Getting stuck at proof-of-concept
One of the most frustrating problems that I’ve seen biotech data teams run into is building a model or an analysis capability, then watching it sit on the shelf, not getting used. Maybe you get a publication or a press release out of it, sometimes an internal presentation or two. But actual contributions to scientific progress are often much more elusive.
I’m not talking about analysis that the bench team asks for. For those projects, there’s already a spot in the pipeline for it to fit into, and the support to make it happen. The projects that get stuck are the ones that introduce a new capability the bench teams didn’t think was possible. Heavy on AI/ML. Using data in novel ways. Precisely the things that AI-enabled biotechs are supposed to be doing.
For a biotech data team to move beyond just supporting the bench teams, and actually drive the science, they need to push past these blocks. And to do this, they need to address three main reasons they tend to stall:
The problem you’re solving can’t be acted on - It’s interesting, important, even relevant. But at the end of the day if it doesn’t help your bench teams make a decision, your project won’t make it past proof-of-concept.
You keep missing the bus - Maybe the lab team isn’t ready to try it in an actual experiment. Maybe you want to be a bit more sure before you ask them to. Or maybe the timing just hasn’t worked out. For whatever reason, it never seems to happen.
The data isn’t available or isn’t consistent - You spend half your time tracking down metadata, and most of the rest cleaning up excel files. By the time you’re ready to run the analysis, the bench team team has moved on.
I’ve explored these issues in various ways in the past, particularly with the Reciprocal Development Principles. (Download the book today!) But in the next few weeks I want to revisit them, from the specific perspective of how you can overcome them to push projects from proof-of-concept to actual impact. Stay tuned!
*** And now the ad: If you have ML projects that are stuck at proof-of-concept, Merelogic consulting can help. Sometimes it’s a few small tweaks to how your team operates. Sometimes it’s larger changes to your tools, infrastructure and projects. If you want to explore what this might look like for your team, send me an email at email@example.com . ***
Thanks for reading Scaling Biotech! Subscribe for free to receive new posts and support my work.