FORTUNE 50 RETAILER SEES OPPORTUNITY TO IMPROVE CUSTOMER EXPERIENCE WITH DC/OS

“Our team implemented a Mesosphere cluster to help the retailer take advantage of DC/OS. This reduced the time to validate algorithms and analyze data from weeks down to minutes.

– Dave Rollo, Strategic Account Manager at Shadow-Soft

Challenge

A Fortune 50 retailer saw an opportunity to dramatically improve customer experience, deliver real-time targeted sales, and improve logistics. The team needed to shorten development and analytics cycles down to minutes. The starting point was an abysmal six weeks to deliver results on highly perishable data – a world away from the organization’s stated vision. It took months to deliver results to a very frustrated data science team.

The retailer also faced a TCO challenge. To meet the new requirements with traditional infrastructure, it would require tremendous compute power to service the algorithms in burst (and then sit idle while not being used).

Solution

Implemented a Mesosphere cluster to take full advantage of the Data Center Operating System (DCOS) technology and create an agile platform. The Shadow-Soft team created standardized containers for analytics programming languages and tools (i.e., R container, Python, and Cassandra). Easy access to data sources was made available to data scientists so they could act fast.

In parallel, a self-service capability was developed with role-based access and security. Teams can now select from a library of containers and data sources and have a working environment ready in minutes instead of weeks. The Mesosphere solution helped the retailer maximize resources at hand and reduce the future spend on infrastructure.

Outcome

Reduced the time to validate algorithms and analyze data down from six weeks to 10 hours. Eventually, this time will be reduced to just minutes.

The data science team can now analyze data faster and pass off insights to a marketing team, who uses the data to send personalized offers to customers via digital and traditional methods.

Request info on real-time analytics