Katrina’s Blog™

News and Commentary

on the science and technology of drugs and medical devices, including discovery, development, manufacturing, and regulation.

Is This the Right Problem?

May 28, 2024
| Leadership

I recently encountered an interesting tweet from Michael A. Fischbach of Stanford University about selecting the right problem. In it, he shared his recent commentary in Cell, “Problem Choice and Decision Trees in Science and Engineering.” [1] According to his Stanford profile, [2] he is “the Liu (Liao) Family Professor of Bioengineering at Stanford University, an Institute Scholar of Stanford ChEM-H, and the director of the Stanford Microbiome Therapies Initiative.” He’s also the recipient of multiple fellowships and awards. Basically, the dude’s got some chops for selecting a good problem. I enjoyed reading his commentary (and the article that inspired him [3]), and I recommend them both to you. What caught my imagination is how the framework Professor Fischbach presents is applicable to startups in the ideation phase (something he references in the article). Let me summarize his commentary and follow up each point with my thoughts on applying it to medical product startups.

 A light bulb on a chalk board with blank idea bubbles signifying how to solve a problem.
Photo by Pixabay

Choose a Problem Wisely

First, Professor Fischbach recommends spending more time upfront to evaluate ideas before choosing your problem. He offers several approaches that have been tested in Stanford’s classrooms. Use “intuition pumps” [4] to generate lots of ideas. Treat each one skeptically, and evaluate multiple in parallel (dare I say in a survival of the fittest style paradigm?) He introduced an intriguing taxonomy of scientific problems in which ideas fall under three categories: perturbation, measurement, or theory/computation. Under each category, the idea is either a logic or technology problem. An example from the commentary is, “A new tissue-clearing technique is measurement technology, while the use of tissue clearing to study liver fibrosis is measurement logic.” The point of classifying your ideas is to understand the niches in which they fall and whether they align with any unusual skills or knowledge you bring to the project.

My experience with founders is that they frequently become attached to a single problem statement heavily based on personal experience. This can lead to tunnel vision on the solution and the customer, which impacts their ability to innovate and communicate the vision to others. Using an idea-generating framework and competing all the ideas equally will make the winning problem statement stronger. Additionally, you will have a pool of nearby ideas that can be used to pivot or expand the business. You will also have a checklist of other ideas considered with a rationale for why you chose the winner, which can make convincing points for funders. You can’t go wrong by taking more time upfront on this evaluation since it will take you 6-7 years or more to commercialize the resulting medical product.

Sticky notes on a cork board showing various problems and ideas.
Photo by Polina Zimmerman

Pragmatically Embrace Risk

Second, Professor Fischbach recommends evaluating and scoring ideas for impact and likelihood of success. He offers two tools for this work. For the Assumption Analysis, you list every assumption you make about a project from start to finish. Then you assign two scores: likelihood of success as a number from 1-5 (low-high) and duration of effort (in months). This list helps you quantify and work to eliminate points of risk and identify projects with appropriate levels of achievability and impact. He also suggests using optimization functions to evaluate impact, comparing questions such as, “How much did we learn?” or “How widely will it be used?” The problem you choose should balance the challenge and outcome at an appropriate level.

I’ve found the framework presented by Alberto Savoia in “The Right It” [5] to be a valuable approach for the early evaluation of medical product ideas in the marketplace context. Speed is a key part of his process, and the steps require low investments in time and expertise (especially when compared to the time you will take to convert your idea into an actual product). Borrow a copy from your library to confirm product/market fit before spending much time or money. The Stanford tools mentioned above can also be useful, particularly in cases where your idea is so disruptive it can be challenging to express. At the very least, using them in your early thinking and planning process will make a document useful for future critical evaluation (see the next point).

A man at a computer in a coworking booth.
Photo by Tim Gouw on Unsplash

Reframe the Problem if Necessary

Third, the professor notes that a project rarely follows a linear path, and appropriate time should be spent on its decision tree, adjusting your plan and activities to improve the outcomes. From the commentary, “The key to navigating a project’s decision tree is to move back and forth frequently between two types of work: getting stuff done (level 1) and evaluating it critically (level 2).” You can’t do both simultaneously, and the Stanford team has found anecdotally that successful scientists frequently switch between planning and doing. When problems arise, they are best viewed as opportunities to learn and improve your project, even redirecting it into a new problem/solution pair.

Learn as you go should be a mantra for founders and startup teams. A founder’s journey is nonlinear, and part of creating something new is finding unexpected challenges along the way. Performing a periodic evaluation of your strategy, tactics, assumptions, and systems should be built into your development schedule. Taking stock and redirecting where necessary takes leadership and pays dividends in team trust and morale. Nothing is more frustrating than continuing to push toward a goal that has lost its importance or achievability since the first proposed. Balancing resilience with practical course changes is a leadership skill worth developing.

An idea wall
Photo by FORTYTWO on Unsplash

A Final Point

Lastly, it’s worth mentioning the problem-ranking approach presented by Professor Alon. [3] His method of ranking problems for ease and interest and then assessing the easy-to-difficult continuum in the context of mentoring new scientists and principal investigators is a thoughtful, educator-first approach. Considering interest as a subjective parameter and exploring how choosing a problem that matches our inner voice gives meaning to the work and is more likely to result in a profound result. His “nurturing schema of research” also applies to practical laboratory management, where any problem can be used as a learning-mentoring experience for individual and team growth.

I hope you enjoyed my summary of the articles and thoughts on how they apply to startup ideation. I encourage you to read both; they are only a few pages long and offer excellent food for thought. Your big idea can only be improved in competition with other ideas. Quantitative data on interest and usability from customers will improve your chances of success.

I help people with big ideas for drugs, devices, diagnostics, and digital health. Reach out to me if you are struggling with yours!

Scrabble tiles spelling out creativity
Photo by Markus Winkler


[1] Fischbach, M. A. (2024). Problem choice and decision trees in science and engineering. Cell, 187(8), 1828-1833. https://doi.org/10.1016/j.cell.2024.03.012. Accessed April 12, 2024.

[2] Stanford Profiles, Michael Fischbach. https://profiles.stanford.edu/michael-fischbach. Accessed April 12, 2024.

[3] Alon, U. (2009). How To Choose a Good Scientific Problem. Molecular Cell, 35(6), 726-728. https://doi.org/10.1016/j.molcel.2009.09.013. Accessed April 12, 2024.

[4] Dennett, D.C. (2013). Intuition Pumps and Other Tools for Thinking (W.W. Norton & Company). Referenced as Table 1 in [1].

[5] Savoia, A. (2019). The Right It: Why So Many Ideas Fail and How to Make Sure Yours Succeed (HarperOne).


Latest Posts