We do research to bridge the knowledge gap between reality and our model of it. Without writing a single line of code, user research sessions rapidly generate a lot of insights. However, running a research session is hard, we are distilling insights from anecdotes shared by imperfect historians. To get meaningful data in a short amount of time, we need two things in our arsenal:
- Questions to uncover the problem
- Tactics to keep the conversation focused
Questions to uncover the problem
There are two types of research sessions that I frequently run:
- Early stage research
- Getting feedback on ideas
Early stage research
The goal of early stage research is to uncover what’s holding people back from doing what they want. I have had success with Chuck Liu’s three questions:
- What are you trying to get done? Why?
- How do you do it today?
- What could be better about how you do this?
What are you trying to get done? Why? Use this question to gain context about their goal. Helpful follow up questions to get a full picture of the environment in which they are looking to solve problems:
- Why is this goal important?
- Who are the other stakeholders?
- What happens next after the goal is achieved?
How do you do it today? Use this question to develop an understanding of their current workflow. Additional follow up questions to understand the details:
- When was the last time you did this task? Was it representative of what you usually do?
- What are the key steps?
- Could you walk me through it?
- What are the artifacts that accompany these steps?
- Who is responsible for each step?
What could be better about how you do this? Listen to answers to find opportunities and ideas. Questions to get people to reflect:
- What is harder than it should be?
- What takes the most time?
- What unofficial workarounds have you had to build?
I find it helpful to conduct research by keeping in mind the summary I would share with the team in the end. The research summary looks like:
In order to _____ (goal), users _____ (action). Today, they achieve this goal by using ____ (current approach). The challenges with this approach are: 1. ____ 2. ____ 3. ____ Ideally, they would prefer a solution where ____ (ideal solution).
*In order to* know when to run an analysis, *scientists* need to know the status of their request. *Today, they achieve this goal by* emailing people to ask for updates. *The challenges with this approach are:* 1. Tedious back and forth between scientists to get requisite data 2. Hard to keep track of all the requests being fulfilled 3. Difficult to split by a request between assignees *Ideally*, they would prefer a solution where* they can see the status of all their requests, broken out by sample.
Getting feedback on ideas
The best way I have found to get input on a proposal is from a walkthrough where people think out loud. The key objective is to get people to share how their world will be different with the new solution.
Typically, I frame these sessions by talking through the key features in the mock, and then asking users to walk through their own task.
Let’s take the last time you did this task. Could you walk through how you would do that task using this design.
As they walk through the design, I look for moments where they pause or look confused. These become segues to ask for more details and tease apart their mental model. Helpful questions to guide the conversation are:
- How would you start? What are you looking for?
- What do you think should happen?
- It seems like you had some concerns about this. Could you tell me more.
- What could be better?
Keeping the conversation focused
Research sessions easily veer off track. It’s challenging to communicate effectively with someone you just met, and harder still to get the insights you need. Some tactics I have found helpful to keep the conversation focused:
Be concrete Have people walk through the last time they ran into an issue, instead of asking about how they usually do a task. Responses to “How do you generally do this task?” are often indicative of how think they do the task, rather than how they actually do it.
Asking people to walk through the last time they did the task brings out how they executed the task, and the people and systems they needed to interact with to get the job done. Instead of “How do you generally do this task?”, ask for when they last did the task.
When was the last time you did this task? Is this task representative of what you frequently do?
Get into the weeds by asking for artifacts such as the spreadsheets they used, or documents they created. Even better, do a screen share or in person visit where they can walk you through their flow.
Focus on the problem, thank people for suggestions Good problems are common across different teams. Suggested solutions are unique to each team’s circumstance. Don’t expect people to be able to tell you exactly what they need, instead observe and reflect with them to tease out the challenges.
Echo back Paraphrase new insights and echo it back to people so that they can confirm or clarify. It ensures that you have the correct takeaway. As an added bonus, paraphrasing forces you to evaluate whether you truly understand their needs.
Cherish noes and moments of hesitation When people hesitate or look confused, you have a rare moment to unpack their concerns. Instead of explaining the product further, ask them to describe the issues they see.
It seems like you had some concerns about this. Could you tell me more.
For questions on how the product behaves, focus on what’s prompting their questions. Gently reflect back by asking questions such as, “How do you think that would work? What would you like to happen?”
Don’t sell the product Explain what the product does but don’t try to convince people of it’s value. When people see you being defensive about the product, they stop giving honest feedback.
Don’t be shy about asking follow up questions when you don’t understand When you don’t understand something that a technical user is saying, it’s not always clear whether it’s because you lack domain knowledge you are expected to have, or don’t have context about an organization’s internal jargon. A question that helps get more context without worrying about sounding ignorant is:
“I know different teams approach this differently, how does your team do it?”
Explain what you are trying to understand People get frustrated with questions whose purpose they don’t understand. Ask for their patience when you get in trouble and explain your motivation. For example, the motivation for a lot of my questions is to understand how common a problem is, or how frequently they will need to use the feature they requested.
What research questions and tactics have you found the most illuminating? Let’s continue this conversation on Twitter @mbe