Get Started

How Many Organizations Are Led by “Data People”?

Scaling AI, Featured Shaun McGirr

In our previous article on building a strong data culture, we outlined an ambitious agenda for change — recognize how the company makes decisions and what needs to be changed, embed business translators and data product managers into the organization, and build an ideas factory instead of becoming too attached to early experiments or models. 

Here, we’ll take those ideas a step further to identify who can be the change agent to put that agenda into practice and the leadership qualities they will need to actually get it done. Warning to CDOs and data executives: these tips may encourage you to step outside your comfort zone and act differently to increase the likelihood of these changes becoming reality (and, spoiler alert, you will end up thinking like an organizational leader and not necessarily a data leader!).

leader

1. Ruthlessness About Your Own Limits

As challenging as it may be, data leaders can’t get too attached to ideas that aren’t sticking, especially their own. Instead of trying to generate ideas (and eventually projects, data, and models) within the data team, setting up the conditions for an ideas factory can bring others into the fold, help decide what’s valuable, and instill a feedback loop with the business. To be sure, we don’t just mean any ideas — the phrase ideas factory indicates a larger scale, but the only way to achieve this is if the business is generating ideas from real problems to solve. That can go against every fiber in a data leader’s body, who has possibly delivered value in the past by assembling (and protecting) experts from too much business influence. 

If a data leader is too protective about their team, models, or roadmaps and dwells in the past instead of figuring out how to change the future, step three of our prior article is going to be especially difficult. If, instead, the data leader puts on the cap of an organizational leader (e.g., making a tough decision to nix a precious use case that business stakeholders will not adopt), they will help their own teams achieve cultural affinity with the wider organization, changing everyone else a little by changing themselves a lot. By adopting the mindset of those they should aim to serve (i.e., members of the C-suite with P+L responsibilities), data leaders will better lead the organization through significant change, instead of being laser-focused on solely their projects and how to make everyone else just like them.

ideas

2. Comfort Relinquishing Control

It might feel counterintuitive at first, but data leaders need to tap in to other experts throughout various pockets of the organization to help implement a robust data strategy, and not just because they can’t be everywhere at once! While data leaders probably know exactly who to pick for business translators and data product managers, they need to accept these individuals likely won’t stay on their central data team forever — they might be moved around internally to serve the greater good. 

By pushing these expert capabilities outside of the central data team, the data experts are letting go of gatekeeping control, which might be disconcerting at first. But the reality is that they won’t fundamentally change how other people behave if they hold the keys to that change forever. The translators (who discover demand in the business and reflect it back to data teams) and data product owners (who deliver concrete value, co-sponsored by business and data teams) will add the most value for data leaders when they are “let go” — by escaping their own comfort zone they will ultimately find more business value, as long as they don’t feel a central gatekeeper is still calling the shots.

keys and control

3. Willingness to Accept Unexpected Sources of Value

It’s unrealistic to believe the centralized data team (only ever a small percentage of the company’s total headcount) will eventually convince every business user to join the “one true way” of the central data team. Instead, data leaders can use their capabilities to empower the business, influencing others to leverage data science, machine learning, and AI — sometimes building upon a project that originated within the data team, but not always. 

The data leader should be happy to be surprised, not dismayed, that business users are building off something themselves. Blocking that innovation, because it “wasn’t on their roadmap,” erodes goodwill, cuts off sources of business value, and taints a valuable internal talent pool for future recruiting to the central team. Yet, in many conversations with data leaders, we hear genuine reservations about making AI everyday, because of the age-old problem of any expert looking out at the masses: “They might reach the wrong conclusion and I will feel responsible for the consequences.”

This view is mistaken, and not just for its inherent elitism. In his book “The Difference: How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies,” author and professor Scott. E. Page says,

Progress and innovation may depend less on lone thinkers with enormous IQs than on diverse people working together and capitalizing on their individuality.”

Page demonstrates that groups with a range of perspectives tend to outperform groups of like-minded experts, on average — and establishing a data culture is the same kind of challenge. It cannot be delivered by a crusader alone! Think of how, almost out of nowhere, TikTok became “the new YouTube” simply by joining live filming with snappy editing. The underlying technology and ideas already existed, it was their adoption by a different set of people that created a cultural phenomenon.

If you believe that the best practices we outlined for building a strong data culture are solid, foundational steps to pursue, we encourage you to think critically about the leadership qualities necessary to move them from concept to practice — seeding an organizational change and helping it grow organically within the organization, accepting you can only prune and water. 

You May Also Like

How Does One Acre Fund Use Dataiku to Increase Its Operational Efficiency in East Africa?

Read More

You Won’t Scale AI Without Enlisting Non-Experts to the Cause

Read More

Bring New People and Roles Into AI Projects With Dataiku 10

Read More

3 Career Options for Data Scientists Who Want to Maintain Their Skills

Read More