Author’s Note: A handbook on this topic will be published in late 2019. I’m thinking of calling it Assumptions Aside: Creating more bias-free understanding of the people you support. To get notification when it’s published, sign up for my kind-of-monthly newsletter, Indi in Your Inbox.
People seem to fall into a hole when writing personas, even when they’re doing it based on research: they use demographics to divide between segments and to represent thinking styles. This is a problem. So, to start, here’s a little quiz. Below are four personas representing the types of people community and technical colleges want to support better. Which of these descriptions helps you understand these people at a meaningful level?
A) A Better Life
Growing up, I saw the struggle my mom went through — finding a job, searching for places to live that we could afford, putting everything into her job for years only to be let go because her boss was paranoid. We lived with my grandparents after that, because Mama was a broken person. I wanted her to feel better, so I offered to do work around the house for grandma so that Mama wouldn’t feel so guilty about staying there. Then I got a job and my own place, and life happened. And as it often does, life got tough. I had to move back in with my grandparents. I was depressed, but my cousin was there for me. He said I could do better. So his encouragement and my experience with my mom made me determined to change things for my generation of the family. I don’t have a lot of time during the day, so I was happy to find out the classes I wanted to take are offered at night, when my family at home can take care of each other. My cousin is always checking in with me, inspiring me and keeping me confident. Right now I’ve been thinking about which degree to get — which one will have the most job prospects when I graduate. With a college degree I can get a good job. A good job means I’ll be able to afford a solid place for my family. We will have more opportunity. It will be different.
B) Kailee — 20 years old, Hendron Kentucky
Kailee is from a low income bracket. She is a single mother. The father left the area three years ago and does not contribute child support. Kailee lives in her aunt’s house, along with her unemployed mother. She contributes to the rent with the money she earns at her part time job at Walmart, which is 20 miles away. Kailee likes her co-workers at the store, but doesn’t like that it’s such a long drive. The toddler and everything else leave her little time, so she takes night classes when she can at the local community college. She also applied for the campus daycare option that is available for low-income students, and was recently accepted. She found out about the campus daycare in one of the newsletters that she received as a student. She loves kids and wants to open a daycare herself some day. She has been taking night classes for two years.
I want to make a difference for people. I’ve always been this way. In middle school, after I learned firsthand how energetic I felt when my blood sugar levels were in control, I was spreading the news, helping kids at school learn how to eat right. I encouraged friends who were getting stuck in life. I’m the kind of person who always notices when somebody is in dire straights. At my job at the convalescent home, I noticed one nurse always mouthing off to a particular patient, so I decided to report her because that patient didn’t deserve the nurse’s anger while trying to recover. A little toddler who was visiting his mom in the home looked sickly, so I asked about getting him to a doctor. And it turned out the food situation at home had gotten bad while his mom was out. I stick up for people. I know this job is a step on my way to a career that is focused on helping people. As soon as I had enough money to cover the tuition, I started taking classes at the local college so I could get my nurses license. And then when I get that, I’ll go for a specialized degree from Vanderbilt. Then I can work at just about any hospital or care facility that I want.
D) Anthony — 19 years old, Elizabethtown Kentucky
Anthony is from a low-income bracket. He knew college was the next step for him, but he didn’t know what to study. He saw a college counselor to help him decide what career to aim for. He is good at math so the counselor suggested bookkeeping. He has been taking classes toward this goal for a year, and he likes his professors. He gets along with his classmates and enjoys attending study sessions with them. But he’s gone to see the counselor again because bookkeeping doesn’t seem like it would let him be around people much for his job. He likes being with people and helping them out.
All four of the descriptions above are meant to bring the persona to life. But are they meaningful enough for you to imagine new approaches for the college system, unhampered by existing solutions?
“[Personas] clearly illustrate the problem space you’re dealing with and the people whose lives you’ll affect … to all stakeholders, including the design team.” Andrea Resmini
I urged people in a tweet to try rewriting their personas without reference to demographics. Demographics can cause assumptions, shortcuts in thinking, and subconscious stereotypes by team members.
