How to do great user research

Great user research fits behind all the best digital projects. We distil the best insights from our years of running user research projects to explain how to deliver them effectively and tease our useful findings.

User research is an essential step for creating great digital projects, yet to add value it needs to be done right. We’ve run user research for a wide range of ambitious and award-winning projects, helping to shape our approach to developing websites for the likes of The Science Museum Group, The British Museum, Cancer Research UK, the University of London, and more. For us, user research isn’t a nice to have, it’s an essential step for delivering great digital projects.

Why do user research?

The best websites sit upon a solid foundation of excellent user research. This is what informs their structure and design. Ultimately, websites exist to serve users, and user research uncovers how they can most effectively do that.

In our experience, there are three key reasons to conduct user research for website projects:

  1. Collaboratively make informed decisions: Great user research helps to unite our internal teams and key stakeholders, because it develops a shared understanding of user goals.
  2. Eliminate assumptions: Projects can start with a whole host of assumptions about what users need/want and who they are. You need replace these with evidence. Assumptions lead to wasted time and effort focused on the wrong things.
  3. Justify digital spend: Robust research means you can build a strong case based on real evidence. This can be key when building a case for change in a large organisation. Website projects represent a significant investment, and good user research can give leadership teams the information they need to make informed, smart decisions.

How to conduct effective user research

Our extensive experience with user research projects has led us to the realisation that the best way to start them is to think about what you want to know, and work backwards from there. Don’t get jump straight into deciding which methods to use before you’ve worked out the problem you’re trying to solve. Visualise the perfect dataset that answers the most pressing questions you have. What does that dataset look like? What kind of information does it contain? Once you understand what you’re trying to create, you can work backwards to understand the types of research you’ll need to do.

It’s important to be specific. Asking broad and vague questions will get you nowhere. Very general research will waste time. A good example of getting specific is some user research we did with the British Museum. We focused specifically on the storytelling content they were producing, and we asked two simple questions: Are users engaging with the story content generated by the museum? And do they want to add their own stories? With that level of specificity, we were able to focus, select the right research methods for those questions, and uncover useful insights about how to boost engagement.

What to watch out for

The user is an invaluable source of insights, but you also need to have a clear distinction between identifying problems and designing solutions. A common mistake with user research is to ask participants who have identified a problem what they think the solution is. The user is not a design expert. When you look back at user suggestions, their solutions may be entirely different and contradictory. Experts should take the insights uncovered by the research to design solutions. This requires expertise. You can’t rely on a user to suggest the right solution.

It’s also useful to make a conscious effort to avoid attributing blame. In a sense, the user is always right, even when they are wrong. If they do something that seems nonsensical to you, there may be a good reason why they did it. This can often be solved with the right design.

Blend different research methods

You want a range of both quantitative (quant) and qualitative (qual) research methods in order to successfully generate rich insights. The qual goes deeper into the issue and lets you explore, uncovering unexpected questions you didn’t think to ask. The quant lets you establish concrete findings that you can be confident will generalise. It’s good practice to substantiate claims by using multiple methods of user research to corroborate. That means you can build confidence in the findings and be sure what you’ve uncovered will stand up.

You should build each round of research on the last, so don’t over-plan each round in detail before you kick off the project. Build in time to iterate from what you find out. If your survey uncovers a problem with a certain area of the site, focus the observation tests on that section.

How to run a great user survey

Surveys are a great tool to get a breadth of quantitative data on a website’s users. Surveys are also relatively easy and low cost, so there are few barriers to using them. Just like with your research project as a whole, the best way to design your survey is to start by thinking about the problems you’re looking to solve and the data you’d want to help make the decisions. Then the questions will flow from that.

You also need time to think through the questions from a user’s perspective. Remember that they won’t know industry specific terms or definitions. Always spell out questions so they are not open to the user’s interpretation. For example, if you ask, “do you visit this website regularly?” this won’t really tell you anything. The user does not know what counts as ‘regularly’. It is much better to be clear about these thresholds and offer a multiple-choice selection.

It’s also important to be alert to potential bias in how you set up your survey and how you can mitigate that. Be aware that it is likely your most engaged users who are responding to your surveys, so think about how you can address that with who you invite to take partFor example, if you have a list of highly engaged members, inviting them to take part in the survey will not accurately reflect your average user.

How to run great user observation tests

Observation testing is where you observe users navigating the site in real time, giving users tasks to complete and asking questions about how they found the experience. We find them a useful method when done right. They help you uncover things you wouldn’t otherwise have thought to ask, or get perspectives from users you’d otherwise have missed. [TJ1]

The single most important thing to do when running observation testing, is ensure you give users a clear goal. If you invite them just to look around the site and provide their thoughts, you won’t get useful insights. This is fundamentally because real users don’t come to websites to ‘look around’, they come with specific goals in mind.

You’ll also have better sessions if you align the goals you set the participants to what they are likely to use the website for in real life, as that will generate more authentic user behaviour. Think about the criteria you used to select the participants and try to tailor the activities you assign them to what they’d be likely to use the site for.

Remember to layer your research methods and build on your previous findings. We find it’s useful to use the survey to identify any problem areas of the website, then use the observation tests to focus in on them and ‘get under the hood’ to get to the root cause. So set the users goals with that in mind. Another tip that may seem counterintuitive is to deliberately recruit the more critical respondents from your survey to take part in the observation test. It can be more useful to hear from these people as you are more likely to uncover problems. Use the survey as a screening tool, and recruit those who reported having some issues. You can also use the survey to help you recruit specific groups of users you’d like to reach, such as a user with specific needs, or from a specific industry.

When running the sessions themselves, follow these tips to maximise their effectiveness:

  1. Tell the user that you can’t help- When they get stuck, you need to let them work it out on their own. Telling them upfront you can’t help sets expectations.
  2. Encourage thinking out loud- That gives much richer insights that you would get otherwise.
  3. Prompt users to be honest- Most people will naturally be polite and sugar-coat things. You don’t want to hear nice things about the website, you want to uncover actionable insights.
  4. Record the session- Because you’re asking questions, your mind will often be focused on thinking of the next thing to ask. Recording the session means you can look back and spot things you’d have missed first time. Always secure the users permission for this.
  5. Ask follow-up questions- At the end of the session, ask about what they did to probe why they did it. Be reactive and responsive. Keep it flexible so you can react, so don’t read questions from a fixed script.

Keep research and testing separate

Testing and research are two different things that often wrongly get conflated. Research is about drawing out insight identifying problems to shape the eventual solution. Testing is done to establish if the solution works. They need different approaches and different mindsets. Similarly, don’t do user research to validate your assumptions. You’ll invite naturally introduce bias. You need to have an open mind and not be looking to prove a hypothesis.

Turning great insights into great results

Conducting user research is an essential first step for delivering a great digital project. If you’re embarking on a user research project and you’d like an expert opinion, or if you’ve developed some great insights and now you need help developing the solution, we’d love to hear from you. Get in touch via [email protected]