5 min.
User research: laying the groundwork for killer insights
1L’art de la gestion de projet2Un projet à succès commence par une bonne gouvernance3Cascade, agilité, demandes de changement?

User research: laying the groundwork for killer insights

  • TECHNICAL LEVEL
Content & SEO Client experience & UX

It’s true, we talk (a lot) about user experience, and we are always looking for ways to create a good one (obviously). If you read my colleague Pierre-Olivier’s article, you also know that “the Google algorithm is getting better and better at figuring out whether or not your users are satisfied.” And to do that, you have to get to know your users, talk to them, get into their heads…

 

And to do that, you have to roll up your sleeves, and start doing research!

There’s no shortage of methods. Scientific testing in labs filled with sensors, combining eye tracking, measurement of emotions and the reaction of your digestive system (I’m kidding about the last one, but it’s not so far off), week-long safari-style ethnographic immersion sessions to observe users in their natural habitats… In short, there are enough options to keep entire UX research teams busy for a very long time.

But of course, we also have to be realistic, because these same research teams, devoted as they are to the cause, quickly slam headlong into the realities of time and resources, which are often limited. Fortunately, UX researchers also impress with their flexibility and pragmatism, because after all, better to do a short and sweet test with your mother as a subject, than no test at all, right?

Right, at least in part. Not everyone has the luxury to engage in months and months of intensive research before they start production, so a fast, well-executed test is sometimes the best solution.

Now let’s get back to the above-mentioned test with your mother, because today I want to put the question of process aside, and concentrate, at least for this article, on how you prepare your research and choose the users you’ll focus on. After all, if I wanted to open a steak house and decided to interview my colleague Matthieu, who’s just getting back from his third edition of Vegan Fest, I suspect we’d quickly learn it was a waste of time for both of us.  

If you’re in a hurry, I’d suggest scrolling directly to the end of this article, where I’ve included a summary of all its key points. If you like what you’re reading, you can always read (and reread) the details later!

The qualities of an ideal participant

Generally speaking, your user research Mr. or Mrs. Right will have a combination of the following characteristics:

Comfortable enough to share information with you

This is particularly important in the case of qualitative research like interviews or open exploratory questionnaires. Far better to have a person who talks too much, even if you have to channel their chatter, than a person who gives three-word answers to each of your questions. To figure out who’s who, there’s a simple technique to evaluate a participant’s level of openness by asking a slightly personal test question (in Just Enough Research, Erika Hall proposes the example of “Tell me how you usually decide what you’ll do with a day off?”).

The tone and level of detail of the response will quickly tell you how likely the candidate is to deliver the insight you’re looking for.

They’re invested, but not employees

Ideally, your research would focus on people who are likely to actually use the product or service you’re working on (thank you, Captain Obvious!), but you’d be surprised to learn how often it happens that a project is based solely on internal interviews with the sales team, or a prototype tested by colleagues.  

But you said we should use any means at hand! Yes, but if you want to give yourself the best chance of success, you should still take the time to avoid people who are too close, in favour of people whose judgment and priorities align more closely with real users. Remember that your users couldn’t care less about your internal designations, processes or political issues. What counts for them are their personal challenges, situations and needs.

Representative of a target group

Most of the time, your target audience can be divided into subgroups. Be careful—I’m not talking about demographic groups, but rather about groups that can be distinguished by specific behaviours, needs, situations and challenges that you’ll need to address.

User research might uncover new characteristics that you hadn’t anticipated, but if you have a minimum idea of your target audience’s expectations, you should be able, intuitively, to set up your initial segmentation.

OK, now that you have an idea of what this rare gem of a person looks like, let’s take a look at how to capitalize on that to string together a necklace of reliable, useful insights for your project.

image-2

Prepare your research: It all starts with what you hope to gain

And you’re off! Your project has started. You’ve probably already decided to do interviews or user testing, because after all, the goal is to talk to people, right? In addition, now you know the profile of your ideal participant… At least you almost do, because the characteristics I listed above apply to all research projects, but you also need to know which participants are most relevant for YOUR project.

As with anything, you don’t have much chance of arriving at your destination if you don’t know where you’re going. By the same token, to optimize the time you invest in your research and the quality of your results, it’s not a great idea to recruit blindly and hope to end up with useful information by crossing your fingers.

As with reverse planning in project management, I suggest waiting five minutes before you jump headfirst into your tests and observations, and start by defining why you’re doing this research.

Define the gains you hope to make through your research to determine its objectives.

The first thing to define is the expected outcome (or gains) of your user research: you didn’t decide to do research for the sake of research, you did it for a particular reason and to respond to a particular need. Maybe you’ve decided to target a new audience that you aren’t as familiar with, maybe you want to evaluate the quality of your current site or product, or maybe you want to know where you stand in relation to the competition.

It may seem obvious, but we’ve quickly skipped past some important steps. So grab a coffee or a tea, and ask yourself, “What is the point of this research? What results are we expecting? Who will make use of this internally, and how?”  

These kinds of questions will allow you to define a clear objective for your research project.

