On 2021-10-20, we concluded our 11 custom interview sessions for my side project together with 2 partners. Let me quote the problem hypothesis we wrote in one of our documentations:
“We believe people who are eager to improve themselves and busy do not have the time to read lengthy books when they just want to get the main ideas/ quick tips on some topics they are interested in because they are occupied by other priorities and it is not necessary to read the full book to get the essence of it.”
Based on this assumption, we built a website containing 15 mins summary of 50 books. It is still here: airchapter.com.
My past experience tells me that it is not ideal to jump into developing features after features. Most of the time, we were discouraged because, after you have spent so much time, no one is signing up. We think, this time, we need to talk to our customers to understand whether our assumption/ hypothesis is correct. And most importantly, whether it is painful enough to our potential customers.
What we have done is simple. We just find exiting users and potential users, through Instagram, and friends’ networks. It looks easy. But it is difficult to do it right. Even we read two books on how to do customer interviews along the way, including Lean Customer Development: Building Products Your Customers Will Buy and The Mom Test, we made several mistakes. But, at least, we know we made some.
I think the customer interview is effective because it saves us precious resources and time before jumping into something that may or may not work. During the whole process, we also learned a great deal.
What we have learned?
We have to qualify users before conducting interviews. Out of the 15 interviewees, only 7 of them are our target customers (we defined what is our target customers before). In the beginning, we were worried that users and potential customers are not willing to conduct interviews. So, we accept everyone. This proves to be ineffective because some interviewees do not even read books. Honestly, I do not know why they have signed up. And there is no need to worry about not having enough interviewees. I feel that people are generally helpful. If we cannot find anyone, it is our problem. How can you find real customers, if you cannot even find potential customers to have a zoom interview with you?
Do not use the future tense in questions. Sometimes, we asked very bad questions, like if we build X feature, will you subscribe? What are the factors that affect your decision to subscribe? Have you ever found it painful to do X? The answers to those questions are useless. It proves nothing. The customer replied, if there is a notetaking function, I may consider subscribing. I am pretty sure if we go to build this feature, still, no one will subscribe. Have you ever found it painful to read a full book? They will reply, “yes, sometimes.”. Again, it proves nothing.
Do not ask closed-end questions. Closed-end questions are misleading. They invite the customers in the direction you wanted. For example, do you think having a note-taking feature is good? do you find reading a full book time-consuming? Bad questions.
Observed emotions. We did not pay too much attention to the emotions of the interviewees. The reason why we believe it is important is that to understand whether the problem is painful enough, we need to observe the tone and emotion of our interviewees. If the problem really bothers them, they may complain to you.
There are a few more things we learned, like the equipment we used, the venue, the group size, etc. But the above 4 points are the most important learning points.
The essence of asking good questions is to put our idea and ego aside. We are not here to seek confirming evidence but disconfirming evidence.
Result from this customer interview
We used 1 month to build the prototype: airchapter.com. We used another month to do 11 customer interviews. We concluded that although the problem hypothesis is confirmed it is not painful enough to our customers. Therefore, we move on to the next thing. We feel happy since we saved time.
The next problem hypothesis we are looking at is:
“We believed that the current learning channels do not serve both learners and content creators well. For learners, they are easily distracted by other content and there is no systematic way of learning on those platforms. For content creators, current platforms are not friendly to knowledge-based content creators which may affect their income or popularity.”
Find the underserved customers and better serve them!