Computerworld

Red Hat Takes Customer Service to Next Level With Predictive Analytics

You might think that Red Hat's core business is software. But to company leaders, it's actually customer service. The better the support, the greater the competitive advantage in the open source market.

To take customer service to the next level, Red Hat wanted to apply predictive analytics to its large customer service data sets in order to make its support function more proactive.

Tom Mirc, senior manager of business systems at Red Hat, was charged with overhauling Red Hat's customer service portal. But his team had its hands full with running the Red Hat's systems -- applications support, operations management, bug fixes -- and had little time for rethinking them.

"I have a very highly qualified development team, but they were bogged down in operational response," Mirc explains. "There was no way I could pull them out of their current roles."

Red Hat Chooses Catalyst IT Services as Outsourcing Partner

So Mirc looked outside the company for help. He had a few requirements for an outsourcing partner. They had to be capable of doing exploratory and leading edge development work. They needed to understand how customers interacted with current systems. And they had to be able to hit the ground running.

That led Mirc to Catalyst IT Services, a Baltimore, Md.-based agile development provider. "We looked at a variety of staff augmentation options as well as the rural outsourcing market," says Mirc. "Catalyst had a unique value proposition; they could find the talent we needed and mobilze them on an ultra-aggressive time frame."

Raleigh, N.C.-based Red Hat had built its own offshore teams in Pune, India, and Brno, Czech Republic, to handle IT operations and application development. But Mirc wanted a team closer to home. He had partnered with a rural IT service provider to assess the company's quality assurance infrastructure, but knew they weren't capable of a scalable Hadoop deployment.

Catalyst Helps Red Hat Improve Cusotmer Service with 'Subscriber Stickiness'

Catalyst got to work building what Mirc calls "subscriber stickiness" into the portal (itself an amalgamation 18 different application integrated via services-oriented architecture) -- from enabling customer to integrate directly with back-end APIs to serving them customized content to building in automated diagnostic capabilities.

They were able to identify customers at an account and user level and begin to analyze their patterns of behavior. The goal was to provide customers with answers to problems they didn't even know they had yet.

"We want our customers to understand there are many attributes of their support subscription. They may think of Red Hat support as just an insurance policy when something goes horribly wrong," Mirc says. "It's more than that. We want to give them the ability to proactively solve their issues or anticipate them."

Using additional behavioral data from Omniture and Eloqua as well as the portal itself, Red Hat also wants to determine where customers are in their lifecycle in order to market more effectively to them.

While Mirc's goal is to transition more of his development staff to more strategic work, he insists they could not have made the shift to big data analytics support as quickly as the business needed them to.

Catalyst's developers "were able to provide that innovation support for our business customers allowing my team to maintain its operations role and [eventually] migrate to more strategic roles," Mirc says. "They were able to quickly conceptualize business objectives and were technically versatile enough to deliver on them in a span of weeks, not months." In fact, Mirc had promised the business a scalable proof of concept within six weeks.

The biggest challenge was getting internal back office administrators to meet that pace; it took longer than expected to get the outsourced team the Linux laptops and credentials they needed to get going. If he had to do it over, Mirc says, he would have started working with procurement and facilities folks a few weeks earlier. But they made up for lost time by integrating Catalyst directly with Mirc's own agile development teams and processes.

"Even if they were not here in person, they were a daily presence in our scrum meetings, interacting directly with our personnel on our communication channels, and had access to the data they needed," says Mirc. "It was a different way of working but it's an extension of our domestic sourcing approach."

Stephanie Overby is regular contributor to CIO.com's IT Outsourcing section. Follow everything from CIO.com on Twitter @CIOonline, Facebook, Google + and LinkedIn.

Read more about it organization in CIO's IT Organization Drilldown.