Quantifying Proto-Personas: Enhancing Your User Insights

In user experience (UX) design and product development, understanding your target audience is paramount. A key tool in this process is the development of proto-personas.

Let’s explore these concepts and how they can transform your approach to human-centered design.

What is a Proto-Persona?

Proto-personas explain the “why” behind the behaviors of our patients, healthcare providers, or other user/buyer data. Proto-personas help us to understand more about the lived experience of a particular group of people. They give them personalities and explain their actions. They are both defined by behaviors, attitudes, and journeys—not demographics like age or gender. Theyexplain the “why” behind quantitative behavior data (such as Google Analytics data) you might observe in your product, website, app, or service.

You might call this a segment, cohort, archetype, or even something else! Proto-persona is a term that we use to describe our specific process/framework of creating an understanding of human behaviors, attitudes, and journeys. 

The “proto” specifically speaks to the fact that these are built using qualitative data only. We consider them to become true “personas” once they have been validated through a quantitative study.

Why Quantify Proto-Personas?

While proto-personas are incredibly useful for getting started, they are inherently subjective—more art than science, we always say.

Though they are based on qualitative data which is valid in its own right, there may be a desire to validate the groupings uncovered at a larger scale.

For some companies, this might not be necessary, but quantitative data can help you create an even more precise and nuanced understanding of your audience and its scale, allowing you to design more effective and tailored user experiences.

How to Quantify Proto-Personas.

Quantifying proto-personas involves several steps:

1. Collect data through a quantitative survey.

• Write a survey.
  ‍
  • When creating your proto-personas, you defined several attributes that differentiate them. These act as the starting point for your survey.
    ‍
  • Next, you want to add questions that are relevant to your study. If you’re focused on building a new product for patients, you can ask about comfortability with technology or AI, for example. If you’re exploring messaging, you can probe about their priorities. Whatever questions you add should complement the remainder of your qualitative research.
    ‍
  • We love using Likert scales to evaluate how strongly held their beliefs are. Pairing a strong statement like, “I will never trust AI with my healthcare information” with a scale from Strongly Disagree to Strongly Agree will show extremes that differentiate your personas.
    ‍
• Define your quotas.
  ‍
  • Screening quotas are used to make sure you include a representative proportion of participants. It’s important to include demographic quotas that roughly match your target audience, but you can also include a quota on a persona attribute.
    ‍
  • If it’s most important to get a broader understanding of people with different attitudes, including a quota on a specific proto-persona attribute will make sure those opinions are heard.
    ‍
  • On the other hand, if it’s more important to understand the scale of each persona, stick with quotas on demographics.
    ‍

• Use a surveying tool, such as Qualtrics, SurveyMonkey, or Pollfish.

One trick we used when defining personas for our AI & Us consumer research report was to use an existing, third-party data point that was relevant to our persona attributes.

(You can download that free research report here, btw.)

In this case, we used Visual Capitalist's global study on AI Sentiment. To capture a relevant proportion of attitudes, we based our quotas on the results of their question, “Do you think Artificial Intelligence (AI) will mostly help or mostly harm people in your country in the next 20 years?”

A rule of thumb is to collect at least 200 responses for each group you think exists to ensure your sample is statistically significant. So if you defined 5 proto-personas, you’ll want a sample size of at least 1,000.

2. Analyze the data with machine learning.

You have proto-personas at this point. Use this process to validate them.

• To analyze your survey results, you first want to use a clustering algorithm to define the different groups within your data. These will likely be similar to the proto-personas you’ve already identified, but you may find new ones or realize that some can be combined.
  ‍
  • Using a clustering algorithm like BigML, import the survey data for your key proto-persona attributes. For best results, we recommend that your attributes include a mix of behaviors, priorities, and values.
    ‍
• After you’ve used machine learning to create your clusters, use them to create crosstabs with each cluster and look for areas where the cluster is over- and under-indexed against the average. This will show you where your personas are truly unique.
  ‍
• After finding the over- and under-indexed traits, you can start ‘Qualitative Encoding.’ It’s similar to affinity mapping, where you’re grouping similar traits to find themes within each persona.
  
  For example, one cluster may have the standout attributes of “very health conscious”—health and wellness is their biggest interest, and they are always looking for new products and services to try— so you could encode that as a very health-conscious patient.
  
  By looking through all the unique traits within a persona, you can pair the quantitative survey data with the qualitative research and research participants.

‍

3. Iterate on your existing proto-personas based on what you learned.

• Come up with descriptive names. Avoid using a person’s name, because something like ‘Declining Dorothy’ or ‘Trendsetting Tracy’ will automatically generate bias in your audience’s mind. We love pairing words like Explorer, Leader, or Learner with relevant adjectives.
  ‍
• Further refine your personas by creating value statements—a creative writing exercise that mimics their attitude and key behaviors or beliefs. As an example, the very health-conscious patient mentioned previously could have a value statement like:
  • "I am determined to be as healthy as possible. I’m an expert at finding new health and wellness hacks to try—it’s borderline an obsession. You find the best new health products when you go off the beaten path a little bit, so I’m willing to try just about anything."
    ‍
  • Each point in this statement is derived from the qualitative encoding and backed by data points from your survey.

Quantifying proto-personas can be an extremely useful undertaking if your proto-personas need a bit more weight behind them.

If you need expert assistance in this endeavor, we’re here to help! Our team specializes in user research and persona development, especially in healthcare, ensuring your proto-personas are not only grounded in data but also actionable and effective.

Reach out to us today to start transforming your proto-personas into powerful tools for your business success!

What's next?

Have a project you’re working on and need some support? Reach out to us.

‍
Do you just want to chat about product, UX, research, process, and methodologies? We’re down for that too. Let's chat.‍

Do you want to avoid talking to another human being right now? We get it. Sign up for our Curious Communications newsletter to stay up to date on all things UX and other curiosities. We’ll hit your inbox every few weeks.