Site icon IT World Canada

3 business challenges faced when adopting big data analytics on an open source platform

As Steve Holder, national practice lead for analytics for SAS Canada explains the points on his “open source integration” slide, he notes he selected a starry background after being inspired by the lunar eclipse event earlier in the week.

For many in Toronto, the eclipse was obscured by an impenetrable layer of cloud. The night sky was blacker than usual for star gazers that hoped to catch the rare celestial event. Perhaps as Holder looked up at those clouds blocking his view, they reminded him of the way a non-flexible data archive can obscure information from a business. Open source integration however, would clear the proverbial skies and reveal the objective.

At the SAS Live: Analytics event hosted by the analytics software vendor and open source contributor Hortonworks, Holder made his case for running SAS Analytics on top of the Hadoop data platform architected by Hortonworks.

“A lot of organizations are struggling with data, figuring out how to make it the right size, the right shape, the right depth,” he said. “Data is getting so big you can’t run it off the side of your desk anymore.”

The advantage offered by SAS is its predictive capability, he continued. Data architects within business teams can apply modeling to data stored on a Hadoop system in such a way that it’s possible to stream analytics in real time. It moves beyond just moving data over to file storage and into the realm of advising on what to do next. That’s what SAS refers to as “next best action.”

Hortonworks and SAS are working together to serve clients in the financial services, telecom, retail, and manufacturing verticals. While executives from both companies highlighted the benefits of their product, they also made clear the complicated nature of shifting legacy systems to a new model. Along with customers, three main business challenges were illuminated during the event.

1. Do you have a big data problem?

“At what volume do I have a big data challenge?” is a question that Vamsi Chemitiganti, general manager for financial services at Hortonworks often hears. The answer? “It depends.

“I define volume at the scale at which your internal infrastructure becomes brittle,” he says. “Start with a business problem and work your way back through the technology.”

Solving a specific business problem with big data analytics capabilities should serve as the starting point on assessing technology, Chemitiganti says. He points to a survey of 183 bankers worldwide that was conducted in the last couple of years. It shows that 76 per cent of bankers say their business driver for adopting big data is to enhance customer engagement, retention and loyalty. Bankers also believe that better understanding customers with big data will help them make more money.

Just as Amazon.com knows all the preferences and interests of its customers as soon as they log on to the site, banks want to try and guide their customers through online services rather than leaving them to flounder. Especially now that customers are interacting with banks via several channels – online, mobile, the call centre, and in branch – tracking the same customer across those touchpoints is more difficult.

“A lot of banks think this is about slapping a mobile front end on to your existing infrastructure,” Chemitiganti says. “It is not.”

2. Start small and build your team as you go

When Chris Dingle came on as the director of customer intelligence at Rogers Communications Inc. in 2013 he got involved with the company’s SAS and Hortonworks project. Rogers had been a 25-year customer of SAS and it was about to deploy the first implementation of SAS on Hortonworks for the media side of its business.

Dingle’s team started off small – just two people – and went about putting together recruiting people that could help prove the business case for a wider implementation.

“It took a little bit of time,” he told the audience. “You need the ability to listen to people and to work across different teams… it’s a team sport when you’re doing this right.”

Today Rogers has about 50 people across multiple teams working on the project. Scaling up that sort of personnel was done with some training and support provided by SAS along the way, he says. Rogers participated in a program that gave his staff direct interaction with SAS experts and helped them get excited about deploying it.

3. Get the business and IT working together

Bob Messier, the senior director of business analytics at SAS, says the company makes a point of ensuring its customers are ready to have healthy collaboration between the IT department and its lines of business. An online questionnaire is required to be filled out by new customers.

“As you go on that journey of modernization, there tends to be a lot of tension between business and IT,” he says. “It tends to help flush that out.”

After that, the next step is to sit down for a meeting with both the business leader and the IT leader, he says. It begins the discussion about the cultural shift that needs to happen if the project is going to work.

It might sound like a lot of work and negotiation, but if you can clear up those clouds and get one good clean shot of the moon, it might be worth it.

 

 

Exit mobile version