As analytics aimed at helping companies make use of all their data proliferate, companies worldwide are embracing these new technologies and finding ever-more-innovative ways to use them. Those that fail to do so risk being left on the sidelines.
“Know your customers.” For years it has been a business mantra that is easier to say than to follow. Now companies can put that maxim into practice with the help of big-data analytics. But as many firms embrace the endless possibilities that big data provides in improving their internal processes and their relations with customers, others will have to step up their game or risk getting left on the sidelines.
A client complaining about her bank with a group of friends over coffee will probably go unnoticed, but one who posts those complaints on social media is likely to attract a better offer by a competitor. An electricity provider can inform customers in real time how much power they use and show them the actual cost of keeping the lights on. Banks, utilities, grocery chains and asset managers can understand their markets better than ever, thanks to a technology invented about 10 years ago that is slowly taking over the way business is done.
“Corporations are now able to understand who the customer is and what forces are impacting his or her behavior in doing or not doing business with them,” said Phani Nagarjuna, founder and chief executive officer of Neuvora, a start-up offering analytical services to Fortune 500 companies. “We tell these companies what they have to do to minimize the risks of their customers leaving them.”
Big-data analytics make all this possible. ‘Big data’ refers to all the digital information—facts and numbers, unstructured as they may appear in real life—that are continuously updated and ever-changing and too numerous to be contained by a single server and analyzed by a single computer. Big-data solutions, essentially, are software solutions aimed at analyzing large data sets. From a corporate perspective, this could include all the data produced via internal workflows, externally between companies and their counterparties (including customers, banks, regulators and so on) and beyond. Gartner Consulting defines it as “high-volume, high-velocity, and/or high-variety information assets that require new forms of processing to enable enhanced decision-making, insight discovery and process optimization.” Analytics solutions make that data available in innumerable outputs that can then be used to increase efficiency, drive sales, reduce risk or meet any of the myriad other goals that are still being defined by companies, analysts, consultants—and innovative start-ups—worldwide.
Such technology is becoming an embedded feature at many organizations, and soon companies across different industries and continents will find they cannot do without it if they want to stay competitive. At stake is not only the ability to get customer insights as granular as each individual but also the opportunity to save money with better forecasting, cheaper data management and smarter organizations. Similar to what happened when the World Wide Web became commonly used in the mid-1990s, big-data solutions are about to become an everyday tool for all.
“Big data is now seen as the new frontier of competition,” says Mark Torr, senior director for EMEA and AP Analytical Platform Centre of Excellence for SAS. “But ultimately it will become a commodity, and everybody will have to deal with it.”
The technology to analyze big data originated with Google’s studies on how to deal with the growing information available on the Web. The key engine that now handles that processing and storage is an open-source solution set named after the elephant
puppet of the son of its co-creator, Doug Cutting. The elephant was named Hadoop. At the time, Cutting worked at Yahoo! Apache Hadoop solutions are now distributed by IT firms Cloudera, Hortonworks and MapR. And countless software vendors make use of the Hadoop framework to provide analytics solutions to their clients.
Large corporations—manufacturers such as Daimler, utilities such as British Gas, and banks such Netherland’s ING, Austrian retail bank Bawag and Spain’s Santander—are all using Hadoop to crunch their vast data. According to presentations of their executives at the 2015 Hadoop summits in London and Brussels, some were looking to cut costs and most were seeking new ways to connect with customers and smarter alternatives to run their business.
“Santander has pooled aggregate information on clients, unifying all available information, including structured and nonstructured data and transaction activity. It plans to add to this public information on clients outside of their relationship with the bank. This information will be used to perform fraud analysis,” said an official source at Santander. “Our key vision for the Group is to obtain a 360-degree view of each individual customer.”
The financial industry with its multitude of data was the first to embrace big-data analysis and take up Hadoop after Cloudera launched the solution (or, more accurately, group of solutions) in 2008, says Steven Totman—who defines himself as a big-data evangelist and the first Hadoop vendor. “Big data is not something new. In all these organizations it tends to be the data they already had, but that they never took value from before.”
DRIVERS FOR CORPORATE UPTAKE
The push to contain costs has been a big driver. “The cost model of storing data in Hadoop is completely different,” says Totman. “In the warehouse space, people would spend $50,000 to $100,000 per terabyte at fully loaded cost, while in Hadoop [the same thing costs just] thousands of dollars if not less.” Another source says the cost of Hadoop is between one-tenth and one-twentieth that of traditional storage and processing technology. “A big deal, a once-in-a-generation kind of change,” is how Thomas Davenport, professor of management and information technology at Babson College and author of the book big data @ work, describes the appearance of the new technology.
