Four Secrets to Creating a First-Rate Data Science Team

As companies of every shape and size pour money and resources into data science and business analytics, it’s no surprise that these organizations are aggressively recruiting data scientists to gain a competitive edge. In fact, data scientists are in such high demand that the profession was dubbed America’s hottest job.

But realizing a need for a robust data science team and building one are two very different things. At Honeywell Aerospace I’ve had the challenge and pleasure of building a data science and analytics team of close to 100 people over the past two years, and I’ve learned a huge amount about what it takes to build a team that delivers results.

Our data science team helped airlines reduce maintenance-related delays by up to 60 percent, with 99 percent accuracy in detecting problems. Going a step beyond predicting component failure, our analytics team also provided prescriptive recommendations on part replacements and repair procedures that prevented scheduled disruptions. We’ve celebrated some successes along the way, but we’re just scratching the surface of what’s possible.

As companies grapple with how to successfully organize and mobilize around analytics, I’m sharing my top four secrets to create a distinguished data science team.

1. Align with business priorities

 Particularly for companies that are new to the realm of data science and advanced data analytics, there is nothing more important than proving your team’s value from the get-go. Start the group off small, focus on delivering a healthy return-on-investment, and grow the team to match the value demonstrated.

 As that team grows, it’s critical that analytics initiatives are aligned to top business priorities. Conduct workshops with business leaders to develop a list of high-value analytics projects and ideas to support the strategic priorities of the company. Once you have these analytics roadmaps in place, you’re off and running in the right direction.

 2. Identify executive champions

 The reality is that in many large organizations – even at some of the best tech companies in the world – a small percentage of employees truly understand what data scientists do, and how they do it. The key is driving a clear understanding of needs and potential impact with decision makers by working directly with top executives.

 The responsibility of providing that education lies with the data science team. Simply and effectively coaching executives on the value of data science will help them advocate for necessary resources to support the team.

It’s critical to identify executives who would best serve as your team’s champions. Look for first adopters, risk takers and the kind of people ready to use new technology to drive transformative change. Oftentimes in large organizations these are product line or functional business leaders, with big, seemingly unachievable goals in front of them who are ready to support and trial new ideas to drive breakthroughs.

3. Build the correct organizational structure

 Focus on building unique expertise to solve complex problems vs. building generic analytics capabilities. One way to accomplish this is through a flat team structure. Avoiding silos lets the team of experts work better together, and encourages them to challenge the status quo to push toward continuous improvement.

With my team in the Aerospace business, we focused on hiring data experts with aircraft data sets and data scientists with a background combining aerospace engineering, prognostics development and advanced industrial data science. In addition, we hired people with experience in business analytics, statistical modeling, economics, actuarial sciences and more. Building the right layers of expertise helped us quickly develop expert capabilities that improved the business.  

4. Establish processes from the start

 Accelerating the time it takes for your team to create value is critical. Developing a proper way to conduct idea generation, project intake, qualification and prioritization sets the team up for success. In addition, identifying specific metrics is necessary to measure analytics progress. From the start, it’s important to establish standard processes or project intake and qualification of analytics opportunities. This will ensure the team prioritizes and executes the right project opportunities.

Of course, evaluating processes on a regular basis is important so the team can adjust as needed to keep pace or stay ahead of changing priorities. What works for a small team in one office will not work for a large global team, so there are organizational, cultural and geographical nuances to consider.

Ready to begin?

 As you work to get your team in place, running efficiently and delivering impactful results, remember that analytics is a team sport. While the data science team creates the insights, the business takes the action to implement those insights and create value for its customers. A strong partnership with the data science team and business stakeholders reveals the true value of this initiative. When done well, a new challenge emerges – growing your team to share even more insights to further an organization’s competitive edge.