What the tweet didn’t clarify was don’t throw out the persona itself — if it’s based on solid research. Keep the well-researched persona and replace those demographic descriptions with descriptions of the underlying reasoning. Reasoning rarely conforms to demographic lines. (I’ll show some exceptions below.)
The descriptions for Kailee and Anthony contain photos, ages, location, statements-of-fact, explanations, preferences, and experiences with the tools the college offers. In contrast, the descriptions for A Better Life and Casey feature inner thoughts, reactions, and guiding principles for their larger intents and purposes.
To actually bring a description to life, to actually develop empathy, you need the deeper, underlying reasoning behind the preferences and statements-of-fact. You need the reasoning, reactions, and guiding principles. A Better Life and Casey descriptions contain the deeper concepts. They are personas that resulted from one of the studies I did for the Kentucky Community & Technical College System.
The context of this particular study was understanding what goes through people’s minds as they pursue college from difficult circumstances, recruited from within a low-income bracket. Recruiting is where demographics are useful. (The photos for Kailee and Anthony are from the KCTCS website.)
My tweet started a good discussion on Twitter, as far as 140 character discussions go. I have captured much of it below and hope for additional comments and input here.
“Do you advocate using neutral names (like “Pat”) or just de-emphasizing gender? Names are helpful for communication.” Jared Faris
“Gender, at least, is awkward NOT to talk about in scenarios, unless gender-neutral pronouns someday become the norm.” Kim Goodwin
Casey was named Stand Up for Others in the KCTCS study, but for this essay I wanted to demonstrate that you could use a neutral name if you prefer. The gender neutral pronoun “they” is catching on in English, so you can experiment with that. Or you can write your descriptions in first person, which is what I do to facilitate the act of walking in someone’s shoes. And there are reasons for purposely adding a demographic like gender to your description, which I will mention below.
Statements-of-fact, preferences, and demographics frequently serve as distracting barriers. They kick off all kinds of subconscious reactions in team members minds. For instance, phrases such as “low-income,” “single mother,” and “good at math” mean something to you because of your own life experience, people you know, and things you’ve read. It takes extra mental effort to get out from under what you already think in response to these phrases. Additionally, the phrases represent what is important to your organization in terms of its direction. “How can we attract more low-income students? Or single mothers?” Your ideas will derive from the existing frame of thinking. It will limit your creativity.
“I agree that people put all sorts of irrelevant crap into their personas. It dilutes their credibility. … Demographics are seldom necessary when it comes to designing structure and flow.” Kim Goodwin
In the stories I gathered for the study, I noticed that no one described themselves as “low-income.” They didn’t dance around the topic of money, but their stories were about bigger aims than money itself: A solid future. More opportunity. If you know what’s important to the people, instead of what’s important to you and your organization, then you can dream up much more powerful support for them.
When you get caught up in demographics, other issues also arise. Kailee and Anthony both describe the same group. They exemplify the tendency I’ve seen for teams to make demographically different versions of the same thinking style. The thinking style they both represent is pursuing college in order to make a difference for other people, which is described above as Casey. A double representation of the same group, such as Kailee and Anthony, causes unnecessary consideration by teams during ideation and design. The demographics don’t require different offerings or features in the context of this college system. In fact, the participants in the study who catalyzed the patterns for Casey were different ages, ranging from 17 to 33. They were various genders and ethnicity.
Personas Enable Cognitive Empathy
There are a few more things that came up in discussion about the ideas in my tweet.
Don’t Ditch the Qualifiers
“You are suggesting … to hide away qualifiers. Anything that we articulate contains a degree of bias: if you need to combat it, it needs to be made visible … explicit conversational checkpoints that have to be discussed … defused / counteracted and used to increase the complexity of whatever is being produced.” Andrea Resmini
Andrea has a good point — part of the usefulness of personas is to force discussion around things that are often unnoticed or uncomfortable to address. I agree that these qualifiers are important in this role as conversational checkpoints. I wish to describe the qualifiers more explicitly in terms of the underlying thinking, rather than describing them with shorthand references to things like age. There are times when age might be necessary as a descriptor, but I hope teams will try harder to get at what’s driving the behavior.