The way big data is stored is also new. Data no longer needs to be set up in a certain way when written, as the format is determined when it is accessed—meaning that instead of storing data in tables with rows and columns, as an example, and knowing that it will be read by associated software in that format when it is accessed, you can decide how it should be read when you consult it. To explain the difference, one data specialist at Cloudera uses the comparison of a kitchen where all the ingredients for a specific dish are already cleaned and chopped, as opposed to a pantry where the ingredients are stored and so can be mixed and matched for different recipes. When a central bank issues a demand for a new type of stress test, it is handy for a bank to have raw, whole data at easy reach and be able to blend them in different mixes.
BANKS PLAYING CATCH-UP
Banks have been using big-data analysis to better manage risk, assure security with encryption and meet compliance requests. But they are now moving beyond that, using this trove of information to better court customers. They are playing catch-up, however, with other industries such as retailers. “It is clear that other sectors have gotten a head start in the race,” said the source at Santander. “Big-data technology has matured in the last years, enabling us to embrace it more securely.”
Big data is now seen as the new frontier of competition. But ultimately it will become a commodity, and everybody will have to deal with it.
~ Mark Torr, SAS
But banks are notoriously slow to embrace change. “I’m dealing with around 80 banks around the globe, and to all of them I am asking why they are not proposing a mortgage when someone uses their calculator online. Only five of them are in the process of considering this now,” says Marc Andrews, vice president, industry analytics solutions, for IBM. In June 2015, IBM launched 20 prepackaged, industry-specific analytic tools that can be used to deal with front-end clients. One of them is directed to retail banking, another to the wealth management industry. These tools are providing users with the ability to predict customers’ behavior based on different elements, like the time of day they shop or their propensity to spend on certain products.
Banks aim at predicting the life-changing events of their customers in order to offer them a service before they imagine they need it, Andrews notes. They try, for example, to predict when a customer will become engaged to his or her life partner so they can proactively work with the client on new account options to avoid losing the customer when the couple decides to merge their bank accounts.
The fairly low cost of the technology allows firms to experiment. “At a relatively cheap price you can try some of your new ideas, and if those do not work, you move on,” says SAS’s Torr. He says this allows companies to find their “gold nugget that may revolutionize the industry.”
MORE INNOVATION YET TO COME
Airbnb has wrought havoc on the tourism industry without owning hotel rooms, and Uber is changing the taxi business without owning cars. Retail banking services may not need a bricks-and-mortar banking branch to operate, but they still need a banking license, and innovation in the sector has so far been limited. There are some examples, however, that may open the door to further surprising developments.
One such example is Cardlytics. Built off banking transaction data, Cardlytics allows merchants to offer targeted discounts to customers based on analysis of their spending habits. Home Depot, for instance, could offer clients who made purchases at rival Lowe’s discounts to shop for specific products—related to their Lowe’s purchases. Free of charge for the banks, the service gives them a way to gain customers’ loyalty and make some revenue while sharing advertisers’ payments, says Brandon Horne, director of engagement at Cardlytics.
Cardlytics, started by two former Capital One managers in 2008, is in partnership “with over 2,500 financial institutions, many of them from a relationship with Fiserv, FIS [or] Digital Insight,” says Horne. They hold direct relationships with Bank of America, Regions Bank, Santander and Lloyds.
Sharing banking data with merchants represents the natural evolution of the Cardlytics business, Horne says, adding that while maintaining the safety and privacy of each customer’s data, they can provide indications of where the customers’ interests lie—anticipating what they will buy next, based on their purchase history.
Big data is also changing the world of analytics for manufacturers and financial services firms. David Kedmey, president and co-founder of EidoSearch, a numeric search engine for data, explains: “Big data has changed analytics and predictive models immensely because of the variety and the quantity of data available.” EidoSearch has about 100 customers in the financial industry, including hedge funds and assets manager. It forecasts thousand of items with a billion pattern comparisons every day.
At 4i, whose customers include Carlsberg, Colgate-Palmolive, Walgreens, Delmonte, Avon and many other brands, managing partner Eugene Roytburg says his firm can tell a manufacturer how to change its product to better serve customers. One example: a shaving cream that also takes care of the skin—a product created to build a demand that did not exist before it was offered.
The term “big data” emerged in 2011, but its implementation has just begun. “Big data as a term in the US is losing a bit of cachet,” says Thomas Davenport, “but we are still in the early state of the implementation and even earlier in Europe and in Asia.” A May survey conducted by Gartner showed that only 28% of the participants invested in Hadoop, with 18% planning to do so in the next two years. Whether Hadoop-based solutions continue to dominate the landscape going forward remains to be seen, and some industry watchers have argued that the system has some fundamental flaws that are starting to be overcome by other technologies.
But regardless of what system provides the backbone for big-data analysis, two things are clear: Making use of the vast reams of data that are produced every day is becoming easier. And such analysis will drive how companies interact with their customers forevermore.
