Best Practices Summary:

• Leverage cloud platforms to bring data across siloes together and make the data available near real time for better performing data science models

• Use modern infrastructure in the cloud to help bring data science models from testing to production more efficiently

• Create a data science community to engage a broad reach of teams within the organization

Case Study:

A large retail company faced challenges delivering insights for marketing, promotions, operations, and other functions due to siloed data infrastructure with slow batch processes and a lack of ability to scale. Costs were high due both legacy infrastructure and the onerous process of gathering data for analytics and data science work. The difficulties interfered, for example, with the company’s ability to apply data science to personalize the customer experience on mobile, web, and store and thereby grow revenue and market share. 

To address the data challenges, the Director of Data Services and Platform Engineering helped create a cloud-based data exchange that brought formerly siloed customer, retail, and employee data together into a single data lake platform with standardized formats. The data could much more easily be served to the data science communities as many time-consuming, batch ingest processes were reduced. Formerly, external data scientists used batch data for simple customer segmentation that had limited personalization impact. With the cloud-based approach more customer-related data could be fed into internally managed models in near real time. The models were more quickly and effectively fine-tuned to create better targeted customer recommendations. The cloud platform also enabled a seamless process to bring data science models from testing into production. As more models generated more recommendations in the company’s mobile app customer experience sales increased and loyalty improved. 

The benefits of the cloud data platform weren’t restricted to customer engagement. The data exchange platform enabled standardized Internet of Things (IoT) data to be automatically gathered from equipment in stores via telemetry. The IoT data was then analyzed to inform equipment maintenance, optimize store labor allocations, and improve product inventory. These analyses saved costs while improving the customer experience. 

As the internal data science community grew, fueled by the data exchange, the company benefited from a broader reach as teams contributed to and leveraged the data for insights. Everyone became more interested and involved in the data. The data science community established an onboarding process, which enabled teams to self-serve the data. What started out as a need for personalized recommendations turned into a forcing function for more automation, which in turn empowered broader analytics insights across the data science community. 

The data exchange platform was very successful in driving business value through better customer recommendations, equipment maintenance, labor allocation, and product inventory management. At the same time, it fueled the data science community and drove greater adoption of data science and insights across the enterprise.