UX research tips for customer development - notes from my session at Leancamp London 2

Notes and a list of resources from my talk at Leancamp London 2. I also added some bits I forgot to mention.
(I gave a longer talk with a similar topic, but different focus and audience at Lean UX Machine Tel Aviv, the mini Leancamp in Edinburgh, and at Lean Startup Machine London and Pakistan.)

We make assumptions about our customers and users. I love that lean startup encourages validating these with real data, but quantitative and qualitative (i.e. talking to people!). We aim to understand our customers' needs and problems, how they are reaching their goals now, and what our product or service could do to help them.
We have to think on our feet every day and decide on the best, most effective way to find out more about our customers and check if our hypotheses are correct.

To get better at this, to collect better data, I urge you to look into UX. A wealth of techniques, materials and advice. Here is how I structure my UX research toolbox:

There are methods to get quantitative or qualitative results, and generative (i.e. creating new insights) or evaluative (such as a usability test) feedback. I learned at university that for a good experiment you need to ensure triangulation. Which simply means: use more than one method. For example, if I'm using an evaluative method like a usability test, I combine it with a generative interview at the beginning of the session. If I'm doing early-stage qualitative customer development interviews, I might also do a survey about a specific aspect I'm investigating.

Research tips

1. State your goals clearly
Write down and/or visualise your assumptions, and what you need to find out.
We used the business model canvas as a tool to capture our assumptions about the customers and value propositions.
If you hypothesise with your team about your customers, make it tangible! We found behavioural variables useful: what attitudes and behaviours do you expect, i.e. do you have to validate? Creating a provisional persona (also referred to as proto-persona) can be helpful to visualise your assumptions about customer goals and behaviour.

Our assumptions were grounded in some research and background information, and especially trying to make the persona made it pretty obvious what we didn't know, and what seemed like risky leaps of faith. In an early stage, when you are still searching for problem-solution fit, an important goal for me is to find out if people care, if there is any emotional response to our idea. Indifference is the worst.

2. Think before you talk!
Please don't just run outside of the building. A small amount of planning will help you to get better results. 

For planning interviews, I like to use a topic map to lay out the problem space. Example: If you are, like I am right now, interested in older people's attitudes towards cooking, high-level topics on your map might be frequency of cooking, types of food, social interactions, frustrations and more. Agree with your team what topics you need to find out about. This makes it easier to ensure interviews stay useful and on track. 

I tend to use prompts rather than writing out questions. Remember, an interview is not a survey! You want to make people talk to you and ask open questions. To get the conversation started, it's useful to have a softball question ready. This is a question that is easy to answer. As an example, when we were approaching older people on the street for some short guerrilla interviews, we made sure we caught them while grocery shopping and simply asked them what they were buying (or had bought) today.

Be smart about how you go about finding out more about people.
Example: we wanted to find out about people's relationship with their family. This is a sensitive topic, so by asking them to share a story about a family recipe, and if/how they have been passing that on, we got to hear a lot about their personal life without appearing intrusive. 

Finally, please please please, especially before you do something quantitative and unmoderated like a survey, usability test your questions! Ask someone to fill in your survey in front of you before you send it out. If they struggle and ask you for clarification, you need to improve your questions.

3. Consider where to find people with relevant experience to talk to
Lots of UX folks use Gumtree and social networks for recruiting (if the customers are lurking in the general public). For more general topics, this can be enough. If you want to make sure you get the right people, you can send them from Gumtree or Facebook to a short survey. Especially with interviews, you want to make sure that you get people who have relevant experience that they can remember and share with you. 
There are great tools such as ethnio out there to catch customers 'in the act' while browsing your website, and recruit them right there and then for a remote interview or usability test.

If you need to get hold of people in the wild, think about when and where they are bored = have time. You want to do guerrilla interviews with young mothers? Don't look for them in the supermarket or on the street, look on google for the nearest playgrounds. Think about where people are waiting, queueing, smoking? If you're after businesses, when is the least busiest time?

4. During the interview: LISTEN!
Practice asking open-ended questions: start with who, what, when, where, how, and why. Say, 'tell me more about that'. You want people to tell you stories. Give people time to think. 
Janice Fraser from LUXr shares some great all-purpose questions in this interview
  • Has there ever been a time when you had x experience?
  • Could you tell me about that?
  • What was great about that?
  • What was awful about that?
  • Why did you do that?
  • And then, what happened?
  • If you had a magic wand, what would you make the situation be like?

5. Make the findings explicit
I like to team up when doing interviews - one person can focus on talking, while the other one takes notes. 
I analyse notes by grouping them into learnings and new insights. I check my assumptions, and force myself to write down 'yes, x was confirmed by...', or 'no, we were wrong about x, because...'. I also capture for what we didn't get any data.
Janice recommends this quick way of analysis by Rachel Hinman that I find very helpful:
  • What we heard/what we saw (one post-it per finding)
  • What it means
  • Why it matters

During the group discussion at the end of my session, we shared lots of other useful tips. 

  • If you work in an industry where it's hard to get access to end users (example was finance), train your sales team in open interview techniques. Make everybody who has contact with customers your proxy.
  • Be the customer service centre, take their calls. Ready For Zero do this - here's a post from their CEO about this, and UX Director Loren Baxter gave a longer interview for FastCoDesign.

I didn't take notes of everything we discussed, so if you were there and remember, please add in a comment!


So much is out there, but here are some favourites:
  • 'Mental Models' by Indi Young
  • 'Storytelling for User Experience' by Whitney Quesenbery & Kevin Brooks
  • 'Remote Research' by Nate Bolt & Toni Tulathimutte
  • 'Undercover User Experience' by Cennydd Bowles
  • 'Designing for the Digital Age' by Kim Goodwin
  • LUXr resources and materials by Janice Fraser and Lane Halley
  • The many useful articles on User Interface Engineering's site