Yes, Personas Focus on Demographics (And How to Fix That)

The intent and design of personas contradict each other — despite what persona advocates say

Figure 1. (Left) The work by Georgia O’Keeffe is described as sexual, despite O’Keeffe’s protests that they are not. (Right) The painting “Nature forging a baby” seems violent, but is it?

“If one person tells you you’re drunk, and you feel fine, ignore him. If ten people tell you you’re drunk, go and have a lie down.”

This heuristic makes a great point: if only a few people claim you’re wrong, you can probably ignore them; however, if lots (or the majority) of people say you’re wrong, you should reevaluate your position.

Persona advocates should heed this advice when it comes to personas and demographics:

“If one person tells you personas focus on demographics, and you disagree, ignore him. If ten people tell you personas focus on demographics, go rethink personas.”

The debate of whether or not personas (which were invented some 25+ years ago) focus on demographics is a topic that just doesn’t go away — despite what is said by Alan Cooper (creator of personas) and Kim Goodwin (author and advocate of personas).

Figure 2. Goodwin, author of “Designing for the digital age” and proponent of personas, contests the idea that personas focus on demographics.

We should investigate this issue and do our best to resolve it. As professionals, we should aim to always improve our thinking, tools, and processes. And if a tool is broken, we should admit it, and make appropriate changes.

This article aims to resolve the “do personas focus on demographics” issue. It does so by:

1. Explaining that a persona makers’ intent is irrelevant when it comes to the impact personas have on others.
2. Showing how the way which data are presented to people affects how those data are interpreted.
3. Illustrating that, yes, personas do focus on demographic data.
4. Show how personas can be corrected so they do not focus on demographics.

And even if you don’t care much about personas or whether or not they focus on demographics, the concepts discussed in this article will help you become better at communicating and using data models with others.

The Intent Versus the Impact of Your Data Models

The first step in untangling this mess is to understand that data models have intents and impacts. They are not the same.

Figure 3. On the TV show “The Office”, there is a scene where character Michael Scott tries to diffuse sexual harassment; however, he unknowingly commits sexual harassment in the process.

The concept of intent vs impact is well illustrated on the comedy show The Office— specifically the episode titled Sexual Harassment. In one scene, the character Michael Scott observes that a coworker, Phyllis Vance, has been humiliated. To comfort her, he hugs her and talks about how attractive she is.

However, instead of helping her overcome her insecurities regarding her appearance, Scott actually amplifies her insecurities and makes her feel worse. Everyone at the office is similarly embarrassed for both Scott and Vance.

Yes, Scott’s intent was to console his coworker and fix the situation, but the impact of his actions made things worse. In his attempt to fight sexual harassment, he commits sexual harassment. The irony and humor is exacerbated when this is clear to all other characters in the scene, and Scott is oblivious.

The issue of misaligned intent vs impact isn’t restricted to sitcoms and HR polices. It happens all the the time in the art world. For example:

1. Georgia O’Keeffe is known for creating sexually allusive paintings (figure 1) even though she denied her work did so. None-the-less, O’Keeffe has come to be associated with feminism — an association she publicly contested and didn’t want (Robinson 1989).
2. When most people first look at the piece “Nature forging a baby” (figure 1), they think of it as a grotesque and violent. It seems like a woman is about to smash a baby! Having the floor littered with baby parts doesn’t help either. However, it’s an allegory to a “sweet and compassionate” nature (the woman) who always renews life (baby on the forge) despite what “Death and Corruption” do to her (Art Is Often Misinterpreted 2012).

The issue of intent vs impact certainly applies to anyone involved in design and research. That is, the intent you have as a researcher and designer is irrelevant when it comes to how your research and designs are interpreted by others. Sure, you might be trying to get coworkers to empathize with consumers — but in reality your models might be fueling disconnection and reinforcing prejudices.

With regard to personas then, it’s safe to say that persona advocates don’t intend personas to be interpreted as being demographic focused. We know this because they keep saying it. However, their intent is irrelevant. What matters is how people interpret it.

Before we show how personas focus on demographics, we need to understand how our brains process and interpret data. This will help you understand why personas cause us to focus on demographics.

Your Brain Loves Jumping to Conclusions

Figure 4. (Left) A “rational” decision maker would gather all the data first, then make a decision. Our minds, however, are always collecting data and forming conclusions (right).

Humans are not machines. Our thoughts are not algorithms. Our brains are not immutable data bases. We have behaviors that, when viewed from perspective of cold logic, seem “irrational” or “biased”.

