Big data demystified

By Divina Paredes | Thursday, October 11 2012

A ‘solid definition’ of big data? When the amount of information in your organisation ‘puts you out of your comfort zone’, says Paul Kent of SAS.

He cites the case of a grocery store where an HR staff member analysed the employee roll and discovered a large percentage of them would come back to work at a certain time every year. This data influenced the onboarding process for these seasonal workers who are welcomed as existing employees and do not “have to go through and learn all the rules and regulations over and over again”.

He says while improvement through the use of data does not necessarily turn into more profit for the company, “usually this is a side effect of doing a better job on anything”. In this case, the company saved on training cost and lifted employee morale.

Kent says another simple application of big data was demonstrated by a rental business which uses elevator logs to observe which of their customers might be less able to pay. This is done by aggregating the data of the number of people stopping on each floor. He says this “unstructured data” of elevator logs was previously unused and kept in a data warehouse. The company found that when there were fewer people going to an office, that company may need a smaller space, or will be unable to pay the bill.

The lessons from all this is to “think differently” and for companies to modernise their computer resources, says Kent.

“Think of ways to use existing data in more complex ways,” he says. “Have you kept the door open to capture new data sets and make something out of it?”

Big data analytics can also use existing data in more complex and exciting ways. A mobile phone provider in Europe, for instance, looks at the network of people a customer interacts with to aid in its assessment of customer churn. The model takes into account calling patterns and whether the customer’s network of friends are changing mobile phone carriers. “You target more marketing dollars at them or listen to them with higher volumes,” he says.

Analytics in the forefront
Another SAS executive, Nelle Schantz, discusses how analytics is helping the New Zealand Automobile Association lead a more directed marketing campaign for its insurance group.

She says the group used to send out 400,000 solicitations a year “with minimum intelligence”. It now sends 240,000 solicitation letters, a reduction of 40 percent, but the response rate has doubled and increased customer satisfaction.

Analytics has also allowed Bank of America examine data to determine loan defaults on mortgage. They take calculations around loan default that used to take four days to four hours. This gives the bank ability to react to risk more quickly, says Schantz. She says proactive analytics is critical for the finance sector, which could be analyzing up to a billion rows of data. “You can do complex assessments more quickly,” she says.

Analytics has also provided Gilt Groupe a different insight into their customers’ experiences. Gilt Groupe runs “flash sales” for designer clothing and premium home accessories.

Gilt Groupe is active in the social media space so it is interested in mining customer comments and predicting customer churn, says Tamara Gruzbarg, senior director of analytics and research.

Customer experience is key, she says, of their core business model of providing a variety of “flash sales” in different categories. Each sale is live for a limited amount of time and for a limited inventory. She says the industry space for this type of sales is getting “very crowded” as more companies including department stores are offering similar services. “How do you provide the best customer experience and differentiate us from everybody else?”
Right now, they hold 50 new sales everyday. “We need to show you something that is relevant to you and prompt you to buy,” she says.

The group also found that the customers who return the most purchases are also the most loyal. “It doesn’t make sense,”says Gruzbarg the first time she reported this to her business colleagues. But she says their industry is so specific so when customers “start feeling comfortable [dealing with them], they can take a gamble on a purchase and if they don’t like it, they can always return it and come back and purchase some more”.

She says it is important to encourage everybody to innovate and ask questions. Questions from engineers “are as important as those coming from the senior vice president”.

Kimberly Holmes, senior vice president of strategic analytics at XL Group, says it is important to look at data broadly. “There is a lot of data that is not in numbers, we need to think about it in new says and take advantage of new technology,” says Holmes.

She says while unstructured data will not "overwhelm" business data at XL Group, “it is just as powerful as structured data”.

She says one of the “curses of the job” is that analytics can help with every business problem. These can range from decisions on how to hire, who to promote, how to advertise and manage distribution. “I can’t think of a question that can’t be answered using analytics.”

“We have enough work lined up for us in the next two years.”

Divina Paredes (@divinap) is editor of CIO New Zealand.


Follow CIO on
Twitter @cio_nz
Facebook
LinkedIn

Download CIO for your tablet here.

Click here to subscribe to CIO.

Sign up to receive free CIO newsletters.

Send news tips to divina@cio.co.nz