Editor’s note: Jesse Johnson is a speaker for ODSC East 2023. Be sure to check out his talk, Development Principles for Biotech Data Teams, there!

One of the biggest issues that data teams embedded in non-tech organizations face is not technical, but operational: finding an effective working model.

In a traditional technical team’s working model, the developer’s role is to understand and address their client’s needs. But as embedded data teams are increasingly expected to not just support the organization but to drive its roadmap, they find themselves relying on their colleagues to generate data and make decisions that impact their digital strategy. So it’s not enough to support their colleagues; data teams need their colleagues to support them as well.

In the biotech space, where this is particularly true, the Reciprocal Development Principles were designed to help embedded data teams create this kind of working model. The principles build off the core ideas of the Agile manifesto, but update them for this new context. In particular, they help teams build a foundation with three layers: Scientific objectives, deliberate communication, and broader iteration.

Scientific Objectives

One of the biggest ways that data teams limit their ability to drive their organization’s roadmap is by limiting themselves to technical rather than organization-level goals, which in biotech means scientific objectives. For example, if a data scientist delivers predictions on schedule with reasonable accuracy, they may consider their job done. But to drive the organization, they should also be involved in, if not accountable for, how the predictions are used to make decisions and drive value downstream.

Even infrastructure projects that seem far removed from organizational objectives will have a definable downstream impact if you look hard enough. Speeding up a process or enabling it to scale allows tools and capabilities to be used in new ways that magnify their impact. Calling out these impacts and objectives changes the way the team views its work. (If you can’t find one, maybe you shouldn’t do the project.) But more importantly, it gives you more confidence to ask for the things that you need from the rest of the organization.

Deliberate Communication

When working in a developer/client model, developers will typically spend just enough time with their clients to understand their needs, or even delegate some of this communication to a product manager. Outside of that, they’ll carefully guard their focus time. Driving your organization’s roadmap, on the other hand, requires shifting habits and mindsets outside your data team. And that inherently requires that everyone on your team spend more time and energy on communication.

It’s still important to guard your team’s focus time, but in this new context, you need to balance that with building a deeper understanding of where and how your colleagues are willing to shift. And this requires a new set of skills and tactics, particularly deliberate empathy – the habit of looking for the underlying constraints, assumptions, and priorities when your colleagues resist or push back in unexpected ways. These skills will allow your team to begin shifting the organization in new and important ways.

Broader Iteration

The one thing I haven’t mentioned yet is the concept that people most associate with Agile: fast iteration cycles. This is just as important in Reciprocal Development, but the iteration needs to include both the technical components and the processes around them. The key idea here is that bad software is a symptom not a cause. In other words, if a team is having problems that seem to be related to software, there’s usually a deeper problem around habits, processes and how the team conceptualizes their work. So if all you do is introduce better software, the team’s processes and mindset will ensure that the team either uses it in a way that supports the old habits, or doesn’t use it at all.

You can’t fix the processes without better software tools, but you also can’t change the tools without changing the processes. So you need to iterate on both simultaneously: work with your colleague teams to make small changes to processes that create space for better tools. Introduce or update the tools that support the new process. Then repeat. Note that this is more than just driving adoption; we’re deliberately changing how other teams in your organization work and conceptualize that work.

Conclusion

The Reciprocal Development Principles are a list of specific principles that will help you and your data team build out these three layers of foundation and begin shifting your role from supporting your colleagues to driving the larger organization. To learn more, you can read about them here or check out the Scaling Biotech newsletter where I’ve been exploring the concepts in more depth.

I’ll also be presenting an overview of the principles at ODSC East in Boston, May 9-11th 2023, in my session titled Development Principles for Biotech Data Teams. I’ll also be giving away copies of my upcoming O’Reilly Report “Leading Biotech Data Teams.”

About the author:

Jesse Johnson is Vice President of Data Science and Data Engineering at Dewpoint Therapeutics, a drug development Biotech startup founded in 2019 around a scientific field called biomolecular condensates. In this role, Jesse’s diverse set of experiences from academic math departments, engineering teams at Google, and data science teams at large, medium, and small life science companies provide a unique perspective on the ways that data and wet lab teams communicate differently, or sometimes don’t communicate at all.