companies, teams, and individual UX practitioners.

DISCLOSURE OF BIAS & PRIVILEGE: Please read here.