I sometimes make a further suggestion to client teams who have years of experience working directly (via research) with the diversity of the people their organization supports. I suggest they abandon “persona” (a representation of a person) and replace it with “behavioral audience segment” (a representation of a group). (Note: I have begun calling these “thinking styles” to emphasize that a person can change to a different group based on context or experience.)This change allows those qualified teams to get away from names and photos. I don’t suggest this for everyone.
Note: “Behavioral audience segment” is the name I use, although there may be a better one. In its defense, Susan Weinschenk uses “behavioral science” to mean what I am trying to represent. And “audience segment” is a common way to express a group an organization is focused on.
Anti-Empathy on the Team
I have seen a kind of anti-empathy occur on some teams who demographically relate to one description and subconsciously dismiss the other descriptions.
“One client used a photo of a young blonde-haired woman. That persona would get dismissed as ‘The Blonde.’” Sophie Dennis
“I work 99% of my time in an international environment, and can tell you that there’s no end to personal, unintentional, sometimes deeply-rooted bias.” Andrea Resmini
I worked with the UX team at an airline over the course of 15 months. We did 10 studies and found patterns for day-of-travel personas. One of the day-of-travel personas we found was initially called The Grumbler. I was uncomfortable with this name, but agreed to adopt it because the team all thought it was memorable and descriptive. However, one day a team member dissed The Grumbler as a person who had a mal-adjusted attitude toward life. At that point I said to the team, “The name of this persona has to be one that the members of this group would be proud to use to describe themselves in the context of a travel day.” It has to be a description the people who are represented would embrace. So we went back over the transcripts and found a common way people described themselves: frustrated. We changed the name of this group to The Frustrated. It took some time for the stereotype of a grumbling, mal-adjusted person to go away. I kept telling team members that one day, on one trip, they themselves might suddenly become The Frustrated. In order to avoid thinking of yourself as superior or biased toward any persona, be careful with the names and descriptions. Make choices based on your team and stakeholders.
“Assumptions and bias exist in the team. Diverse teams is the solution.” Giles Colborne
“[personas] only work when everyone on the team has been in on the research.” Dana Chisnell
“[personas] work best when co-researched and co-created.” Jeff Gothelf
“Given that not everyone can conduct research, what is the best tool for sharing findings?” Peter Merholz
“Or if there is turnover in design teams — a good way to carry research insights forward?” Kevin Hoffman
“The best tool is exposure.” Jared Spool
Lack of Empathy on the Team
Sometimes lack of empathy takes other forms. For teams who have not been in contact with many people, or for teams whose own world does not mirror the diversity present in these people, then you’ll want to employ a trick.
The trick is to purposely add demographics to the description that are not common, in order to lift the team out of stereotypical assumptions that come along with using the demographics that have higher correlation.
“I use them to challenge assumptions.” Anne Gibson
“Use it to challenge assumptions. E.g. for low digital skills, use someone younger, not the lazy 60+ stereotype.” Sophie Dennis
“Demographics sure _imply_ values/attitudes/aptitudes. When chosen consciously, they’re great mythbusters for stereotypes.” Petteri Hiisilä
“Photos are badly done at least half the time … but the right set of photos can help the (often very male and white) team remember that the whole user base doesn’t look like what they expect.” Kim Goodwin
So for a persona representing a style of thinking that distrusts digital communications because of privacy issues, because it adds to the load of thing in their daily life, and because it just isn’t necessary, try choosing a 35-year-old mother of two young kids — so long as that demographic does exist. Her reasoning for not having an email account or social media is the same as certain 75-year-old males reasoning. While older people probably correlate more to this line of reasoning, using a younger person will help keep the team focused on the thinking, not the demographic.
