This blog was written by Ankit Mahadevia, CBO of both Rodin and Spero, as part of the “From the Trenches” feature of LifeSciVC.
In a period of surging optimism for our industry, it’s easy to forget that historically most of the therapeutic projects we undertake don’t make it to the market (see the data here if you absolutely have to remind yourself of the odds).
A vast body of literature (examples here, here, and here) attempts to provide a life preserver in this sea of uncertainty by counseling us on “fast failure.” The basic idea is that, given our odds, shutting down troubled programs early when costs are low greatly increases overall success of a therapeutic enterprise. This is especially useful in a cash-constrained venture backed biotech, and in my observation, a critical discipline for investors to see in any team that wants to grow a pipeline. However, while the why is well covered in my observation, very few authors go into the how especially for early stage businesses.
I’ve collected a few observations and lessons learned as a startup team member on the very messy business of “failing fast.”
- Think about the end at the beginning After the data comes in, the variable nature of some animal models and diversity of enthusiasm within a Board for a program can inspire heated debate on a program’s prospects. In my experience, thinking about the program’s end in terms of “no-go” criteria when you start allows for a rational debate and consensus before the end is near. Setting a bar that retains a winner and weeds out lower quality programs requires a tough set of judgment calls (subject of a blog post in and of itself). The process requires discussion, debate, and most importantly refinement as a program goes on. In particular, this process is very difficult if done on the fly in the Boardroom or in a crisis. One new company I’m currently involved with killed two projects very early in the R&D cycle (after an average of 5 months of effort and $125k direct cost per program)– in each case we stated and refined our goalpost (a specific efficacy profile) at every meeting with our Board and academic collaborators on the project as we progressed, and this common understanding made the “decision to discontinue” smooth even for the academic collaborators this impacted most.
- Team spirit (or lack thereof): Explicit go/no-gos are crucial in executing on a quick kill, but intangible signals also help. In a setting where there’s difference of opinion on a program decision, how hard do your scientific colleagues closest to the program go to bat for it? Large gaps in enthusiasm among team for a program ought to drive deep discussion about the end game. In my opinion, team views on a program’s prospects should not be a guessing game; I’ve written previously on key attributes of new team members and a “truth-seeking” mentality is a critical one. Especially in an emerging organization, driving to the right scientific answer over the right answer for an individual is crucial and it takes a specific mindset and culture.
- Sooner rather than later: @LifeSciVC has written about the sunk cost bias in venture capital. It applies to program decisions as well. The more capital goes into a program, the stronger the collective tendency to think about advancing a program until it recoups that investment. A kill decision can be a tough discussion that’s tempting to defer if a program has a currently limited impact on the budget. However, pulling the ripcord (or at least having a discussion about it) early can save a team from putting good time after bad (and, of course, good money after bad).
- Listen to the market (with big caveats): This is the trickiest parameter of all. The entrepreneur’s mentality is to forge ahead even in the face of rejection. However, especially if other programs in your pipeline are getting rave reviews, the cold shoulder from a prospective investor or partner can be a helpful data point. A few caveats here – first, I’d distinguish strongly between the classic non-specific “too early” response and real, specific, constructive feedback. Second, regardless of what feedback you receive, it cannot trump your team’s inside knowledge and instincts about whether to continue to sail against the wind. In one company setting, our team killed a project in the orphan space based on the totality of market feedback despite tantalizing preclinical data. The primary feedback was from multiple conversations with two partners with the deepest expertise and wherewithal to partner in the relevant therapeutic area. The concern was valid, and the path to address the concern extended all the way to the clinic. With real tangible feedback and no early “go/no-go” to guide us, we made the tough decision to move on to other projects.
The decision to shut down a program that could ultimately fill a patient need is never easy. There’s no magic way to create a culture that appreciates the need to fail fast and often. I’ve written before about recruiting team members that have the risk tolerance to thrive in startupland. In my view, along with truth-seeking behavior, it also pays to recruit colleagues with a high failure tolerance. Especially in a startup setting, early decisions on program investment set the tone for future ones, and for a culture built around embracing failure as a key first step to success will pay future dividends.