Placing Data Scientists Within Your Organization
There are a number of different ways in which Data Science (DS) teams can be structured, but if your organization has chosen not to use an Embedded team model (where Data Scientists are assigned and report up to business unit leadership), then the question of where in the organization your DS team should report remains a pertinent one. This is especially tricky because DS straddles the border between technology, operations, and sometimes even marketing or finance, and needs to be able to move easily between each of these realms.
Ultimately, however, in every organization, your DS team needs to be accounted for as both a cost base and resource. Under whom should they report?
Parsing Your Choices
As with so many things in the DS world, you’ve got some choices to make when it comes this question, and each one has its own trade-offs.
Technology – CTO
Having your Data Science team report within the Technology component of your organization makes sense on quite a few levels. After all, your DS team will be working very closely with cutting edge technologies, seeking to push the limits of the infrastructure in place to produce better and more applicable results. Additionally, as with all software development processes, it is important for business owners to collaborate closely with their technologically-minded counterparts. In the sense that productionalizing insights will require business partnership, situating your proponents within the technology team will facilitate that goal, especially if you have a team rich in Hackers.
Placing your DS team within your technology organization, however, might not be a great idea if your organization has strong silos between business and technology. If the technology team is commonly referred to in a monolithic fashion (i.e., “Technology’s working on it, but I don’t know when they’ll be ready”), these silos will hamper the crucial ability of your DS team to collaborate with the business, and close off other business units from coming to them for help.
Operations – COO
If your organization is production-oriented, and you expect your DS team to focus heavily on producing solutions that optimize process efficiency, subsuming your DS team within your operations organization might be the right choice. In doing so, you will increase their exposure to the everyday operations of the business unit, which helps develop domain knowledge that can be crucial in solving problems.
At the same time, providing visibility and accessibility for operations staff helps cross-pollinate ideas. Perhaps one of your junior operations staff has the start of a great idea, but they might otherwise be too intimidated to reach across lines to share it. Situating your DS team within the operations organization lowers these barriers and promotes the open exchange of information and ideas.
One thing to watch out for when placing Data Scientists within the operations team are the tendency of operations teams to implement under-the-radar workarounds to business problems that might otherwise mandate a technological solution. An environment in which this is regularly tolerated in lieu of open discussion with technology teams may isolate your DS team and reduce their effectiveness. Placing them under operations may also limit your individual DS staff’s ability to develop technological sophistication in line with their long term goals.
Marketing – CMO
Many venture-backed startups hire “Growth Hackers” for the purpose of kickstarting the explosive growth necessary to achieve market traction and return positive results for early stage investors. Employing one or more Data Scientists or even a whole team for this purpose is certainly option, especially if the operational and technological aspects of the product are otherwise simple (think Snapchat). Bringing traditional academic credentials into this field can yield positive results through initiatives such as psychographic market segmentation and studying the impacts of gamification on user experience and interface design.
If the biggest problem for your organization is growing it quickly, then placing your DS team within marketing may be the right solution. However, bear in mind that this structure typically requires a fairly specific skillset. You may be more likely to attract Academics and Quants, which can lead to unbalanced teams and reduce the breadth of problems your DS team is able to solve, and the ease with which they may be implemented in technology.
Finance – CFO
Conversely, if your organization seeks efficiency through financial structuring, as may be the case with banks, placing a DS team within Finance may be the solution. Traditionally, within larger banks, this role might be played by a team of accountants under the controller. However, if you happen to work within an organization willing to experiment with this structure (off the top of my head, I can’t think of any… but I’m sure they’re out there), the same role might be augmented with a few Academics and Quants, and within the scope of conventional finance the lack of technology acumen on the team might not manifest as a gap in the skillset.
One thing to watch for when using this structure, however, is that innovative financial engineering often leads to innovative financial regulation. Staying abreast of regulations and having a strong legal and compliance review process is crucial for ensuring this team of people tasked with toeing the line of possibility don’t pass it entirely.
Making Your Own Team – CIO (Chief Information Officer) or CDO (Chief Data Officer).
The industry refers to the “I” in CIO as interchangeably “Information” and “Infrastructure”. For these purposes, we’ll be looking at the “Information” aspect. What if you made your DS team their own executive business unit, which reports directly to the CEO? You would then incorporate them under either a CIO or CDO. To some degree, this is the purest implementation of the discrete Data Science team, which seeks to treat “Data Science as a Service”. Business units would bring their problems to the DS team, who would then work together to solve them.
One interesting advantage of this model is that it establishes a growth path for analysts and analytically-minded staff in other departments. Rather than rising to the top of the analytical organization within their departments and then moving laterally between groups, joining a discrete, cross-functional DS team provides an excellent growth point and an opportunity to develop the talent already within your organization. This also helps infuse your DS team with crucial domain expertise, as these folks have experience on the front lines of your business.
On the other hand, it is worth noting that if your organization is small, or incredibly busy or bootstrapped, incorporating a whole other organization just for data may be more work than it’s worth. As with all innovative business structures, be prepared to build consensus and buy-in from other executives who may see this as an intrusion on their sphere of influence.
Making Your Choice
Given that Data Science is such an amorphous, flexible, and high-potential field of work, it is crucial to provide the “right to win” for your team. One closing remark to consider is that, while it may seem logical to strive to place your DS team within the organization that makes the best use of their skills, it may also be worth making the choice tactically – situate your team within the group that is most open to new ideas. Perhaps it doesn’t make sense at first, but once your team investigates the concept more thoroughly, it may turn out to be a great one.
If you haven’t already, read some of the other pieces that I have written on the topic of managing data scientists, or get in touch if you’re interested in exchanging some ideas.