Think Carefully About Exceptions
There are some instances where using a demographic is appropriate. These instances have to do with how people react or reason based on what they are seeing about a person — usually fraught with assumptions.
If you are studying the way people think when they are facing discrimination, then of course my tweet is completely inapplicable. Discrimination is based on perceiving these demographics, so the demographics cause the thinking when a person is in such a situation. (Context is key to your personas, and I write more about it later on.)
Speaking of age, it seems to be the qualifier that people most often cite as something that causes behavior or causes a style of thinking. In most cases this assumption is wrong — aging does bring different life events with it, but it doesn’t necessarily cause a change in underlying guiding principles, reasoning, or reactions. Experiences over decades might influence a change in guiding principles. Physiological changes in the body might cause different reactions and decision-making. But simply taking another trip around the sun does not. What’s at issue is the meaning of the word “age.”
“I think it’s curious that you would frame age as a personal characteristic. Age the number, or age the phase?” Yerbasuena
So often, age is stated numerically. I hope teams will change this to a description. Indeed, a “phase of life” might work, or a physiological or health description, or a life event. I like to work in the underlying reasoning wherever possible. Don’t ignore age, but don’t describe it using numbers.
“Except physical limitations are a huge factor. Example a 90 year old whose digital acuity is a big obstacle for using iphone.” Fellene Gaylord
“That has nothing to do with age. It has to do with physical ability. Many 90-year-olds can operate an iPhone fine.” Jared Spool
“True. But they will use it differently than a youngster with limited digital dexterity. Why not study that?” Fellene Gaylord
Describe what you’re seeing in your data with the philosophy, approach, intent. A person who is 25 with motor difficulty in their fingers has same problem hitting buttons. The aforementioned 35-year-old mother who doesn’t have a need for email or social media is a good example of guiding principles and reasoning.
“New parents vs. near-retirees have clearly different needs in a financial planning tool.” Kim Goodwin
The wording “new parents” and “near-retirees” are good descriptions to use in place of numbers when in the context of financial planning for the next two decades. A “new parent” and a “near-retiree” might both be 40 years old. What’s more influential in terms of supporting them with financial planning tools is their guiding principles and reactions. So taking it one level deeper, depending on the data patterns, you might end up with “near-retiree who confidently looks forward to the next two decades, with spending plans that allow for the things they want to do” versus “near-retiree who is worried about making financial ends meet” versus “near-retiree who intends to do a little belt-tightening and worries how it will affect lifestyle and happiness.”
There is a big exception for youth, at least in terms of understanding the underlying reasoning and guiding principles. In my experience, young people have a hard time explaining their thinking because many haven’t really developed or examined these underlying thoughts. The listening sessions I have conducted with people in their late teens have frequently met with difficulty in these regards, although not always. Deb Gelman probably has more to say on this, since my experience is fairly limited.
“Will anything you’ve included [in the persona description] reinforce unhelpful assumptions about this population? The person shopping for that minivan could be a dad just as easily as a mom.” Kim Goodwin
If you are about to say, “And most commonly actually is the dad,” then ask yourself why you want to say that. The correlation between gender and who actually does the shopping for the minivan is unimportant. What is more helpful to your organization is understanding how one set of thinking patterns shows, for example, the tendency to make trade-offs in a purchase decision (“if the price is right, then I will choose the minivan with the higher safety record” or “among all minivans with the highest safety records, given a certain price range, which one has the best maintenance record?”) versus the willingness to wait until all the criteria are met completely. So why bother with the gender correlation in this case?
Additionally, be careful about the context of your persona. If it’s about a purchase decision, that’s different than the use of a minivan, or even possibly different from the decision to get a car and to have that car be a minivan. (I talk about context and scoping a little later.)
There are a couple of exceptions to use gender, of course. Physiological differences in bodies in the realm of healthcare cause different behavior. There are health and illness correlations that run along gender lines. Hormones can influence actions and styles of thinking. Take this into account if your are creating personas for healthcare or something related to physiological aspects.