One such example of seemingly irrational behavior is WYSIATI (What You See Is All There Is). This is a process where our brains jump to conclusions. Here’s a three minute video of Daniel Kahneman describing it:

Figure 5. In this video, Kahneman briefly explains how our minds are always forming conclusions — regardless if the data are incomplete or not.

The key is this:

If we [see some elements] of a story, we construct the best story we can out of those elements. And we’re not really fully aware of what we don’t know.

For example, if I tell you here is that leader of a nation and she’s “intelligent” and “strong”. Now, if I ask you at this point if she’s a good leader, you already have an answer: she’s a good leader. But now, the third word could be “corrupt”. So I haven’t told you anything about her character. You were not waiting. You took the information that you had and you made the best story possible out of it. That’s how our minds work.

The fact that our brains are always jumping to conclusions with incomplete data isn’t a new idea. You’ve no doubt heard about the importance of “first impressions” when meeting someone. As we gather data about someone, e.g. how they dress, their physical appearance, mannerisms, voice,… we form an opinion of them. Moreover this first impression, as it goes, can be hard to shake.

This happens because of what psychologists call the primacy effect. In short, it claims that data which are presented first to someone is often remembered best. For example, if someone reads off to you a list of 20 numbers, you’re most likely to remember the first few numbers, and forget the others.

When it comes to our data models then — first impressions count! You must be aware that what the viewer first sees about your data models will:

1. Be weighed the most
2. Determine how all subsequent data are framed, recalled, and understood

Now. With intent vs impact, WYSIATI, and primacy effect out of the way, it’s time to explain the relationship between personas and demographics.

Do Personas Contain Demographic Data?

Figure 6. A collection of personas from Cooper’s book “About Face”. Besides containing demographics data, these models reinforce prejudices and stereotypes. The the “cool” black guy drives a sexy sports car. The white, conservative soccer mom drives a minivan. The “redneck” drives a pickup truck with his dog in the back.

Before answering “do personas focus on demographics”, we need to get on the same page as to what are demographic data, and if personas contain them.

Technically, a demographic is a “quantifiable characteristic of a given population” (Ramachandran 2017). So, for example, data about me such as caucasian, with blond hair, blue eyes, and lives in New York are not demographics. This is because we’re just describing me.

However, if we said that some group of market researchers are caucasian , with blond hair, blue eyes, and live in New York, then these are demographics. This is because we’re describing a group of people (i.e. a population). If you have studied statistics, you can think of demographics as being similar to parameters.

In summary, the two criteria that define data as demographic are:

1. They are quantifiable
2. They describe a characteristic of a population

Now, let’s see if anything within a persona satisfy these criteria.

To avoid the claim that this article uses straw-man personas to criticize personas, let’s use personas which come directly from persona creator Alan Cooper and persona advocate Kim Goodwin.

Here is an example of a persona from Alan Cooper (Cooper 2004) and one from Kim Goodwin (Goodwin 2011).

Figure 7. A persona from Cooper’s “The Inmates are running the asylum” (left), and a persona from Goodwin’s “Designing for the digital age” (right).

Let’s stick with “Ethan”. Now, in that persona, are any of those data quantifiable? The answer is yes. They are:

1. Age (9 years old)
2. Gender (Male)
3. Ethnicity (caucasian)
4. Hair color (brown)

Before moving on to the next criteria, we should address a counter-argument persona advocates might make. They may claim that because these data are not represented by a number, then they are not demographic data. This argument is wrong. It is wrong because the criteria for demographic data isn’t that it is quantified (was counted), but that it can be counted (is quantifiable). “Age”, “gender”, and so on are attributes of a population that can be counted, and thus, are quantifiable. This makes them candidates for demographics.

Figure 8. Persona creator, Alan Cooper, claims that personas don’t contain demographic data and that demographics are antithetical to personas. He is wrong. Personas not only contain demographics data, but their design causes viewers to focus on demographics.

Now it’s time to test for the second criteria of demographic data: are these describing a characteristic of a population. To answer this, we’ll use a description of personas from About Face (Cooper 2007) :

Figure 9. In Cooper’s “About Face”, personas are meant to describe a group of people — not one specific person. This satisfies the second criteria for demographic data: the characteristics must describe group of people, not an individual.

In Cooper’s own words, these data are not describing attributes of an individual, but of a group. This satisfies the second criteria for demographic data.

