Assumption-based personas and how they affect recruitment and research

This image has been doing the rounds on social media again and it makes a great point about personas.

Obviously Prince Charles and Ozzy Osbourne are very different people, but the truth is that they share quite a few demographic details. This is the thing about personas: they rarely are as simple as a compilation of locations, ages and top-level demographics. Therefore, the above is not what we would consider a successful persona.

Personas do not – or should not – equal a hard list of rules. Instead, they provide design and product teams (and suppliers such as People for Research) with an overarching idea of who they need to include in research.

In summary, the importance of personas cannot be underestimated in user research, but sometimes they’re not as great as we think they are and we might be too close to the action to see their flaws.

Misconceptions with personas

We’ll start by digging deeper into the obvious example above: demographics are not personas. Although this information is important to capture during the screening process, targeting participants exclusively based on a demographic profile will potentially mean missing out on the right people. Personas should lean more towards experiences, interests and behaviours.

But, you might be thinking: “user researchers stopped building demographic personas ages ago”.

The problem that we currently see quite often is more related to how the personas are created (the source of the information) and their relevance for user research, depending on which team has created them.

As personas are used by multiple teams within an organisation, some teams’ priorities are put first and other teams’ needs are overlooked. Marketing and sales both require and provide different information when creating personas, as do the product development team and the design team, while the C-Suite team might go in a completely different direction.

The interpretation of what is important gets lost across silos. Cooperation between teams and the ability to create different sides of the same persona depending on internal priorities is key – this is the right time to have a wider discussion with all teams involved about the types of people you need to understand or speak to, what motivates them, where they spend their time, what behaviours they have and how you can optimise products and services to make their lives easier.

First problem identified and out of the way. Time to handle our second challenge...

Grouping people

As soon as users are grouped into personas, we end up making assumptions and ultimately exclude some users’ needs to accommodate the rest of the group. We understand grouping people is essential to understand general journeys through websites, purchase decisions, etc. However, we see the tendency to create these groups off the back of the demographic information, and not based on their behaviours and experiences.

For example, imagine you are looking at both pet owners and new parents during the toy purchase process. The simple split would be to look at one group versus the other, but depending on what you’re looking to get out of the research, it could be more beneficial to group them based on their shopping behaviour, allowing you to create groups such as ‘savvy shoppers’, ‘luxury buyers’, etc.

We recommend this useful article by the NN Group on why some personas just don’t work.

It’s all well and good talking about how terrible personas are, but businesses and agencies still need them to find the right participants for user research and more.

Creating personas

Let’s start by changing that wording. Personas are not created, we do not sit in a board room and invent groups of people that behave a certain way or experience specific things. Personas are discovered after research and then crafted after further discovery through interactions with users.

The second, often overlooked point is that not every user problem can be solved by a single product or service, or by creating something new that will miraculously work for every user experiencing a similar issue. A solution created in isolation won’t work most of the time.

The final problem is the temptation of allowing individual user issues that match your expectations become part of your personas. Don’t let confirmation bias change your personas if the individual issue or need you’ve identified is not part of a trend.

Research cuts through

Running some user research – such as high-volume surveys – at the beginning of the process is key to establish a relationship with your users, helping you to create solid personas that contribute to business growth and user-friendly products. If you haven’t worked on your personas in a while, research also helps to update existing profiles and communicate any changes to your team.

Look beyond the scope of your product and get to the crux of users’ interactions, behaviours and experiences. The users’ problems or needs don’t have to be solved at this stage, just understood and acknowledged.

At this stage, one of the elements of bias you need to look out for is potentially leading participants with badly built research questions. Rather than trying to prove or disprove personas, take the time to write a good screener, get colleagues with different perspectives and goals involved, and listen to people to get the best results.

How much input do researchers have?

As mentioned before, personas are very rarely created by a single department in a company. Rightly so, there will be personas for marketing/sales, business development, customer service, the C-Suite, etc. – and each of these groups will create their personas based on their own needs and objectives.

The problem is when personas are siloed, instead of created collaboratively across different departments with the involvement of researchers and UX design teams. Cooperation between departments is essential during the creation process and, although they will be have different sides to them, these variations of the same personas should be consistent.

What are your thoughts on personas in research? It would be interesting to know how you use them and whether your research is persona-led, used to uncover personas, or none of the above.