Gender in different cultures can affect behavior. If I were working with a client exploring a gender-divided culture, I would still want to dive into the guiding principles and reasoning behind it all. I would still want to describe it using those attributes. Depending on the team I’m working with, I might call attention to it with purposeful gender description. But the majority of the description would focus on the thinking style.
Again, it’s the definition of a word, “ethnicity,” that is at issue. If you define it as “pigmentation of skin, color of hair, and color and shape of eyes,” then it does not cause differences in behavior or thinking or guiding principles. (Unless you’re studying discrimination or something similar.) If you define it as “culture,” then probably. What does your data indicate? Concepts your parents teach you form your guiding principles. What your society approves and disapproves can affect your reasoning, reactions, and decision-making. So maybe for your context, differences are important to describe. But try to describe them in terms of guiding principles, reactions, and reasoning. Not everyone within a given culture thinks the same way.
When I see ethnicity highlighted in a persona, usually it is done shallowly, as skin-color, and is used as a head-nod toward diversity. What are your real reasons for highlighting skin color? You could use it if these reasons have roots in culture, affecting thinking in the context of your study. Or the reason could be because the team needs reminding that everyone doesn’t come from the same guiding principles and reasoning as themselves. Or maybe there’s a guideline in the protocol at your organization that requires it. There are reasons to mention it, but they are rare compared to the typical situations that I encounter.
I’ll see personas with city of residence listed. There might be reasons to list a location if proximity to the coast or mountains or desert influence behavior within the context of what your organization is trying to support. Or access to natural resources. Or tropical or freezing climates. Or natural hazards. It’s all relative to your context.
“I’d argue location is necessary to avoid defaulting to western / urban / class cultural assumptions, but can fall into stereotype easily if you’re not careful.” Barry Saunders
“I think … only include of affects behaviour or motivation. So e.g. urban v. rural could be relevant.” Sophie Dennis
“Urban/suburban/rural can matter, in terms of resources available.” Laura Creekmore
These examples would all be subject to the same metric: do they relate to your context? If you are supporting people who are growing vegetables or teaching class in an open air building in a rural area, yes perhaps the rural location would affect the thinking and approach of different personas. If you are supporting people who are tracking wild animals or filming a wilderness adventure, the environment would affect the thinking and decision-making. What’s the thinking when weather threatens to affect equipment or food supply? What’s the thinking on energy sources and recharging? There are all sorts of ways to approach these things. Additionally, the thinking and decision-making within one context, e.g. wilderness filming, would still form into different patterns representing different personas. The environment does not equal the persona. Reference the environment as you describe aspects of the reasoning, reactions, and guiding principles for each persona.
The Grand Exception
When doing this kind of research, I am cognizant of the fact that it’s a situation where an organization is seeking to support an audience segment. I have not had experience doing this kind of research in situations where it’s an artist — someone (like a chef or a jewelry maker, or perhaps even a news organization) is creating something for their own reasons. To create beauty, to thrill people, to get a message out. An audience forms who admires this created body of work. I think this situation is different enough from the support scenario that it merits calling out.
Empathizing without a Face?
Using a photo on a person requires that you choose demographics. But without a face, how can the team empathize? Empathy was one major concern that people voiced in the twitter discussion. Here’s a sampling of these voices:
“If you removed all those things, what would be left would not be a persona.” Charles Lambdin
“What’s left would no longer be a persona. Personas use specifics to generate empathy.” Jeff Chausse
“… seems hard to empathize with a list of attributes.” Mark Montri
“I see your point, but personas without gender and age will make it hard to empathize.” Jörg Linder
“But do you have a persona at all without such characteristics? An abstract bunch of data is harder to relate to. … personal characteristics in personas are a feature to enable us to relate to users.” John Wood
If you infer from my tweet that demographics would be replaced by lists of data, these concerns would indeed be valid. But the examples in this essay demonstrate that I am replacing demographics with the inner thinking which enables stronger development of empathy.
