Are you safe or are you comfortable?
In this week’s newsletter, I’m continuing to explore arguments for why biotechs should invest more in the mundane aspects of their data systems. And I’m going to frame it using what Seth Godin calls the Icarus Deception, which distinguishes between safe decisions and comfortable decisions.
The idea is that safe decisions are choices that maximize your chances of success (however you define that) while comfortable decisions are the ones that mimic choices that made you (or others) successful in the past. We often conflate the two because during times of stability, the comfortable decisions that worked in the past are most likely to continue working, making them safe. But in times of change, those things often stop working. And when the comfortable decisions stop working, they’re no longer safe.
Right now, biotech is in one of those times of change. (Maybe it always has been, but it is now too.) The cost of developing a new drug is becoming untenable at the same time that new data sources and AI/ML tools are creating potential avenues to bring it down. Meanwhile, the macroeconomic situation has completely upended the expectations that investors have for biotechs seeking funding. As runways get shorter, decision makers lean more and more on safe decisions.
Or do they?
The problem, of course, is that it’s fairly easy to recognize a comfortable decision. But in times of change it’s almost impossible to recognize a safe decision. So decision makers looking for safe decisions default to comfortable. And the outcome isn’t always what they were looking for.
So what does this look like for biotechs?
Well, making a company page with futuristic graphics, fancy transitions and a sprinkling of buzz words is relatively inexpensive. So, not surprisingly, it’s comfortable. Hiring a comp bio/data science/ML team is a much bigger investment and it feels like putting your money where your mouth is. But it’s far too easy to treat them like you would a marketing team: Let them work their magic on what the core business produces (lab data), but don’t change the core business. So it’s ultimately still comfortable. You’re not safe yet.
As many of you reading this will have experienced firsthand, when you only make these kinds of peripheral changes, the data team ends up continually blocked because what the lab produces, when left to its own devices, isn’t what a data team needs. It’s the worst of both worlds: Investing resources in the more future-safe option while keeping enough resources in the comfortable option to make sure nothing changes.
To actually use AI/ML in any meaningful way, to take advantage of all the new data sources, to become (more) data driven - biotechs need to make changes to the core business, i.e. the lab and the experimental process: Designing broadly scoped experiments to generate training data for models. Pulling data scientists into planning processes from the very beginning. Capturing and sharing metadata in a consistent and detailed form.
Even before the current market situation, there were very few biotech organizations that truly understood what this meant, and even fewer that were willing to do it. Today, as biotechs look for ways to extend their runways, many are retreating even farther back into the comfortable ways of working.
And sure - when the goal is survival, you do what you have to do. Maybe it’s temporary. Maybe you’ll shift back once the funding market returns. But eventually, every biotech organization is going to have to decide: Are you going to make the fundamental changes that will maximize your chances of thriving in the new world? Or keep making the comfortable decisions that will anchor you to the old one?