Big Data for Bigger Profits
Just because something is heavily hyped, doesn’t mean it’s not a big deal. Big Data is a big deal. Data sets that are too large and complex to be captured, stored, searched, shared, analyzed and visualized by ordinary data management tools are chock full of potential value, particularly in terms of the ability to combine structured data (such as credit card and deposit account records) with unstructured data (Web click-throughs and text-based content).
For banks, Big Data has the potential to promote new waves of productivity, more revenue at much lower costs, greater customer satisfaction and fraud and risk prevention. The challenge is to achieve these goals without running aground on the hype that tends to accompany so much of the discussion around Big Data. Here are some suggestions:
Build to use. Building for Big Data isn’t cheap. You might wish to make information you already have understandable and usable faster. You already know, for example, which customers have called to say they will be late with their mortgage payment. Undoubtedly some of those customers are slipping into the “will-never-pay” category, while others may be just the opposite: highly responsible borrowers who definitely will pay. With Big Data you can examine other information about those customers so that you can take swift action on the former while avoiding insulting the latter.
Or you might wish to use Big Data to get more accurate and detailed information about customers. Their deposits are a treasure trove of customer information but with “traditional data,” a deposit of ten checks can be just that – a single deposit in banking systems. Big Data can analyze that same deposit at the item level and gather important data about your share of that customer’s wallet and resource and activity levels at the branch.
In another common use, Big Data can get down to, if not that elusive “segment of one,” at least very narrow segments. For example, a customer’s activity on the bank’s Website could determine what message that customer finds on returning to the site. The mobile transaction confirmation sent to the customer’s phone could be personalized with an offer tailored to that person’s situation. Traditional data has already dramatically diminished the massiveness of mass marketing but Big Data gives you an even narrower focus with the dual result of cheaper marketing and increased sales effectiveness.
Some banks look to Big Data as a way to refine their decision-making. As banks have gotten pricier to acquire, for example, active acquirers are looking for more and more information on the selling bank. All these are legitimate and likely uses for Big Data, but each use involves different infrastructure considerations to settle up front before a costly build-out.
Integrate, don’t replace. Big Data’s possibilities can tempt bank business heads into embarking on rapid development, business by business, sometimes circumventing the data infrastructure already in place. This temptation is facilitated by the fact that many of these tools were built as open source and are thus more accessible.
Resist that temptation. Your data warehouse, repositories and other data services are proven, value-producing solutions built at significant cost. Big Data’s new tools are complementary to what you already have and need to be integrated with your existing infrastructure. Otherwise you run at least three risks: sacrificing your current analytics, creating information that is not “apples to apples” with your current analytics and building yet another silo. The rework to amend these hasty decisions is always costly, so take the time to plan a thoughtful integration.
Share Big Data’s value with the customer. Customers are increasingly astute about the amount of information you have about them and its value. Through airline and other loyalty programs, they are familiar with being rewarded by programs that also serve the company’s good. When GE Appliances wanted to improve the productivity of their service workers by cutting the time it takes to diagnose problems with appliances, they put ethernet ports on their appliances and outfitted the workers with laptops loaded with analytic software to plug into the ports. Not only did the company achieve its goals, but now customers don’t have to wait as long for repairs and are spared repeated visits for misdiagnoses.
In the past, for banks to learn commercial customers’ needs, they had to ask for meetings with them, perhaps multiple times, ask a lot of questions, and show them several draft plans. Today, if you can crunch a lot of Big Data before even calling on customers, you can present them with well-researched options at the outset, saving them time and improving your offering.
But don’t creep out customers. Just because you are in a position to connect a lot of dots doesn’t mean that customers will appreciate seeing them connected. A well-publicized example was Target’s controversial use of Big Data: How Target Figured Out a Teen Girl Was Pregnant Before Her Father Knew.
The more information you glean from customers, the greater your obligation to handle it carefully. You can probably connect enough data points to predict that a customer is ready to buy a new car, that another is looking for security openings to launder money, that another’s marriage is breaking up, that a college fund account is emptying out much faster than usual, or that a family member is critically ill. In each case, propriety, the law and decency prescribe different courses. Getting them wrong will backfire, with customers getting stingier about giving data or leaving the bank if they feel uncomfortable.
Be ready to invest in the necessary talent. When companies begin looking at Big Data options, it doesn’t take long to discover that they have skill gaps to fill. They need statisticians to comb through myriad systems to format complex data for business ease-of-use. They need data scientists, regulatory experts, data security experts and so on – obviously not skills typically found in one person. According to Apollo Research Institute, the ideal worker in Big Data has been described as “‘T-shaped,’ that is, having depth in at least one specialization, but also a breadth of understanding across a range of specialties.”
Big Data is still data – it’s not answers. It still requires good communicators who can link data and findings to decisions and the bottom line.