Creating personas is an on-going process. Part of that process should be actually talking to your customers. Market research, narratives from sales people, or customer support… that stuff is great, but nothing beats getting feedback straight from the source. By sitting down face-to-face with your customers, you’ll gain invaluable insights that will help you reach the strategic goals of your company. Validating assumptions using the techniques covered in this article can help you craft more effective, informed decisions around your marketing mix, messaging, and design.

Yes, proto-personas are a tool to create a shared understanding but they are often rife with assumptions (that’s also kind of the point). It may be unwise to use them to inform design decisions because it’s essentially an untested hypothesis — the first step in the process. If the ultimate goal is to drive prospects through their buying journey, you’d be surprised how often internal bias can lead you astray.

Psst: Download our free eBook about interviewing customers.

Users and Customers May Not Be the Same People

If you’ve ever designed a consumer product, you’ve probably found that your users are your customers. They find your product, purchase it, and use it. Meet the needs of your users and you’ve met the needs of your customers. But what happens when your (buying) customers are different than your users?

That’s the challenge our client, New Relic, along with many other B2B SaaS companies are facing. It is imperative to understand the buyers’ unique needs and how they use the product (if at all). A DevOps Director at a large company will certainly have requirements (e.g., automation, configuration, training, customer support, etc.) that won’t be shared by a mid-level Engineer. This means you must meet your customers’ needs via content and features so they buy, use (maybe), and distribute your product across their organization.

Create Alignment with a Research Plan

blog_persona_magnifying_glass (1)

Set an intention and remember that you’re trying to expand on the initial exercise! Stay aware of where the answers lead you. What beliefs were challenged? What’s unknown? If your stakeholders were in the room during the proto-persona exercise then you may already know what they want to learn from detailed research; even so, conduct follow-up stakeholder interviews to make sure you didn’t miss anything.

Compile the essential topics you uncovered into a simple table: a column for goals and another for research questions. Think back to agile development: write an epic story (goal) and define it’s individual tasks (research questions) — you’ll soon find out if you’re biting off more than you can chew.

New Relic is keenly aware that as their business scales, they must continually refine their customer research. We recently created proto-personas with New Relic that yielded additional enterprise customer questions. As a result, one of our project goals was to simply narrow the enterprise knowledge gap.

As you can see, one goal can evoke many questions.

Be prepared to refine your research plan so you aren’t asking impossible questions or undertaking a project with no real value. Be honest with yourself; your goals must match your intended output. Our work was the impetus for an updated set of personas and a redesigned section of the website. While we could have conducted other research, it wouldn’t have been relevant.

Recruiting Qualified Interview Participants

blog_persona_chat_bubble (2)

When conducting customer interviews – recruit for behaviors, not demographics. We wanted to talk to enterprise customers with decision-making ability, buying intent, and domain/product experience. However, that’s not to say demographics aren’t important. We also recruited by job role because we knew that they’d exhibit different behavioral characteristics. For example, you may find that your product is bought and used solely by a younger audience so you’d be wise to recruit by age AND behavior.

New Relic’s product suite is immensely popular so we already had a pool of potential participants (don’t have an existing database? try these DIY methods). We created a screener that we sent out to customers who recently signed up and/or upgraded. Humans generally have bad memory — talking to people in the midst of product evaluation ensured that participants told us timely, vivid stories. Otherwise we would have run the risk of forcing people to inaccurately recount their actions.

Every screener question must be about phasing out interviewees you don’t want; leave the information-gathering to the actual interviews. Start with questions that are most likely to eliminate people at the beginning of the survey and work your way down. Due to our customer/buyer focus, we screened participants by asking about their buying interest and purchasing ability. Write your own screener using this handy template from Google Ventures’ Michael Margolis.

Keep this mantra: “If I interviewed another person, would I learn something different?” It’s not about statistical significance, it’s about deep insights. In About Face 3, Alan Cooper suggests that 5 – 6 people per presumed behavioral variable is enough for valid, research-based personas. Speak to just enough people and no more.

You’re trying to have a conversation with another person so you don’t want to sound like a telemarketer reading a script. A discussion guide includes the types of questions you’ll ask along with the topics you hope to cover. Try to cover the same topics with everyone, but feel free to explore related tangents during your interviews (not too many!).

It’s All About Behaviors

What are they doing and how is it being done? Focus on specific events and scenarios so people aren’t left to fill in the gaps — which they will — based on what they think they do. Start off broad by asking questions about the participant’s background. This is also a chance to build rapport before getting into the heart of the conversation. Ask open-ended questions about goals, context, process, and experiences.

We asked customers what they were trying to accomplish with New Relic, which allowed us to ask follow-up questions about how and if their goals were related to product evaluation. One helpful technique is to ask interviewees to recall a particularly poor or amazing experience. Be aware that interviews are likely to give you attitudinal data, but that doesn’t mean you won’t learn how someone may behave. It’s also a good idea to mix methods so you can listen to what people say and watch what they do. During our interviews we also observed participants completing tasks on the New Relic website.

Things to Avoid

You’re trying to get answers to your questions so it can be very tempting to lead people down a certain path. Don’t! Just like in court, lawyers never “lead the witness”. Instead of asking “How does your organization give you product goals?” (implying goals come from someone else) try asking “How did you arrive at your product goal?” Also, never ask what someone would do because it’s proven that people can’t predict their future behavior. Focus on current and past behavior to understand how they might behave in the future.

Creating Revised Personas

After you complete each interview, concisely summarize while noting behaviors that were exhibited. Can you find anything that matches or disproves your assumptions? Eventually patterns will emerge and you’ll be able to informally group participants. Try comparing each summary against one another to understand how many personas you may have. As you compare, group people by the behavioral traits they share.

Draw a range for every compelling behavioral trait you found with each side representing one extreme (e.g., not technical, very technical). Plot your participants on each range — this is more art than science — by relying on your observations. After completing this process, you’ll find that certain people tend to “hang out together”. If the same cluster occurs repeatedly across multiple ranges then you have a distinct persona. Be wary of clusters with conflicting behavioral traits! Either you’ve uncovered a goldmine or you need to go back through your notes and revisit how you plotted participants. For more instruction, check out Alan Cooper’s About Face 3.

blog_persona_poster (1)

List out each persona’s characteristics and categorize them accordingly. Try visualizing the behavioral ranges for each persona. This is especially helpful when people need to quickly compare personas to understand their similarities and differences. Other helpful storytelling tools are experience collages, journey maps, day in the life scenarios, and more. Don’t get too cute by adding fictional information that doesn’t add meaning. Unless it has a direct relationship with the persona’s goals or behaviors then leave it out. Personas should dispel stereotypes, not support them.

Share Your Insights

Talking to customers is an involved process and can take more time than you anticipate. If done well, the insights will inform important business and product decisions. Finessing your final deliverable can help presentations go more smoothly by psychologically providing more credibility to your process. Take some extra time to really dial in the copy and visual language.


Don’t let the vast knowledge gained through this process stop at your desk. Rally your organization behind your validated ideas and design a better solution for your customers.