December 13, 2023 7 min to read

Building Backwards

Building a biotech company solutions first.

Here we explore the business of biotech and how building backwards can help narrow in on a market niche.

Background

The first thing I learned when spinning out our university based research into a startup, Saguaro Solutions is that things that make great projects in academia are not automatically great projects in industry. It’s not that academic research is bad, it’s just that the goals are completely different.

In Presse Lab, my PhD lab that specialized in building machine learning tools for biophysics applications, we were very diliberate about the projects we chose to work on because we knew that it can take years to get a project from concept to publication. Nothing feels worse than spending years working on something only to find out after the fact that nobody is interested in it. My advisor, Steve Presse, was very good at picking projects that the community cares about. He instilled in us a drive to do the same by talking to people and having high level conversations about what the community finds interesting. As a result our lab frequently publishes in high impact journals and wins large grants, two metrics which are important for academic success.

A successful company is not much different from a successful lab in that it needs to be working on something that people care about. The difference is that a company, at the end of the day, needs to make money. So while it might be very interesting that we can image micro-sized parts of a cell never seen before, if nobody is willing to pay for that then it will not survive as a business.

So as we shift from from university lab to industry lab, our approach is to use the same concepts of project picking that we mastered in our academic work, but to do so with a business mindset. So the question is: how do we pick a good project in biotech?

Tech is not Biotech

One tried and true approach to picking good business ideas in the tech sector is the Lean Startup methodology. In the Lean Startup methodology you start with a guess about a problem that people have, then iterively go through cyles of Build, Measure, Learn. In the Build phase you create a minimum viable product, in the Measure phase you see how people respond to it, and in the Learn phase you evaluate how to improve the product. After completing a cycle you go back to the Build phase and repeat the process. The problem with this approach in biotech is that the product lifecycle is completely different than it is in tech. Building a drug or a diagnostic tool is expensive and time consuming. Often there is no such thing as a minimum viable product: either the drug works or it doesn’t.

In biotech, it is much harder to start with a guess then iterively diffuse towards a stable product idea. Instead, we really need to have a clear understanding of the product before we build it. This is where the concept of building backwards comes in.

Building Backwards

The concept of bulding backwards is laid out in two great books Biodesign: The Process of Innovating Medical Technologies and Building Backwards to Biotech. The core idea is that the market need should be identified before the product is built.

Biodesign lays out a mantra for building a biotech company: Identify, Invent, Implement. In the Identify phase you search for market needs, then filter out the ones that are impractical. Once a problem niche has been identified you move to the Invent phase where you propose and screen solutions. Once an appropriate solution has been identified you move to the Implement phase where you create market strategies and business plans. The key difference between Build Measure Learn and Identify Invent Implement is that the Identify phase is done before the Invent phase. This is building backwards.

Building backwards is a means of identifying a problem in the Identify phase. If we have a target number of patients we would like to treat, then we need to be working on a disease that affects at least that many people. If we have a target price point for our product, then we need to be working on a problem that people are willing to pay that much to solve. If we have a target time to market, then we need to be working on a problem that can be solved in that time frame.

Our strategy

What I like about building backwards is that it’s very analytical. If our goal is to use our technology to develop a clinical AI diagnostic tool then we can solve for the best problem to work on. For example let’s just start with the end goal that we have an FDA approved AI diagnostic tool.

• If we have an FDA approved diagnostic tool then that means we have passed clinical trials.
• If we have passed clinical trials then it means that we have paid all the costs to have a clinical trial.
• If we have paid money before generating revenue then that means we have investors.
• If we have investors then that means we are VC fundable.
• If we are VC fundable then that means we are expected to generate at least $X per year (where estimates on $X vary firm to firm and year to year).
• If we are expected to generate $X per year then that means that we are expected to have Y patients per year with a cost per patient of $Z, where $X = Y*Z.
• If we have Y patients per year then that means that W>Y people screen for the disease that our tool targets and that we have captured Y/W of the market.
• If each we get $Z per patient then that means that comparable solutions cost $Z per patient.

So theoretically if we start with a large database of diseases and diagnostic tools then screen them based on factors such as number of patients and comparable products, we can filter down to a list of problems that we should be looking at. But that is the first iteration!

We then need to dive deeper into each problem looking at other factors at a finer scale such as patient experience (whether patients feel there is a need for a better solution), provider costs, and competitive landscape. At this finer scale each larger problem spits into several smaller problems. We can then repeat the process of filtering down to a list of problems that we should be looking at. The end goal is that this iterive process of Needs Finding and Needs Screening will ultimately lead us to a problem that needs solving.

This is a simplification of the process, but I believe it is a good illustration of the concept. The important part is that building backwards takes a very complex and intimidating problem and breaks it down into something algorithmic and manageable. And if there is one thing that physicists turned AI scientists turned biotech entrepreneurs are good at, it’s solving algorithmic problems!

Call to Action

Here I outlined how building backwards can be used to build successful companies in biotech, but I believe that this same approach could be applied to other industries, even other fascets of life. I challenge you to think of ways you can apply backwards thinking to your own life. If you have any ideas, please feel free to share them with me!