Big Data. Nate Silver. Moneyball. Buzz words? Yes. But there is something significant behind them — namely, a realization that using data in new ways can yield improvements in all types of personal and business scenarios. Tongues are wagging and even Hollywood is hopping on the bandwagon. But what is Big Data? And what can it do for your credit union?
Big Data might be best described as the phenomenon of looking at large amounts of existing data — as well as uncovering new sources of data — and then analyzing that information in order to discover new truths and build more effective strategies.
A common example comes from the movie Moneyball, which took its premise from a non-fiction book of the same name by Michael Lewis. This true story is centered upon the Oakland Athletics baseball team and its general manager Billy Beane. At the time, the Athletics were at a disadvantage because they were not a wealthy team and were always out-bid for the most famous players. To compete and eventually win against high-paying teams like the Yankees, the Athletics began using data mining to discover which players had unrecognized offensive talents and were undervalued by the baseball industry.
Statistics, of course, had been used in baseball for decades, but Beane’s thesis was that the metrics being focused on, such as batting average and runs batted in, were antiquated and thus obscured other player characteristics that could more useful in winning games. Helped along by some imaginative statisticians, Beane looked beyond the traditional areas of focus — including how often a player reached base (not only from hits but also walks and hit by a pitch) or reached third base — in the belief that such indicators would uncover inexpensive but exceptional talent.
Another tenant of Beane’s was that scouts using all the traditional methods of observation and evaluation were too subjective and their picks would actually do worse than those selected purely by creative data mining. In both cases, Beane was eventually proven right.
Nate Silver has also become a Big Data buzz word. He is a statistician, author, and blogger who, by virtue of analyzing data, has become a highly successful predictor of political races. Not incidentally, he also started early in life analyzing baseball statistics. In the 2008 presidential election, Silver accurately predicted how 49 of the 50 states would vote. In the 2012 election, he improved his record by correctly predicting all 50 states. He also predicted 31 of the 33 U.S. Senate races in 2012.
A third telling example of Big Data — summarized to Callahan & Associates and others by Chris Howard at a recent partners meeting of Credit Union Financial Services (CUFSLP) — comes from President Obama’s recent campaign. Conventional wisdom in campaigns calls for focusing on undecided voters with a high propensity to vote — i.e., those who are most likely to show up on election day — but a low affinity toward your particular candidate, and then converting them to a high affinity for that individual.
The Obama campaign strategists altered this process. They tried to find people who leaned toward Obama already, but had a low propensity to actually go to the polls. In other words, they sought to convert low-propensity, high-affinity voters into high-propensity, high-affinity individuals who would end up casting their ballot for Obama come Election Day.
The strategists looked in some unusual places to find these individuals. For example, they discovered a high correlation between people who shopped at Whole Foods and people who favored Obama, so they specifically targeted those shoppers. They also discovered that people who ate at Cracker Barrel had a low affinity for Obama, so to save money they avoided messaging to neighborhoods high in Cracker Barrel customers. They also discovered what Obama supporters liked to watch on Netflix and then urged people who liked the same movies to go to the polls.
The Romney campaign used Big Data techniques as well, but less creatively and obviously to less effect. Romney's people looked for likely voters and targeted those who favored the Republican candidate. They did achieve the goals they set out for — securing many of the high propensity voters by which most historic elections were won or lost. But they were trumped by an Obama team that had worked to redefine what the winning goals for this election would be. Obama strategists went beyond the norm to get traditional non-voters to the polls, altered the pool of voters as a whole, and thus secured the victory.
These individuals were savvy in another way as well. They understood — as did the Romney team — that they needed to secure only as many voters in the several swing states as were needed to actually win. With the necessary number of voters in hand, the election in that state could be won and extra resources allocated accordingly. In short, strategists from both campaigns learned to put their Big Data findings to the most efficient and effective use.
What Big Data Is And Isn’t
These examples have not been lost on any number of entities, including credit unions and the for-profit financial institutions they compete against. Can credit unions take advantage of Big Data techniques to further their mission, to attract more members, and to build stronger relationships with the members they have? The answer is yes. But in order to do so, credit unions must drop some old notions and begin to think more creatively.
What do you need to reap the eventual rewards of Big Data? To begin with, you need data. And, in fact, credit unions have more data on hand than they generally tend to recognize or acknowledge. Most credit unions have at their fingertips all kinds of useful information, including data from the Census Bureau, as well as information on credit, transactions, demographics, spending habits, size of households, and where members spend their money. Other information might be gleaned about shoppers at certain stores or kinds of stores, people interested in affinity or rewards programs, or correlations between income and mortgage size.
Credit unions should start by assessing what data they already have and what they can get access to. The next step has two parts: How do you ask the right questions, and how do you avoid being overwhelmed by data; that is, entering a static zone in which data-mining answers are placed on the table but end up leading nowhere.
Part of this second step is understanding that Big Data is not an Information Technology (IT) function. Entering and exploring the realm of Big Data is a strategic and executive-driven function rather than an IT one. Big Data is really not about the data. Rather, it is about ways of looking at and using that information in innovative and non-traditional ways. It is about shifting through data in a way that offers you new insights on how to move your credit union forward.
A New Ocean To Sail
Big Data is a tool, a technique, and a resource that can be well or ill used. But we don’t think it’s just a bunch of transient noise. It may be similar to the personal computer industry in 1980, poised for tremendous work but without a clear understanding of what exactly the use for the technology was going to be. Remember that in 1980, PC manufacturers promoted their products as simply a place to store cooking recipes and Christmas card addresses. Like those manufacturers, people today have yet to uncover the many different uses of Big Data and make the demands on their information resources that will lead to transformative culture shifts and enhanced cooperative strategies down the line.
We at Callahan’s have been heavily involved in data since our founding in 1985. We believe strongly in the concept that a proper and creative use of data can help change any credit union’s outcome. It can help a cooperative work smarter and more efficiently, attract new members, deepen relationships with existing members, and much more.
There’s plenty of room for exploration and perhaps even cooperative exploration by which two credit unions, or groups of credit unions, can partner together on Big Data investigations and initiatives. There’s a new ocean here to sail.