How many people need to be interviewed to confirm the hypothesis?
I first came across the user interviews 6 months ago to validate our product idea. I wanted to talk to as many people as possible. Like, I was sure that the more people you talk to, the more valid your outcomes will be. So, excited, I texted like 50 people and appointed 20 calls for the next 5 days. That week I can associate with the total chaos. I held 4–5 interviews per day, they were all mixed up in my head. I didn’t have time for the most important part — processing results and making conclusions. I postponed the interview analysis to the next week to find myself in dozens of hypotheses and needed to discuss them again with the majority of the previous respondents.
Time passed. We came up with mockups and later on, prototypes of the product, that needed to be tested. This time I was prepared and set only 6 calls during the week. After 4th call, it was already clear that respondents have almost the same feedbacks over the interface and product flow. I highlighted the most common recommendations and implemented them into mockups. Only after finishing it I appointed the next 2 interviews. Less pressure, more effect.
So, here are my outcomes:
1You don’t need lots of interviews. You will see a correlation even on 4–5 people if they are representatives of the narrow target audience. For example, we have product and service tech companies as our clients. I considered this difference and conducted separate bunches of interviews with founders/recruiters of each company type.
2 If you already see repeatable feedback, better implement it in your mockups before the next interviews to get the complex product overview. My scenario is: interview 5 people — implement their advice — interview 5 more people.
3 The quantities of interviews depend also on the number of screens you are ready to show. In this article, I described the approach of dividing the functional into separate iterations for the test drive. In that case, you will conduct generally more interviews but it may still be 4–5 people per iteration.
4 We have both tech companies and tech candidates as our clients, so I have been paralleling interviews with them from the beginning of product development. I can sum up that all the tips listed above work equally effective to B2B and B2C clients.
Please, clap if you find this story useful and would like me to write more about product management in a startup :)
