Are personas useful for product development?
One common tool in product design are personas. The idea behind a persona is to have a more detailed description of your target customer. This supposedly helps you in getting into your customers perspective. But are they useful as a tool?
What is a persona?
A persona is a made up character that should represent your customer. They have some level of personal details to them, to make them feel tangible. When we talk about personas, we usually talk about a buying persona. You could use the tool for other things as well and create a user persona, an employee persona etc.
For illustration, let’s create a buying persona for our services here at UTXO Solutions:
Name: Karl Müller-Schmidt
Role: Head of product management at “Medium Sized Incorporated”
Responsibility: Has to make sure, that the companies products are successful. Manages a small team of 6 product managers.
Struggles: The whole company didn’t invent something truly new in the last 20 years. Budgets are shrinking. His boss isn’t very open to experiments.
Technological environment: He has used ChatGPT a few times and knows TikTok from his son. He feels pretty up to date on methodologies, since he worked as a product owner for over 15 years. He also took part in a Design Thinking workshop recently.
Social environment: He is married to his wife since 13 years and they have two kids. A son (10) and a daughter (8). They live with their dog in a medium sized suburban house. He has one best friend, he does a lot with and a small group of people he meets from time to time. Most of his colleagues respect him, but he has a really annoying employee (Tim)
Goals: He want’s some great products to be developed to justify a promotion.
Quote: “Let’s get some great things done!”
With all this information, we could now develop our services. Always with a small voice in our head “What would Karl want?”
How good are personas?
Personas are kind of the go to tool for inexperienced product developers. They have the lowest level of abstraction.
You go from: “What do I want from this product?” to “What does someone else want?” In a way, personas are better then nothing. However, that’s roughly it.
Look at Karl again. Do you feel, like you are Karl? Probably not. I guess, that not even the person I had in mind, when describing Karl, would feel like Karl. So our product wouldn’t feel, like it is made for you. But why?
In my opinion, Personas fail in four different ways:
- Demographics don’t tell you anything about decisions
- No one is ever the standard customer
- The form doesn’t allow true data
- You can’t integrate them in first principle development
Demographics don’t tell you anything about decisions
We buy products, because we want to gain some progress, some improvement in our life. Personas don’t tell us anything about this. They may tell us, what goal our customer is trying to achieve. But they don’t tell us anything about his metrics for progress. When we want to develop something amazing, we need to know, how our customer decides, whether it is amazing.
Personas are mostly demographics. Fictional name, average age, job, family situation etc. But people are different. Statements made about someone because of their demographics usually offend them. It is like saying:
- “You are over 45, so you can’t understand new technology “
- “You are under 25, so you have a false sense of entitlement and a bad work ethic”
- “You are a woman, so family is more important to you then your career”
You can’t know how someone will behave, just from looking at him.
No one is ever the standard customer
Thinking in personas leads us in building for standard customers. Someone, we designed to be a certain way. But there is no Karl Müller-Schmidt like this out there. We would basically build for someone, who doesn’t exist. And we hope, that there are enough people, that are enough like Karl to buy our product.
But then I ask: Why not build for them? Why not build for people, that follow a pattern. Something, that is flexible enough, to account for individual differences?
You don’t want to build a product for your own needs, because you won’t buy it. So don’t build one for someone, that doesn’t exist.
The form doesn’t allow true data
Karl is obviously made up by me for this article. But most personas are build with at least some customer research. I have been part of persona projects. This is usually some big questionnaire. It has questions in it like:
- Some questions for the technological and social environment
- Other demographic questions
- A quote that describes you
What this means is, that you do a research for the persona. And then you claim, that there is research behind the persona. However, the data is not relevant. The relevant data is: How and why did you decide in the past in relevant situations?
But the persona is, as the name suggests, a person. What we are looking for is a behavior. We want to engineer our product for a decision process and a usage behavior. The form of a person simply doesn’t allow this.
What it does though is, that we infer the behavior from the persona. And this invites the whole spectrum of bias. We don’t take the information from the user on their behavior. We make the behavior up from what we think this kind of person would act like.
You can’t integrate them in first principle thinking
First principle thinking is a technique to reduce assumptions. When you build a product, you have decisions to make. What features should you build and what marketing messages should you use. The best results come, if there is a direct line of causality from the knowledge about your customer to the decisions. When every piece of knowledge directly leads to a new piece of knowledge. Until you know, what to build and how to market.
Personas don’t work for this, as you have at least two break points for the causality:
The first is in creating the persona. You aggregate the information from the user research into a persona. But there is no clear line of selecting the elements from the research into the persona. If you research behavior, you have to transfer it into a person(a). If you researched for a person(a), you didn’t research behavior at all.
The second is, when you transfer from persona to product. As written above: You infer behavior from demographics. This invites bias. And error.
So what’s the conclusio?
When you work with personas, you work with more assumptions then necessary. These assumptions look like data. But they are not. Demographics can’t tell you something about behavior. You build for a customer, that is made up and never there. The form invites for error. And it is practically unusable for a causality driven product development process.
Yeah, you could say, they are better then nothing. But I would even disagree to that. Having no solution and being aware of it, gives you a drive to find one. Personas satisfy the need for a better framework without solving your problem.