The data science dilemma: Should marketers build, buy, or both?

Data science, artificial intelligence, machine learning and algorithms have become the business buzzwords du jour. And with good reason. Organizations adopting data science-driven automation see up to 10x improvement in business process speed, 72 percent reduction in customer churn, and 40 percent reduction in new customer acquisition costs.

It’s the so-called sexiest field of the 21st century as both enabling technologies and brands are fighting for the top talent that will allow them to modernize and scale their marketing efforts. But much like the big data that fuels the practice, data science for marketing is challenging to get a grip on.

Having a data science team doesn’t mean that marketing gets to take advantage of the expertise. Macro level insights and micro level optimization can’t always both be a focus.

What should be built? What should be bought and managed? When do you combine building and buying?

If you’re going to build a data science team and data products, be sure you’re using your resources to build something that’s truly valuable and differentiating. And be sure you can continuously support the products and innovate. It’s not enough to set it and forget it — your data products and the algorithms and models that serve as their foundation must be updated as they are used.

Take Stitch Fix for example. They’ve built an 80+ person data science team because data science is critical to their total product and competitive differentiation. This is a company where success relies on machine learning and algorithmic product selection based on multiple levers — wisdom of the crowd, predictive analytics, and both explicit and implicit data — and where data science is critical to business processes across the organization, including supporting marketing.

Every modern marketing organization — and company for that matter — is data-driven. But that doesn’t mean that data science is core to the business model itself. When you’re looking to apply data science to marketing or business functions, there are often technology partners that have already built the solution.

RevZilla is the nation’s leading retailer for motorcycle jackets and gear. The founders were frustrated with the lack of customer-centric experiences for motorcycle enthusiasts — they didn’t just want to shop on price, they wanted to buy from retailers who shared their passion and would advise, not just sell. Instead of waiting for this to happen, they created RevZilla.

Every individual in RevZilla’s audience has unique preferences for riding. From sport touring, adventure, and even racing, the gear needs for their buyers are incredibly wide-ranging. While the RevZilla team maintains a “local bike shop” feel in every store visit, phone call, or customer service email, the team wanted to scale that approach to their total marketing program.