Cognitive empathy requires not a face, not preferences and demographics, but the underlying reasoning, reactions, and guiding principles. Without these you cannot develop empathy. And if you cannot develop empathy, you cannot wield it — you cannot walk in someone’s shoes. (I speak about the differences between and uses of emotional empathy and cognitive empathy in this Creative Mornings talk on Empathy.)
“Ok I think I’m seeing it now. Hard to do *and* keep the familiar/quick ref nature, but worth the effort.” Ant Miller
Context, Context, Context
“The point of personas is empathy, and you can’t empathize with a generality. Specificity matters.” Peter Merholz
Scope and context are paramount to your personas. A persona that attempts to represent every aspect of a person overall is pretty much useless.
This tweet is such a fun example of this that I had to screenshot it.
If you are exploring this teenager’s study habits and thinking that went into preparing for the last few maths tests she took, perhaps your philosophies might be similar. In terms of the context of studying, the scope of the last few maths tests, you might both be diligent about studying for several days before the test and feel slightly sick to your stomach in anticipation on test day. If your thinking matches along these lines, you’d both belong to the same persona.
Personas don’t work as generalizations; they need specific context. This is because personas aren’t meant for overall uses — they’re meant for ideation and designing solutions within a specific scope that your organization is concerned with this month or quarter.
When I worked with the airline team, we saw another example of this with our day-of-travel personas. When some people traveled on business, they belonged to one persona, but when they took their family on vacation, they switched to another persona. Some people switched personas after certain experiences. A person is human, and humans don’t remain exactly the same all the time, unswayed by mood or circumstances. Which might seem frustrating when you’re trying to capture the idea of someone in a persona … so that’s why you narrow down the context to a certain situation and focus the scope on what goes through peoples minds as they move toward a particular intent or purpose.
This means you will have several sets of personas within one organization. When I started working with them, the airline had an existing set of personas around the context of purchase decisions. We created the set of personas around day-of-travel. They were working on personas around deciding whether to take a flight or achieve the intent some other way. And we had enough data for me to create some personas within the context of checking bags.
If you can embrace the ambiguity that you will never be able to pin down humans in a complete, concrete way, then you’re ready for the mindset that you’ll always be following new scopes over time and creating new personas as you get budget and interest from your stakeholders.
“The issue with abstract data is while you avoid being too specific, you risk getting too vague. You get the amazing elastic user.” John Wood
This risk is valid, and it’s why we need the context, the scope, and the inner reasoning, reactions, and guiding principles.
Easier Than You Think
As you went over the examples in this essay, you might have had the sense that it would be fun to know a person this well. To talk to a person and get to this depth, as a part of your job, would be awesome. Or else you had a reaction along the lines that it would be a lot of work or take a lot of persuading.
It’s actually pretty easy.
It does require a different mindset. You need to slow down and try to understand what’s beneath what someone says by asking them. You can only get deeper by establishing trust and rapport. You also have to turn off the impulse to solve problems in your head as you hear about them. You have to let go of your product and your organization entirely because this isn’t about them, it’s about the person and what the person is aiming for in the greater scheme of things. And, all this is better to do by phone than in person. (In most cases.) You don’t need to see their artifacts or observe their behavior, because the knowledge you are after only exists in their minds.
This kind of information is collected in small studies, maybe 10–18 participants each. You don’t need to do it all the time — it is completely separate from your development cycle. You will never know everything about people, so you approach it like natural science and explore one little aspect at a time.
No, you cannot collect this kind of information via survey. There’s no way you could think up in advance all the possible ways people reason, react, and guide their decisions, in order to put them in checkbox format on a survey.
To learn how listening sessions are different than interviews (and more fun and easier), get a taste of it in this tour guide story or read my book, Practical Empathy. (I am also available to help get your personas back on track or to mentor your team through the process of collecting data for a new set of contextual personas.)