Would you like to predict the future? The proper use of data—both internal and external, might let you do just that.
To make better predictions, companies need to access and analyze myriad external data.
The data provide insight, which drives which products to make and which markets to enter.
Retailers use analytics to forecast purchasing behaviors and customize advertising.
There are enormous opportunities for brokerages and agencies to improve client services, grow revenue and develop more cost-effective processes and systems. Some carriers already realize this and have invested in robust predictive data capabilities. The time has come for the rest of the industry to follow suit.
Many brokerages and agencies think they’re already using data analytics. In such cases, they may only be analyzing internal data. Analyzing these historical metrics and events can lead to a range of possible outcomes. And although it’s a good start, it’s only half the race.
To make better and more wide-ranging predictions, companies also need to access and analyze external data, including real-time information on geopolitical happenings, macroeconomic conditions, weather patterns, market information, customer demographics, competitor moves, and social and cultural trends—literally millions of data sets, all of them flowing over the Internet. In the past, businesses were overwhelmed trying to make sense of all these data. By using a computer to analyze and make the data searchable, this is no longer the case.
Up until recently, predictive analytics was confined to accessing and analyzing only structured data, the information residing in a company’s database or on line-item spreadsheets providing metrics on sales, claims, revenue and so on. What’s been missing is the analysis of unstructured data, the information that doesn’t fit gracefully in a spreadsheet—stuff like corporate email, digital documents, photographs, Web pages, and video and audio files.
The goal is to access as much structured and unstructured data as possible—really big data—and then create algorithms to draw out interesting patterns and correlations, highlighting information truly useful to the business. From this information comes insight. And insight is everything in business. It drives which products (or services) to provide, improve or abandon; which markets to enter or exit; and which geographic territories to engage in (or not).
With insights drawn from your data, internal operations become more efficient and workforces more productive. Customers can become more satisfied with your service, opening them up to cross-selling and up-selling opportunities. Lastly, data analytics can narrow the odds of someone buying something, positioning the seller to offer the right message to a consumer at the right time. Learning that a non-client is about to increase the number of its factory workers before competitors know gives your firm a leg up on marketing its workforce safety program first.
A data scientist once said predictive analytics takes the world of business and compresses it into a perfectly packed snowball of glowing insights. In today’s 24/7 global business environment, the organization with the most recent, relevant and accurate insights can respond more decisively to this intelligence, making sharper decisions that leave competitors in the dust.
Follow the Leaders
Many industries are doing just that. Banks and stock brokerages, for instance, are using advanced analytics to better understand their customers in relation to their products, distribution channels and rates. Based on this information, they can market products more aligned with their customers’ progressive life stages, such as when account holders get married, have children, send their kids to college and head into their retirement years.
Retailers are using analytics to forecast customer purchasing behaviors and then are customizing their advertising messages. Similar to what banks are doing, retailers are analyzing digital data on a current customer’s demographics, social media interactions, and Internet search and browsing activities. These data are then added to external market and economic data to extrapolate ways to generate more personal pitches to a consumer.
Another industry turning data into dollars is real estate. The online real estate firm Zillow is accessing previously unobtainable government information like census and permit data to make better predictions on home values. Previously these data were unformatted, unstructured and outdated. Zillow correlated these new data with its other metrics on the market value of a home to improve its forecast accuracy. The result? The company nearly halved its median error rate of a home’s final sale price from 13.6% to 6.9%.
Healthcare providers are using data analytics to reduce hospital readmissions—one of the biggest factors in the high cost of care. Algorithms identify specific social and economic patterns among certain patients then predicts those at high risk for readmission. Once patients are identified as such, the hospital targets various interventions to reduce the possibility the person will be readmitted.
Professional baseball, among the most data-intensive businesses anywhere, is also using data analytics. The Boston Red Sox created models that extract data on single-game ticket sales and revenue from merchandise, food and parking to determine how many ticket takers, ushers and security guards the team will need for a particular game, as well as how much food will be needed to feed spectators. In other models, the team also incorporates data on the weather, the opposing team, the day and time of week, the team’s standings and the roster of able players.
These different data sets, when analyzed using proprietary algorithms, identify unusual patterns and correlations. Dollar Beard Night was a promotional campaign in which anyone who showed up with a beard at Fenway Park received a $1 ticket to the game. As expected, attendance went up. What was unexpected was a spike in hot dog sales. The team had the same number of hot dogs it always had on hand for a full house, but it was caught short without enough wieners. Who knew? Beards and franks.
The Boston Red Sox created models that extract data on single-game ticket sales and revenue from merchandise, food and parking to determine how many ticket takers, ushers and security guards the team will need for a particular game and how much food will be needed to feed spectators.Tweet
‘We’re Not Selling a Thing’
The diverse opportunities for using predictive data analytics are many. Like the Boston Red Sox, brokers already have a wealth of customer data at their disposal, including information on policies, client retention, the duration and costs associated with claims, and wide-ranging personal lines and commercial lines customer demographics, including clients’ financial and credit histories. This is extremely valuable data, assuming something is done with it all.
