You don't need a platform to be data driven
*** A quick request: I’m starting to explore ways I can specifically help folks who have recently moved from Tech into Biotech. If this describes you and you’re willing to discuss your experience with me, please send me an email (jesse@merelogic.net) or reach out through Bits in Bio slack. ***
Last week, I wrote about how decision makers (in biotech as well as elsewhere) often conflate safe decisions with comfortable decisions. This week, I want to try and disentangle two other concepts that are often conflated: Data-driven biotechs and Platform biotechs. Here are my definitions:
A platform biotech is an organization whose strategy is to delay pursuing their first new drug (or diagnostics or whatever the product is) in order to build a machine that will (eventually) churn them out much faster than if they had put all their eggs in the first basket. These are the biotech version of Lincoln sharpening his axe.
A data-driven biotech is one that deliberately invests in processes and infrastructure that will allow it to use data/AI/ML/etc. over intuition to speed up or otherwise improve whatever outcome they’re pursuing.
Both of these are investments in the sense that they’re a deliberate decisions to take on additional risk in exchange for a higher potential reward. Any research-stage biotech is taking on risk by exploring new, unproven science. But a platform biotech takes on additional risk by delaying the validation of that core scientific risk. A data-driven biotech takes on additional risk by investing resources that could otherwise go to core science into data and data infrastructure.
This may seem like an unnecessarily subtle distinction to make, particularly for those of you who are mostly here for my rants about metadata, but bear with me.
Right now, the elephant in the room when biotech leaders think about how to invest their limited budgets is the fact that it’s much harder to raise VC funding today than it was a couple of years ago. There are a lot of reasons for that, including interest rates and the counter-reaction to the surge of funding during the pandemic. But there’s also the fact that investors have become much more skeptical of the platform strategy. They’re looking for startups that can get to the clinic as quickly as possible.
In times of uncertainty, everyone wants to minimize risk. And the risk associated with deliberately putting off validation seems like an easy one to drop.
This make it very difficult for startups to continue with the platform strategy. But it also means that even non-platform biotechs are generally interested in reducing anything that feels platformy, or doesn’t obviously get them to the clinic faster.
And one big thing that often falls into that bucket is investment in being data driven.
Decision makers associate the data-driven strategy with the platform strategy because this data/AI/ML in biotech has historically involved a lot of early-stage and speculative development requiring the kinds of long-term investments that basically force you into a platform approach. So many (most? all?) of the early example of data-driven or AI/ML-driven startups were platform biotechs.
But does it have to be like that today?
A lot of the technology around data, AI and ML is still at a point that requires some serious investment - particularly the LLMs and generative models that have only made it into the mainstream in the last few years. But there’s also a lot that has already been developed to the point that it can generate value today. This includes a lot of the mundane things that make data more accessible to decision makers even without AI/ML.
So if your goal is to build a data driven biotech, you basically have two choices: 1) Wait for the market to change and join a platform biotech that’s also wants to be data driven or 2) Start making the case for why the data tools that are already available can be used to get drugs to the clinic faster today.
Which one are you doing?