I always think about how to get more from conversations. I often felt like sometimes we would have conversations with top-notch experts in our field had heads filled with specialized information about how drugs are developed at pharma companies, or how billing works in healthcare, but I would never be quite satisfied at the end of our conversation.
I started to try adding structure to my meetings, where I would create a slide presentation for each interview. Each slide contained a hypothesis I had about the industry, the status about the hypothesis, and the next step. I instantly found the conversations I had to be much more stimulating.
When Kurt and I were first building Neurocurious, a machine learning product for pharmaceutical companies, one of the assumptions we had was that pharmaceutical scientists struggled with knowing whether a potential drug was toxic or not. Before structuring our interviews, we would bounce around the question, naively assuming, of course this is a problem scientists struggle with. We then started making slides that had tables like this:
And we slowly started filling them out. We learned that an assumption we had thought true to be true for many months — that scientists would use our software to learn the toxic side effects of their drugs — was in fact not a problem. When we interviewed an ex-Pfizer director who headed their toxicology team, we learned that pharmaceutical companies, for the most part, have tox figured out. He told us, “Toxicology was a problem 20 years ago, but not anymore.”
While large companies didn’t care very much about tox, we kept interviewing all sorts of people in pharma and learned that smaller startup therapeutics companies do care about their tox, because the large pharma companies that might acquire them want to de-risk the potential drug as much as possible.
We revised our slides:
Over the course of interviewing over 100 executives, business developers, and scientists in the pharmaceutical industry, and iterating on our hypotheses, we now have a set of refined hypotheses about the problems that pharmaceutical scientists struggle with. These set of problems can be used to design a product whose value proposition aligns much more closely with the problems of the pharmaceutical industry, than the value proposition we had originally began with.