That’s just the internal data. There are millions of external data sets out there that should also be accessed and analyzed to provide better services. Sean Allen, the vice president of business process services at Xchanging, a technology solutions provider, can imagine a day when a marine insurance broker provides sensors measuring air and water temperature, engine oil pressure, ship vibration and other factors to a ship owner. “The broker would be analyzing diverse streams of data informing navigational best practices—information that can be of great value to a ship owner,” says Allen. “I could see the same scenario playing out with clients in other industry sectors, with the broker providing a premium discount for those using the sensor.”
While a sensor-embedded ship may be in the future, some firms are already collecting and using data to help clients make strategic decisions. Baltimore-based brokerage RCM&D, for example, is using predictive data analytics to help clients with their workers compensation claim profiles.
“The analytics predict which people, based on claims data and behavioral information, are more likely to take longer to recover from a workplace injury or illness and return to work,” explains Albert “Skip” Counselman, RCM&D chairman and CEO. “This is pure risk consulting. We’re not selling a thing. We deliver far more value when we can reduce a client’s risks than when we just sell them a policy.”
Xchanging’s Allen projects brokers will be increasingly valued for the services they bring to clients rather than just the transfer of risk. “The ones that will be around in 10 years will be those that offer predictive risk services, not just reactive risk transfer,” he says. “Clients want more than an intermediary between them and the carrier.
They want to know the value-added services the broker can provide that address their particular risk profile.”
Aon Risk Solutions is in the midst of providing such services to a major client. “We’re working with one of the largest food companies in the world to analyze their risks and become smarter in how we advise them about these risks—both risk mitigation and risk transfer,” says Lori Goltermann, CEO of Aon Risk Solutions U.S. Retail.
“In this work, we’re leveraging our Global Risk Insights Platform, a giant database from which we can extract analytics on an industry-by-industry basis.”
For proprietary reasons, Goltermann could not provide the client’s name or divulge much about the use of the database. But she underscored why brokers are well positioned to these in-depth analyses. “We’ve got more than $147 billion in premium flow,” she says. “From this flow, we have an enormous volume of highly specific and insightful customer, policy and claims data that no other industry can match.”
Predictive analytics can also help brokers retain clients. A case in point is a client that may be in M&A mode, perhaps in the market to sell the business. An algorithm can be devised to sift through disparate data sets on the Internet, as well as a client’s public information, looking for information that may indicate the company is shopping itself for an acquisition. Such unique correlations may speak volumes.
At the Starting Gates
You have to get your hands around these metrics first to figure out what you want next from a business standpoint.Tweet
Predictive analytics, in which a broker applies proven statistical techniques to help clients understand their opportunities and risks, holds the power to change the broker’s business. Mike Victorson, president and CEO of M3 Insurance, in Madison, Wisconsin, aims to capture predictive analytics capabilities at his agency. “We’re at ground zero right now, meaning we haven’t leveraged any analytics,” he says. “But we fully plan to.”
In this quest, Victorson has hired a new head of IT from the retail industry with a background in e-commerce. The IT head is responsible for helping Victorson gain a better understanding of each customer by segmenting them by their business type, risk profile, insurance policy and claims history, among other categories.
“We now have a fresh look at what different customers buy and why they buy it,” he says. “You have to get your hands around these metrics first to figure out what you want next from a business standpoint. If you start with data analytics out of the starting gates, you’ll be running around chasing your tail.”
Victorson is now in the process of evaluating the next steps, which may include the temporary hiring of a data scientist to “do the measuring and predictions,” he says. “I can see this as differentiating our agency in the future.”
The Chief Data Officer
As Victorson touched on, the skill sets needed to really use data in this way do not necessarily exist in the current agency or brokerage. Putting data analytics into practice requires recruiting employees with statistical, mathematical and computational skills that most firms lack (they’re not in the IT group). Such employees can collaborate with personnel in finance, marketing, HR and other corporate functions to create predictive models that guide more informed and insightful decision making.
A large number of insurance companies have recently created the post of chief data officer (CDO), recruiting data scientists and statisticians from other industries to assume the post. CDOs have different backgrounds and skills than chief information officers and chief technology officers, who tend to focus on technology for the sake of operational efficiency. CDOs, on the other hand, have more of a business focus. Their primary task is to understand the different types of data inside and outside an organization in order to tease out more refined and profitable ways of achieving strategic goals.
The CDO forms a staff of similarly skilled mathematicians and statisticians who can create proprietary predictive models based on specific business objectives. This data analytics group works with IT but is neither part of it nor reports to it. CDOs have line responsibility for their staff on par with a CIO’s responsibility for the IT group.
Without a CDO or some mathematically inclined employee with a background in data science, it would be folly for a brokerage or agency to invest in the development and expanded use of analytics. The solution is to hire the competencies needed to begin this journey. Just because data analytics feels like technology does not mean current IT staff members are capable of taking on this responsibility. In most cases, they aren’t.
None of this will be easy. But once brokerages and agencies begin down this road, they will gain an added benefit—a new image as a technological leader.
“The more technological the brokerage industry becomes, the less it will be perceived by smart and technologically gifted college graduates as stodgy or Luddite,” says Robert Hartwig, the president and chief economist at the Insurance Information Institute. “The opportunity to use sophisticated data analytics in a brokerage or agency is very enticing to college graduates who want to work in organizations that are dealing with technology issues of extreme importance.”
In this way, brokerages and agencies can better fill their ranks with younger generations of technologically savvy minds who can continue the digital evolution.
Move over Google, Uber and Apple. There’s a new techie in town.