Define your research questions

Your objectives are clear—a small victory in and of itself—but we can’t stop there, not now that we’ve made such good headway, because the best way to know exactly what answers you’re looking for, is to write your questions. Your research questions should reflect—or better yet, clarify—your objectives.

Imagine you were creating a recipe site, and defined the following elements:

“As we optimize the recipe section of our cooking site (context), we hope to obtain more information on the way our target audience searches, consults and uses our online recipes (objective) to make sure we are able to define our recipe cards, search filters and calls to action in the most relevant way possible (gains expected from the project).”

Your research questions might, for example, be the following:

  • In what situations do the target audience search for and use online recipes? (Location, devices used, context, level of attention, etc.)
  • How do they search for these recipes, what criteria do they use?
  • How do they select a recipe from among the proposed results?
  • What information needs to be contained in a recipe card?
  • Once they’ve found the ideal recipe, what action do you want the user to take? (For example, read, save, share, categorize, print, etc.)

As always, sometimes, finding the right answer is as simple as asking the right questions!

Choose your research methods

As I suggested at the beginning of this article, there are countless methods available, but you can’t do them all. Look at them as a tool box, from which you pick the tool that best fits your need. I highly recommend reading this book—a collection of tools that has become one of my favourite references.

Your research questions will give you (the most important) part of what you need to know to determine the research method or methods most suited to your project. The rest will be defined by your constraints in terms of budget, resources and time.

Be careful, however, not to make too many excuses! Spending a bit more time on research can win you extra time for design and development, in addition to the fact that solid observations are a formidable weapon for shutting down interminable internal debates. Take the time to review your schedule if you think it’s worth it.

Another word of advice: whenever possible, combine at least two research methods for each objective: one quantitative method that gives you information about the behaviour of your users, and one qualitative method that will allow you to explain that behaviour. That’s the best way to get complete information on which you can base your decisions.

And there you have it. You’ve got your objectives, your questions, you know how you’re going to carry out your research (and it might have taken you less time than it did to read this article). There’s nothing stopping you now… Except, maybe, for two questions: how many participants do you need to recruit, and also—how do you find them?

image-4

Recruiting participants

Here’s a question that comes up early on. And when it does, you’ll be glad to have defined your objectives, target groups and research methods because, among other things, these factors are what determine the size of your testing group.

The number of users you’ll need to recruit depends above all on the method you’re using. For quantitative methods like, for example, an eye-tracking study, survey questionnaire or tree test, you’ll need more participants than what is required for qualitative methods like a semi-structured interview, a logbook or an open survey.

As an example, generally speaking, the minimum requirement for a questionnaire is 30 participants, while an eye-tracking test requires at least 40 to 50 potential users.

However, the size of your sample will ideally depend on the estimated size of your target population. You’ll need to break your main audience down into prioritized subgroups in order to study the particularities of each. That way, if your budget doesn’t allow you to cover them all, you’ll be able to concentrate your efforts on your highest priority personas.

Finally, note that the majority of online testing and research tools include a calculator to define the minimum sample size and number of participants necessary to ensure the reliability of results.

And finally, where can you find reliable users to carry out your research?

Lastly, here are a few possible alternatives for recruiting users who’ll be able to deliver the precious information you’re looking for:

  • Recruit users right on your website: Certain tools allow you to display a pop-up when a visitor arrives on your site, to ask if they want to participate in a study. Ethnio is one example, and also allows you to manage meeting schedules for in-person user interviews.
  • Use the recruitment tool integrated with your online research tools. Optimal Workshop, for example, will recruit testers that meet your criteria for you, for card sorting, treejacks and qualitative research.  
  • Use your internal database of clients and leads, on the condition, of course, that you’ve made sure the requirements for consent are fulfilled.
  • Go to a place frequented by your potential users, and recruit them on site. This solution is better suited to qualitative studies, or those that require fewer participants.
  • Hire a company specialized in recruiting participants for studies.

image-3

In conclusion

If you were in a rush and you scrolled straight down here, then congratulations, you made it!

To summarize, the most important thing to retain is that the key to successful research is to properly identify what your expectations are. If you know exactly where you’re going, you’ll only have to compare itineraries that are compatible with your constraints. Take some time to reflect on and define your research questions, it will change your life, avoid disappointment, and in addition it can show that you know where you’re going!

And if you can’t do all that, is recruiting your mother for a test still a good solution? Not necessarily, because even though she’s always there for you to unpack your Ikea when you move, or rearrange your furniture because, “it looks better like that, don’t you think?”, she probably doesn’t fit the profile of your typical user. And if (even worse), she’s directly involved in the project, I would advise you to drop the idea altogether.

Finally, if asked who was responsible for the quality of the research, I would say it’s as much the researcher themselves, who chooses and sticks to the right methodology, as the other stakeholders (those in charge of marketing, technical, graphics, finance or any other person with expectations regarding the research project) who will help the researcher to identify the objectives of the research and what’s expected of it.

Because in the end, what counts is that your project moves forward, ideally in the right direction!