It’s hard to find a business enabler that’s evolving as quickly as Big Data analytics. This undisputed darling of the loyalty program and retail world has now infiltrated nearly every industry, from transportation and healthcare to insurance and supply chain management. It’s the catalyst for a new marketing mindset, where targeted offers are not only mapped to what the individual customer likes but how many “likes” that person is worth in social media.
While the trends around the use of Big Data are morphing as fast as Lady Gaga’s wardrobe, so are the accompanying challenges. How do you know if the data is ‘good?’ How can you validate information without delaying real-time analysis to the point where the opportunity has passed? And how do you use Big Data responsibly, increasing customer intimacy without crossing the line to customer voyeur?
We set out to find the answers and look at the emerging trends.
Big Data is Watching You—with Your Permission, Of Course
While knowing who you are and your position in life is still important, it’s the ability to look at what you do that is driving the analytics surge.
“There is a wealth of new behavioral data out there today,” explained Carl Madaffari, senior vice president, database solutions at Epsilon. “Take telematics (systems used in automobiles that combine wireless tracking with GPS tracking), as an example. Insurance companies are using this technology to monitor customer driving behaviors and adjusting rates accordingly. The value proposition is this: if you give us permission to monitor your driving behavior, you will be rewarded with lower rates if you prove that you’re a good, defensive driver.”
Once an individual checks that all-important, “yes, I give you permission” box, it opens the door for other types of interactions—like text coupons for discounts at nearby restaurants or special offers from other retailers en route.
“If there’s one emerging trend around telemetric data, that is collaboration and data sharing with business partners. Insurance companies work with restaurants, retailers, car dealerships and other types of vendors who could benefit from the data they’re gathering,” Madaffari said.
But, it’s not just the automotive and personal lines insurance sectors that are getting in on the act.
“Think about the growing use of Fitbits® and similar products that track an individual’s movements, sleep patterns and physical activities—all behavioral data that could help healthcare providers monitor patient care and insurance companies reward customers who adapt healthy lifestyle choices,” Madaffari said. “They can share this data with the local juice bar, and at the end of a run or workout, the Fitbit®-wearing consumer is rewarded with a coupon for a free kale smoothie to finish off his or her day.”
Loyalty Programs Transform from “Earn and Redeem” to “Surprise and Delight”
According to Madaffari, the loyalty program space is in the midst of transformation. While those old, familiar points programs still exist, that market is saturated. So, many consumer-facing companies are trying something new: the sneak attack.
“Instead of purchasing nine coffees and getting the tenth one free, companies are testing the concept of ‘unexpected rewards.’ The idea is to surprise and delight loyal customers with an unexpected discount or bonus,” Madaffari said.
It’s important to note that, while buying a lot of stuff from a company makes you a valuable customer, so does your perceived influence over other prospective buyers. Are your blog posts highly likeable? Does your Twitter following rival that of George Stephanopoulos? Your “net” worth just increased.
“Through the cloud store, we can quickly pull up an individual’s connectivity score and identify not only who is influential in social media but what kind of followers that person has, ” Madaffari said. “So, for example, if a person is heavy into home electronics, is typically a positive speaker when she posts, and has numerous followers and ‘likes’ among technology lovers, these factors might prompt the salesperson at a boutique electronics store to give her a 40 percent discount on her purchase that day.”
Using face recognition technology, an in-store retail system might identify a frequent shopper at checkout, ‘see’ that he bought two red sweaters on line and has looked at a jacket twice but never finalized that purchase. Because he has a large social following among the fashion-forward, the salesperson might offer him the jacket he’s been eyeing at a 30 percent off, just-for-him price.
Cool, right? To some. Other people might find the whole concept a little bit creepy.
“We have to walk a very fine line with how we use the data we collect. It’s the balance between learning what your girlfriend likes so you can be a better boyfriend to becoming the creepy, stalker guy staring into her window so she’s never out of his sight,” Madaffari said.
Of course, all of the Big Data trends aren’t focused on generating consumer response. We’re seeing Big Data make its move to more traditional industries, the most notable of which is commercial insurance. In the past, underwriting wasn’t necessarily a moneymaker; companies relied on investment income for revenue. Today, it’s a brand new world.
“In this era of low interest rates, insurance companies need strong real-time analytics capabilities to achieve the elusive underwriting profit and sustained growth,” explained Amit Unde, chief architect and director of insurance solutions for L&T Infotech. “Going forward, the competitive battles will be played on the data turf. It’s the companies that leverage both external and internal Big Data, predictive analytics and adoptive underwriting models that will come out on top.”
In the past, commercial underwriting was a back-office function, with decisions based on agent submissions coupled with the underwriter’s intuition. Because agents have a vested financial interest in gaining approval at a competitive price (it’s called commission), their submissions sometimes painted a rosier picture than what actually existed.
“With Google Maps and location intelligence services, the underwriter can view a property from all angles and assess distance from a coastline, flood plain or other potential hazards. Online access to hundreds of different data sources—from videos to photos to loss trends and other documents— is now just a few clicks away,” Unde said. “But, without the right tools, mining this data is still a highly manual process.”
Companies can leverage these technologies, the new abundance of data and outsourcing partners to develop or acquire the right tools.
“The right outsourcing partner can help companies establish tangible, repeatable methods and tools to more accurately assess risk and price these policies accordingly,” Unde said. “For example, we offer an underwriting solution that enables insurers to integrate internal data sources and historical pricing data with most public data sources and third-party providers to automate the risk scoring function.”
In the years to come, the use of Big Data and new technologies could completely transform the industry.
“I wouldn’t be surprised if, in the next five years, the next big player in the commercial insurance industry was a new company with a Big Data-driven automated policy issuance and claims payout model,” Unde said. “Automated decision-making has the potential to transform the industry, enabling small players to compete with large insurers, based on their technology.”
The Challenge of Data Accuracy and Interpretation
Of course, no great advances come without challenge. The buzz-kill of all of this: unless you’re working with accurate data, you’re not going to see monumental benefits. “Garbage in, garbage out” still applies.
“Every time we come up with an untapped source of data, there’s a certain exuberance, But then, as we refine and narrow, we see issues with that data. So, more data isn’t always better,” Madaffari explained.
One way to make sure you’re working with clean data is to work with a partner who offers real-time hygiene tools. Also, take the time to sweep your own database to ensure your internal information is accurate and complete.
“In the insurance industry, companies should validate against a set of rules or cross-verify against multiple sources,” Unde said. “However, in most cases, it doesn’t make sense for insurers to boil the ocean to get 100 percent data accuracy. It makes better sense to apply the 80/20 rule to achieve the desired accuracy for the 80 percent of the dataset without having to invest intensive efforts—then asking ‘did you mean’ questions in the remaining 20 percent of cases.”
In the marketing and loyalty world, preventing inaccurate interpretation of intent is always top of mind. For example, is the person who logged on to a medical web site actually impacted by that disease? Is the person comparing auto insurance rates really in the market for a new vehicle, too? If a person bought a petite sweater online, is she really 5’1 or purchasing a gift?
According to Madaffari, it takes more than just technology to get things right.
“Big Data is part magic and science,” he said. “Newer technologies are upping the ante from a science perspective. So, there is a huge need to enhance the magic part of the equation, which is the people who can provide the context and relevance to that data and how it can be used.”