Breaking Down Big Data

No one disputes that Big Data—the accumulation, storage and analysis of streams of structured and unstructured data—has the potential to improve agency operations.

Better data analytics will help agency principals discern trends in retention, policy cancellations and expirations, carrier service and other vital business metrics—improving decision-making and overall profitability as a result.

But for now, Big Data is still taking small steps.

Starting Small

Why isn’t Big Data a big deal yet for agents? Because it’s actually in an embryonic stage for all kinds of businesses.

“People forget that a large company like a General Mills or Exxon is swimming in oceans of data,” says Denise Swink, an independent consultant and CEO of the Smart Manufacturing Leadership Coalition, an organization that promotes the use of Big Data in factory machinery. “We’re talking on the order of 50 million pieces of data a day.”

Sorting through all this data for performance insights is like searching the proverbial haystack for the needle. The computational power to accumulate wide-ranging sets of data and analyze them via sophisticated algorithms is still developing—and most agents aren’t even at the point of collecting data, much less knowing which data to amass.

“Most smaller companies remain stuck in a world of Excel spreadsheets and writing stuff down on paper,” Swink explains. “They need to ensure they have the right, quality data to begin with, and that it is delivered in the right context. Only then can they ask questions of the data to glean insights about their performance.”

Getting It Right

Vendors agree this is where most independent agencies currently sit. “It’s all about getting the right data at the right time in the right format,” says Christa Degnan, vice president of product management, agency business at Vertafore.

Tracking new business is a perfect example. In order to make sense of new business data, agencies must break it down into discrete components, such as the particular segment or line of business, insurance market, zip code and price. The data then becomes consumable, indicating which carriers are taking on certain exposures in specific locations in a particular price range. Agency principals can then evaluate the information and make business decisions accordingly.

All this data—and much, much more—is already is in the hands of agency principals. “Regardless of the agency management system, agents are sitting on a lot of very valuable data today—about their customers, carriers and their employees,” says Michael Howe, senior vice president of product development at Applied Systems Inc. “The problem is accessing it in a way that they can do something with it.”

How to Begin

But before that, agencies have to actually gather the data first, says Ron Berg, executive director of the Big “I” Agents Council for Technology.

“As mechanisms agents can use to analyze their data become more available, so have the number of things they need to track,” Berg says. “If they put in motion processes to begin tracking data like total premium, number of policies, expirations and cancellations, where referrals are coming from and other metrics, they are then able to learn from this insight to develop strategic and tactical plans for growth.”

While Berg says vendors have solid ideas on how Big Data can help agents, he acknowledges agents are “not there yet”—many are just beginning to differentiate between Big Data, data analysis and data warehousing. “They are saying they need Big Data, but they don’t know where to start,” Berg says. “And ‘where to start’ involves workflow changes, which can be painful.”

Undaunted, some agents are venturing forward into the world of Big Data. Nancy Sattler, owner of Sattler Insurance Agency in Lewiston, Idaho, is leveraging Vertafore’s new platform to analyze new business, account cancellations and carrier performance. “I’m able to know which carriers I’m retaining business with and not retaining business with,” she says. “I’m also able to track where I’m able to increase a carrier’s business by 20% and where I stand at the moment.”

This process requires Sattler to ride herd on her staff, making sure they input the right data in the right formats. Without the certainty of clean and accurate data, she’ll have nothing but a muddled view of agency performance. “This has become my new role and responsibility—to make sure everyone here is putting the right items into the system,” she says. “Data out is only as good as data in.”

The system needs to know, for example, where new business originates. “In a month, I might see that we’ve accumulated $40,000 in new business. But for me to analyze it, I need to know what percentage came from the Internet, from ads in the Idaho Press Tribune, from referrals and so on,” Sattler explains. “Then, I can study this information to better allocate our advertising dollars.”

The agency is establishing rules for what data it needs to collect, when and in what context. But this is not the first step in leveraging Big Data to improve agency performance. “You have to know the questions you want to ask of the data before deciding what data you need to answer those questions,” Swink says.

Ask the Right Questions

Agencies like Hoffman Brown Company are asking probing questions like these. The Los Angeles-based firm wanted a better grasp on retentions.

“We couldn’t calculate retentions before,” says Joe Pratts, the agency’s chief operating officer. “Was it 95% of renewals or 80%? We didn’t know how much business we were losing. I’d try to figure this out using Excel and always ended up guesstimating. Now I can see the exact percentage by department.”

Even better, Pratts is able to discern possible reasons why the agency is losing business in a particular zip code or with a specific carrier. “I can narrow down the retentions by desk, knowing that personal lines retentions are 97% with this producer and 88% with another producer, while our average agency retention is 93.5%,” he explains. “I can then generate a view of the possible reasons why the one desk’s percentage is less. I might learn the person’s book was made up of older people who have passed away, or people who have moved out of state. Or maybe the carrier’s service was not good.”

Hoffman Brown can also discern policy expiration dates on a producer-by-producer basis. The data may indicate that a particular producer has a heavy load of expirations coming up in July, while another producer is light that month. “It’s presented in a graphic, making it easy for me to see the disparity,” Pratts says. “I can then tell the person with the light schedule to help out the other producer.” That extra assistance may mark the difference between retaining an account and losing it.

What’s Next?

Vendors further enhancing their platforms will culminate in a range of integrated data analytical capabilities across lines of business, policies, producers, carriers and other sources. Sophisticated algorithms will spit out information on all sorts of valuable performance-based criteria, such as the best customer experiences with a particular type of insurance in a specific zip code on a carrier-by-carrier basis. And this information will beget a treasure trove of industry benchmarks, helping agencies compare their performance to other agencies on an anonymous basis.

“If an agency is losing business involving trucks, it can look at the experience of other agents in its area selling the same insurance to determine where it stands,” Degnan explains. “The agency principal can then figure out whether it is just his or her agency that is experiencing trouble, or all other agencies. If it is just that agency, then the principal can figure out if the agency did something that caused the problem.”

Both Applied Systems and Vertafore appreciate the newness of Big Data for agencies, and are more than willing to provide some handholding as agencies take their first tentative steps. “Analytics are notoriously trickier than meets the eye,” Howe says. “Going forward, we will work with agents as the product expands. We are not going to rest on the laurels of what we’ve already put out.”

Degnan agrees. “Big Data is only going to get bigger,” she says. “Right now we have it in bits and pieces. The goal for us is to help agencies get started—how to leverage what we’re already offering.”

Russ Banham is an IA contributor.