JTBD interviews: How many do I need to get meaningful answers?

Our approach means we don’t need to conduct 100 or 1000 interviews to get our answer, we are able to do it in fewer. 

The interview is the most important part of conducting Jobs to be Done projects. That’s because ultimately it helps us to understand why a customer hires your product. 

But how many do you need to conduct in order to get meaningful data you can then use to develop better products?

Interviews, when conducted correctly and under the right conditions, can be time-intensive.

They’re a big investment. That’s why we’ve created a process – unique to The Re-Wired Group – that will yield the answers you need in just ten interviews.

10 interviews is all you need to make better decisions.

Our process is built around intent and philosophy, grounded in the Dr Taguchi method of Robust Design. This experience and learning is from working on launching 3500 new products. With case studies including Autobooks, SNHU and Intercom.

The interview process itself is largely based in a form of qualitative research. And the way we’ve designed our process relies heavily on methods used in engineering production. 

Intent – it’s all about understanding the why.

The trick of interviews and extracting meaningful information is in understanding the why; why someone does something. It’s not copying down words, the value is in listening to your interviewee. 

We find that if you conduct over ten interviews, the intent gets lost. You have so much information, you end up relying on words or the definition of words. Jobs isn’t about words, it’s about intent. 

Philosophy = a process that works

Most people believe that you need statistical significance in order to get real learning out.  

We use a concept taught to Bob by Dr Taguchi which is based on the Robust Design of Experiments. RDOE’s goal is to find settings that achieve a particular response with high consistency. 

Dr Taguchi taught us that if you understand what the levers or the inputs are to a system, we can design an experiment that induces that variation. 

We want to know if there is a cause or effect to fight the belief that demographics or segmentation causes people to buy our products. By using RDOE, we are forcing people to think about all the different variables at play and to try and capture as many of these variables as possible. 

This approach means we don’t need to conduct 100 or 1000 interviews to get our answer, we are able to do it in fewer.  

This approach means we don’t need to conduct 100 or 1000 interviews to get our answer, we are able to do it in fewer. 

The important first step of Jobs is to really understand the question you’re trying to answer. You need to understand it from the supply and demand of why you’re asking. This will tell you how to set up this experiment and the types of people you need to talk to. 

By using a system like this, you’re not relying on the luck of the draw. 

The types of people you interview are also important. 

You want to create variation around the right demographics and around the question you are trying to answer.

Variation means you want a mix of demographics. These could be gender, size of company, years in business, why you think they bought you, etc.

If you have segmentation already, injecting variation into it isn’t a bad thing to have. 

What’s the real reason behind wanting more than ten interviews?

The question about why you only need ten interviews is really question is rooted in fear.

The fear of missing out, the fear we are going to miss something. Perhaps you think your product is complicated and as such, you need more samples to extract meaning. 

If you follow the process we’ve outlined, the chance of missing out on something is very low.

When conducting JTBD project work, we’ll have some clients who ask for another two or three interviews. We’ll do it, but all of a sudden, they start to hear the same stories, again and again. If this happens you have most probably conducted enough.