To Succeed with Data, Invest in Your Company Culture
By Erika Janowicz, VP, Chief Data Scientist at WestCap
When I first traveled to Europe as a kid, I was in awe of the cultural differences. In Spain, people ate later, napped in the afternoons and spent more time with extended family. I realized at that moment there were other ways to look at the world. Whether it’s delightful or uncomfortable, a shift in perspective yields rich results.
When launching data-driven initiatives, most companies focus on just one point of view: the technological. Data science and analytics projects are seen as the purview of IT, writing code and launching applications in a black box. But data initiatives that fixate solely on the technical are doomed, because they lack the perspectives and buy-in of key stakeholders. In order to be successful, organizations should place equal weight on building a strong data culture.
Why Data Science Isn’t Enough
As a growing number of companies embark on data initiatives, it makes sense that they tend to prioritize technology. Business leaders want ROI, and it’s hard to quantify returns on something as intangible as culture. Siloed teams have been the norm, so changing course requires overcoming institutional inertia.
But overcoming these obstacles is worth it. When tech teams undertake data initiatives in isolation, they lose the opportunity to innovate based on diverse perspectives. Without input from other teams, they may lack an understanding of real business processes, customer needs and organizational goals. That can lead to projects that produce inaccurate results or fail to achieve concrete business objectives. Initiatives fail; time, effort and resources are wasted. And data teams who don’t feel that their work is making an impact ultimately struggle with low morale and high turnover.
How to Build a Strong Data Culture
In order for data initiatives to succeed, companies need to invest in a robust data people culture alongside technological innovation. Here are five steps organizations can take to move in the right direction:
1. Enable collaboration across teams
Instead of operating in silos, departments across the company need to work together to define and shape data initiatives. Create opportunities for stakeholders in different business units — from marketing to HR to finance to communications — to share key priorities in order to inform projects that can support those directly. Facilitate channels for input along the way to keep everyone aligned and excited about initiatives. Set the tone from above that the company values continuous learning and keeping an open mind.
2. Focus on end results
When it comes to data initiatives, many organizations pursue too many ideas at once. A scattered approach can work in good times, but when funding gets tight, the lack of focus can pull companies under. Initiatives should select narrow use cases with concrete business outcomes that support the company’s core mission. Consider how any project will ultimately impact customers. It’s not about chasing shiny objects but solving real problems based on input from users.
3. Break the black box
Many people in non-technical roles assume that data science models operate as a black box: Team members run a request, something magical happens and they get a result. The internal logic is deemed too complicated to understand. In fact, much of the logic used for forecasting can and should be discussed with non-technical team members to ensure that it’s accurate and useful.
4. Democratize communication about data
Demystify data initiatives by describing them in terms that a broad audience can understand. Emphasize outcomes, not algorithms. People don’t want to hear about the intricacies of your natural language processing model; they want to understand how it can improve customer engagement. Data leaders should also set clear expectations about project scope and timelines.
