Design

Why personas in interaction design are a trap

Illustration by Jutta Kivilompolo

Personas are a common tool for interaction design. But are they really that useful? Reaktor’s Lead Designer Karri-Pekka Laakso reflects on why you should reconsider the whole idea.

A user of a service or an IT system is an abstract, ambiguous and elastic concept. It is prone to be interpreted in conflicting ways that suit each party’s needs. To make these “users” more realistic and tangible, we tend to use personas that describe people in more detail by giving them faces, names, job titles, hobbies, desires, and even life stories. 

The idea is that these unambiguous representatives of real people arouse our imagination. They aim to make our communication about the users more precise and effective: no more talking about “what do the users want”, but instead “Does Zaid actually need this? And surely this would not work for Anna?”

In addition, stepping in the role of a persona should help come up with design ideas through empathizing with users. This is the reason Alan Cooper invented them in the first place.

Since personas are a concept easy to understand, they are easily accepted as something we should do as part of interaction design. But are they actually that useful?

In general, I think not. With these four points, I’ll explain why.

1. Personas can distract you from what matters the most: design

Personas are clearly an easy-to-understand, lucrative concept, intuitively preferable to complex models and diagrams. When the basic idea of a persona has been stated, we automatically get it and are able to fill in the blanks. They are a universal concept with counterparts in marketing, scriptwriting, role-playing games, you name it. 

But this intuitive idea of personas is a trap. We get easily carried away with trying to describe our personas with increasing amounts of detail. It’s especially luring to start inventing all kinds of characteristics that seem necessary.  All too often it turns out that they have minimal or even no impact on our design.

Thus the first problem is that writing personas can easily become substitute work for doing the actual design. It may be fun, but personas are not the point. We need the design, and personas are only a means to get it.

2. We overestimate our skills to act in the role of other people

The central part of empathizing with users is to understand their behavior: What are they doing or trying to achieve? What are their motivations? Knowing this is key when evaluating or creating design. 

Pinning down behavior into personas isn’t at all as natural as the “properties” of the users, and thus this part gets easily skipped. And even if it’s addressed, it’s tempting to take some simple goal, stick it on the persona and feel like we have accomplished something: Issa is a gourmet enthusiast. The rest can be deducted from that. Let’s move on. 

But that is not true. Human behavior is complex. People are complex. We overestimate our skills to act in the role of other people. The deduction from a personal “property” to actual, realistic behavior is a bridge too far. Relying on our unverified assumptions and deductions leads to failure: “When you assume, you make an ass out of u and me” (unknown source).

Creating personas that contain behavior is not trivial. You see descriptions like “well-executed personas”, “rich characteristics” – like this, for exampleThose phrases indicate that writing good personas is hard, far from the intuitive character description.

This reminds me of use cases — remember them? They weren’t that bad, you just had to know how to write them correctly (lol). Most people wrote them in the way they found intuitive, which made use cases trivial, not very useful, and in the end a writing exercise nobody found worth the effort.

Personas are good for separating interests of different types of uses, but I can’t see them as an effective or natural way to describe motivations or behavior.

3. People’s behavior is situational, not tied to a single persona

Personas can easily lure us to oversimplify the world and our users. Personas, especially the ones that are not based on field research, are prone to turn out as hollow, one-dimensional caricatures: the stingy single parent, the culture-loving elderly lady, and so forth. In turn, they make us prone to lose the ability to see people, our users, as multifaceted as they are. 

Real people are situational. For grocery shopping, our user might be in a hurry today: they need something to eat for the kids and fast. But on Friday, the same person has time and is willing to buy something a bit special. This is one person, but easily two personas. Yet the two design personas with their name, image, and different hobbies hide this part of the reality from us.

At Reaktor, we recently published a study examining how people interact with everyday AI, such as recommendation algorithms. Turns out, AI keeps disappointing people – and even causing algorithmic fatigue – exactly because it doesn’t allow us to be situational.  Many AI services today are designed for fixed personas. You can read all about the negative implications of this kind of thinking here.

4. Do you put as much effort in validating the personas as you do creating them?

If we rely on personas, we should also validate them. We easily just pour our stereotypes into personas without ensuring that such people actually exist. Ever heard of the US Air Force average pilot? If not, read about it here. In short, the US Air Force tried to design their aircraft for an average pilot, but it turned out that a not single living instance of the average pilot existed. Not even one. 

These problems are all fixable. Perhaps the “you just have to do them correctly” applies to all design tools, making personas no worse than any other tool. Nevertheless, I have a hard time recommending them, though I appreciate Cooper’s original idea. 

So if not personas, then what?

I recommend doing field studies and basing your design on the findings. The whole package deserves its own blog post (perhaps even several of them), but in a nutshell: 

Field studies mean gathering knowledge by doing user observations, contextual interviews, and sometimes even trying to do the users’ tasks yourself. It is enlightening and time extremely well spent, even in domains you thought you knew. 

The leanest possible approach is to use this knowledge directly in design: design for the people and the tasks or cases you’ve seen. You don’t even need to name the people — the tasks are the most important thing. Based on your field visits you know that they really happen. They need to get done. Make sure your design helps in it. 

In the end, what we need is the interaction design. Personas and other intermediate tools are only a means to get to the design and thus transient — if we can do without them, we should. 

Passionate about breaking the boundaries in design? We are looking for a Lead Digital Product Designer.

Never miss a post