To drive the point that demographic data are included in personas, I’ll rewrite Cooper’s persona without demographic data:

A child traveling alone for the first time, who wants to play games, games, and more games.

This persona eliminates age, gender, ethnicity, and hair color. Now, you might argue that “child” and “traveling alone” are demographic data. However, I’ll argue that “child” is qualitative and unmeasurable — there is no objective definition of what “child” means. And while I would call “traveling alone” situational data (because it’s not describing the user directly, but the situation the user is in), it can be counted and it can be interpreted as a characteristic. So, if you want to be that strict, you can write:

Someone who wants to play games, games, and more games.

Now, it’s time to show how personas focus on demographics

Do Personas Focus on Demographic Data?

Figure 10. The article “Creating Personas”, from Cooper.com, shows a picture of someone referencing a persona. Notice how the photograph (demographic data) takes up half the page. It’s a data model that overpowers everything else.

Proving that personas focus on demographics is pretty simple. Just ask: what data do you see first when looking at a persona? Well, you see a name and often a photo.

Names and photos are models of demographic data. A photo might not be a check list, but it conveys the same data. It could even be argued a photo conveys demographic data (age, gender, socio-economical status, ethnicity religion, nationality…) much better than a list would. The same is true of name. For example, compare the name Katie Bennett against Ahmad Abdul-Rahman Saqr al-Fadhli. Differences in age, gender, ethnicity, nationality, and religion can be inferred from those names.

Figure 11. Despite what Kim Goodwin says, the personas illustrated in her book focuses on demographic data. The first three things you see are: name, a picture (age, gender, ethnicity, hair color,…), and age.

Now, just having these data in these models isn’t enough to claim that they cause you to focus on them. However, when we combine the facts that these data are:

1. The first things presented by a persona
2. First processed by our brains (via WYSIATI and primacy effect)

We can safely infer that these data are weighed the most and used to frame subsequent data. In other words, your first impression of a persona are demographic data.

How to Fix It

Figure 12. Researcher Indi Young understands that personas contain demographics data, and that these data are degenerative to your design and innovation processes.

The fix is simple:

1. Expunge demographics entirely from personas (no name, ages, pictures…)
2. Put the demographic at the end of your persona
3. If you have a picture, make it small and/or include images of people with different genders, ethnicity, and ages
4. Don’t use personas

Figure 13. (Left) A persona created by Kim Goodwin. (Right) I fixed her persona. It removes most demographic data, and goals are moved from being the last thing to the first — this should cause you to focus on them more.

Persona Advocates Can Improve Personas If They Apply Cognitive Psychology and Use Empathy

Instead of putting so much of their time and energy into telling people that they are wrong and don’t understand personas, persona advocates should do some research and empathize with those who say personas do focus on demographics.

If they did, they’d learn — as I pointed out in this article — that it’s not people who misinterpret or misuse personas; rather, it’s a combination of how personas are designed and how our minds process data that cause us to focus on the demographic data contained within personas.

If persona advocates truly wanted to help persona users, they would fix personas, instead of just complaining for the last 25+ years that so many people keep using them incorrectly. If they do not take this action, we can look forward to both sides continuing to criticize each-other — which doesn’t help make design and innovation better.

Want to make progress with learning and applying JTBD theory?

• Download When Coffee and Kale Compete for free
• Buy it on paperback or kindle
• Learn about our public and private workshops
• Ask about our Insight Sprints: the fastest way to accelerate and spread knowledge, gather market insights, and develop an actionable plan based upon what you learn
• Or contact me directly.

Further Reading

Describing Personas

References

Cooper, A. (2004). The inmates are running the asylum:Why high-tech products drive us crazy and how to restore the sanity. Indianapolis: Sams.

Cooper, A., Reimann, R., & Cronin, D. (2007). About face 3: the essentials of interaction design. John Wiley & Sons.

Cooper.com. (2015) Creating personas. https://www.cooper.com/journal/2015/6/creating-personas

Goodwin, K. (2011). Designing for the digital age: How to create human-centered products and services. John Wiley & Sons.

Ramachandran, G. (2017). Impact of Demographic Variables on the Use Patterns of Electronic Information Resources Among Aerospace Scientists and Engineers of Bangalore. International Journal on Environmental Sciences, 8(1), 90–104.

Robinson, R. (1989). ‘I’M GOING TO LIVE A DIFFERENT LIFE’. From https://archive.nytimes.com/www.nytimes.com/books/99/04/11/nnp/robinson-okeeffe.html