It’s difficult to get your head around big data. In fact, it might be impossible, even with today’s superfast computer processors and scores of scary-smart programmers trying to do just that.
“I tell people, ‘You think you know how big big data is until you see bigger data,’” said Casey Kinsey, the president and founder of Lofty Labs, a Web development consultancy company in Fayetteville. “There’s always some bigger set of data.”
The Northwest Arkansas Technology Council recently held a panel discussion on big data, and Kinsey was one of the featured panelists. The panel discussion was about how companies can use information gleaned from the production of so much data in today’s world.
Big data has always existed, but slow computers — computers fast enough to analyze millions and millions of units of data were prohibitively expensive — prevented a lot of over-the-counter analysis of the data. Now, insight gained from big data is a fact of business life.
“Every time you visit a website you have some cookies stored that have 10 to 20 data points about you,” Kinsey said. “That data can be analyzed. It’s all limited by how much data you can crunch through a computer. In a decade I would imagine we will be able to handle a vast amount of data.”
Banking on Big Data
Properly analyzing big data information is big business for the banking and finance industries. A lot of financial information is, of course, cold hard numbers. But in today’s social media world, there are all kinds of electronic data — what Rick McGraw calls “unstructured data” such as Twitter postings and blog entries.
McGraw, the CEO of Black Oak Analytics in Little Rock, said so many transactions now are done remotely that face-to-face interactions with customers are becoming rarer. That makes understanding what customers think — through their social media proclamations — all the more important.
“Previously, the banking systems have been designed to handle structured data where there are very precise columns and rows and everything lines up very nicely,” McGraw said. “How do you know how best to serve your customer? You do it by analyzing the data that you have available to you.”
Todd Greer of SpotRight in Little Rock, a data analytics and marketing firm, said big data gives banks a chance to compete for customers in a variety of ways. Knowing a customer’s profile — age, marital status, children, etc. — can let a bank better customize its website when the customer visits it online, or market specific products to customers who are more likely to be interested in those products.
So if you recently have had a child and have blogged about it or shopped online for diapers, then a bank can likely develop your profile with those facts in mind. “Just across marketing in general, you want to target your message for a particular audience,” Greer said. “You could just do broadcast medium with the same message for everybody, but statistics show if you make it more relevant to me, I’m more likely to pay attention and more likely to do something about it.
“It’s because they’ve made an attempt to profile you as a consumer and tailor a message that is more fitting for you. Obviously, it’s a lot of guesswork. They try to put you in big bucket categories of parents of young children or things like that. There is enough data out there to get pretty precise.”
Greer said the other big target that big data helps banks with is risk management with new customers. Those customers fall into two categories — someone new to credit, such as college students, and someone with bad credit history but not necessarily a bad risk currently.
“There’s a race for financial services to find those two categories of credit risk,” Greer said. “How do you find those and go after those types? You want more customers who look like your best customers, and you don’t want to waste your time on customers who look like your worst customers or those who aren’t likely to be your customers.”
Kinsey’s Lofty Labs works with MondoBrain of Virginia, a company that uses algorithms and data analytics in business modeling and predictive analysis. MondoBrain collected data from a host of banks related to lending and then analyzed it to determine how to categorize prospective applicants according to risk.
“For people in this age range, what is the ideal salary range that predicts 100 percent of the time that they will never miss a payment?” Kinsey asked, referring to a sample result of the analysis. “Credit management is risk mitigation, which in its way is a type of gambling. You’re hedging. What’s our risk? We’re trying to hedge our bets, so for every loss we have, we have a big enough gain that tends to justify the means.”
Discovering Trends
Ryan Frazier, CEO of DataRank in Fayetteville, said his company uses social media postings to track consumer spending. That data can give a real-time prediction of a company’s performance — rather than waiting for its quarterly earnings report — through comparing how many people are checking in at a store or blogging about a bank product.
“If you have more real-time flow of data, that gives you a better idea of trends,” Frazier said. “It can give you an indication if they are going to be earning and, if so, by how much. You can see if more or less people are checking in this quarter versus last quarter or this quarter versus the same quarter last year. It can give you an idea directionally of how they are performing.”
McGraw said shopkeepers used to know their customers because they talked face-to-face during transactions in the store. Now, many purchases are online or otherwise impersonal, so businesses learn about their customers through social media.
McGraw said Black Oak works with its clients to better handle the interaction and collection of information from customers. It’s important that people understand they control their data and how to share it.
“There is a creepiness factor in that,” McGraw said. “People don’t like to feel like they’re being spied on. Transparency in data use [is important] because you can’t underestimate the creepiness factor in all of this.”
Greer stressed that data mining is done with consumer privacy rights protected.
“Every point in life a potential consumer touches some area, and there is somebody who is hoping to interact with that consumer,” Greer said. “It’s way more precise now than it ever has been, and it will continue to be more and more precise as people get better analytics. As things get more sophisticated, so does the marketing and targeting.”