In my previous post, I told you about what is a user persona and why you should consider building one (or more). This one would be dedicated to building user personas, validating them, and from now on, with what you have created.
Without further ado, let's begin!
How to create user personas
There are three main types of user personas:
- Proto personas
- Qualitative personas
- Statistical personas
If you have your data or willing potential responders, you can start with qualitative and move to statistical personas:
Qualitative personas are based on user research. You can apply different methods: user interviews, roundtables, usability testing, etc. To make a decent data pool, you'd need to target 5-30 people. After gathering your data, it's time to segment it – find similarities between different respondents and group them by match percentage.
Quick note: we're not looking for a perfect match; highlighted trends are way more important here.
Who to interview:
- Client decision-makers
- Users
- Subject Matter Experts
What to look for:
- Project/product expectations
- Pains and problems
- Problem-solving approach
- How they describe thought processes while working on a task and the specific terms they use
Statistical personas would work for you only if you have a significant number of potential research participants; under significant, we mean over 100.
- Like qualitative personas, statistical personas require user research on a small group of respondents (5-30, remember?)
- Based on the results of the user research, you'd need to prepare polls and surveys for the next step and conduct the study with over 100 participants (there's no maximum to it)
- Analyze the data. One of the most popular methods to do so would be a Latent Class Analysis (LCA)
But what if I have no user data whatsoever?
Here's when proto-personas come to play.
Proto-persona is a user research method based on assumptions about your potential target audience. It's a method that'd work for those who are only starting a new business/launching a new product/does not have any user data.
The simplest way to create proto-personas is to have a brainstorming session/workshop with your immediate team, including people from management and different professional areas. During the workshop, each member offers 2-5 personas who should be in the target audience, and your goal is to mix and match them all together to find similarities.
Important: due to the lack of real-world data, proto-personas tend to be a bit abstract, but you can and should validate them
But how to validate something that's based on assumption?
The answer is simple but a bit unethical – cyberstalking.
There's a thing called the "selfie method of user research." It means you surf the WWW to find your actual clients/competitors' client representatives on social networks, and then you build personas based on them. To validate already existing personas, you'd need to go the other way around – try and find real people within your potential target audience that would match the personas you've created.
Linkedin is the best source.
Who to involve in the proto personas workshop?
Everyone. Try to have as many functions and departments/professional roles involved in the workshop to get the most diverse results and crazy ideas captured.
What happens after happily ever after or after you've created and used your personas for the first time
After you've used your personas for the first time, you don't have to retire them forever – you can reuse or scale them up. Businesses, products, and projects grow – so can personas.
- Reuse old personas, creating growth scenarios. Project Manager David could become Project Director or dramatically change his career and switch to QA, for example.
- Add different methods as you go and gather more user data – switch from proto personas to qualitative and from qualitative to statistical.
Bonus: Common mistakes in building user personas
- The Data pool is too small. There's no such thing as too much data in user research, so it's always better to gather more than you need
- Unvalidated proto personas
- One-fits-all
- Overly detailed personas
- Making a list of things instead of a full-scale persona
Top comments (0)