When you’re in the customer discovery phase of building a new product, few things are more important than speaking to your target users. Problem is, in hindsight many of these conversations aren’t useful and, worse, keep you from building something the market wants. Don’t leave it to luck to gather the information you need.
Make your own luck by applying a scientific process to your customer discovery work. Before a line of code is even written, the best teams already have a significant amount of confidence that there is a market for their product. They avoid the “launch and see” approach by conducting well-planned generative and evaluative interviews.
So, before you get out out of the building and talk to potential customers, know which questions to ask, how to ask them, and even how to react to unexpected answers. All are key to learning the truth about what your market actually wants before you build anything.
You probably realize by now that a successful research interview starts long before the first question is asked. Heed Abraham Lincoln’s advice and sharpen your axe. Take the time to fully understand what your goals are and base your preparations around them. Otherwise, your interview is going to yield the wrong data.
These steps will help you run an effective generative interview, create quality records, reduce bias, and ultimately make better decisions for your future product. Here we’ll walk you through what we learned the hard way after we conducted nearly 100 informal interviews but still built the wrong product. It wasn’t until after we took a research-first approach to customer discovery that we were able to ensure we were building the right thing before we ever built it.
Here's how to conduct user interviews:
Before you draft any questions or pick any interviewees, you need to figure out what it is you need to know. Start by identifying what must be true about the behavior of the future users you have in mind for your theoretical product. Then make a plan to learn more about these assumptions by asking the right questions.
Here are a few of the questions we asked in our user interviews:
Rob Fitzpatrick, author of ‘The Mom Test’ says most founders often get lied to because they ask the wrong questions.
You can’t pick anyone off the street for user interviews. You need to choose an audience that’s representative of your ideal user base to get the most accurate feedback. The wrong people will give you bad data, which will result in inaccurate findings. Garbage data in, garbage data out.
At Grain, finding the right people meant narrowing our scope and focusing on people already doing the pre-requisite actions for our product to be useful, recording calls. We weren’t about to convince people to record calls in the first place, but we could see if they were experiencing problems that could be used to influence how Grain developed.
Choose people who will give you honest feedback, which means users who may even challenge your hypothesis. Here’s a guide on how to recruit candidates for your user interviews.
We get it. The stakes are high, you think you have great ideas, and you’re looking for validation wherever you can get it. This isn’t the place. The people who will stroke your ego should not be the ones you interview.
Start by avoiding falling prey to selection bias when choosing who to interview. Our inclination to talk with only those who will say nice things is natural. It’s not helpful. Rather than choose people who will be “nice,” choose people who are experiencing the problem you’re trying to solve.
You know when you’re having a bad hair day but your mom says you look great? You want to avoid that possibility, too, by asking the right questions in the right environment. If users know they’re talking to you about your company, your product, or you personally, they’ll likely feel inclined to be nice. Learn more about how to set up your environment and create a user interview guide here.
Avoid asking users leading questions, which risks confirmation bias. Don’t ask people anything that will guide them to an answer you want to hear, but instead ask them questions that will yield accurate data. For example, asking, “What kind of problems do you have when recording phone calls?” requires a negative answer. “Tell me about the last time you recorded a customer conversation” lets the interviewee speak to their experience, including any problems.
If your users can take in clues from you on how to answer a question, then your data is biased.
Ask your interviewees about the present, not the future. Too often we’re focused on getting people to speculate about what they might do instead of focusing on what they already do. (We’d know, we were guilty of it too.) It’s the difference between asking “do you record meetings?” and “Do you wish you could record your meetings?” That’s a big difference.
One gives you the truth. The other gives you the customers’ best-guess at an “ideal.” And best guesses often miss the mark. Questions should be aimed at helping you answer — whether your users realize that there’s a problem (or not) and If they do, how they are dealing with it now.
Just because the users you interview are a great fit for the product doesn’t mean you can ask them for ‘what they would like you to build’. Your users are best suited to help you understand the problem, not the end product.
A key example: Henry Ford is famous for saying, “If I had asked my customers what they wanted they would have said a faster horse.”
To put it another way, research interviews aren’t about asking users what they want. They are for learning what your users do now — and understanding what is and isn’t working for them.
So, learn what your customers' biggest problems are now so you can solve them with your product in the future. If you ask them what they want, you’ll end up with a Homer Simpson Car.
If you’re not recording every detail of your interview, the decisions you base on your research are already in jeopardy. Humans are pretty bad at remembering things properly or completely. We have a tendency to remember only the information that helps our business case. Couple of things to keep in mind:
- Make your recordings easy to tie back to your notes: it’s not only about the quality of the records, but also their usability.
- Keep your notes clean: this means highlighting key phrases or comments that stand out to you, writing down physical observations, and most importantly paying attention while taking notes.
- Be mindful of the fact that the interviews aren’t just for you: research interviews are for every person involved in the product development and decision-making process. Recordings and notes need to be easy to comprehend, even for those who weren’t at the interview.
- Share better research notes: Whether you’re collecting ideas for a new product or gathering feedback about an established one, the data from your research interviews needs to reflect the experience of the interviewee as closely as possible. Learn how to take and share better research notes here.
Recording every interview ensures you keep high-quality records for you and others who may be using the data later on to shape your product.
You may have quality data and records of this data, but you’re not in the clear yet. You need to properly analyze and interpret your data, or the entire user interview process was for naught. There are a few steps involved here to keep your analysis from totally derailing.
- Find patterns: It’s acceptable to have some bias here, as it likely spawned your hypothesis. But, you’re going to need a strong case to prove your stance. Look for patterns and themes in the data, like people harping about a certain problem or task they wish they could do better. Over time these patterns will become easier to see. Use those patterns to prove or disprove your stance.
- Perform periodic analyses: Analysis isn’t something you should save for all at once after the interview process. Analyze data between interviews to spot themes or identify questions that aren’t yielding quality responses. You might realize a certain question needs to be reframed, or that all your data feels off because your audience is wrong.
- Use the literal voice of the customer: In the very likely scenario that you’re not the one calling all the shots, you’re going to need to convince others of the direction your product needs to go. Use video clips to show the customer voicing problems they’re having. Using the literal voice of the customer is a concise and influential way to argue a case and drive home the problem at hand.
Finding product market fit can be brutal. Worse though, is going through the user interview process only to find you didn’t even get that market fit.
Accurate records truly are everything in this process. They’re the only way you can look back and find the flaws in your reasoning, poke holes in the hypotheses of others, and get your product to the right place.
Hindsight is always 20/20, but with an effective user interview you can gain the foresight to create a product that helps users solve problems in ways they didn’t even think possible.
Try out our ready-made and tested Zapier templates using Grain. Click the “Use this Zap” to get started.