How a Small Insurance Firm Uses Big Data to Level the Playing Field
Large insurance companies are very good at aggregating data to analyze risk, but not so good at using big data to improve the front-end customer experience. That presents an opportunity for smaller firms, explains a digital marketing expert at insurance start-up Insureon.
- By Linda L. Briggs
- January 18, 2017
Smaller players in markets dominated by huge corporations can use technology to help level the playing field. Insureon, a start-up insurance broker aimed at small businesses, is a good example of that maxim. According to Forbes magazine, the company raised $31 million in venture capital in 2015, "most of which is being used to invest in [its] tech platform."
Upside spoke with Grafton Robinson, who manages digital marketing efforts at Insureon "You'd think that insurers would be great at big data," Robinson says, "but it's not always the case." In this interview, he talks about the challenges faced by small businesses such as Insureon and how the judicious use of analytics, particularly with website data, can help even the odds.
Upside: How effectively is the business insurance market using big data and analytics?
Grafton Robinson: Well, you might say that actuaries at insurance companies were the first data scientists. The entire industry is built on the idea that you can aggregate information to make informed decisions about risk. In the business insurance marketplace, that means carriers decide who they are willing to insure based on the size of the business, the location, the industry, and a dozen other factors.
So you'd think that insurers would be great at big data, but it's not always the case.
Most insurers have been around for decades (some for centuries). They're huge organizations that are used to doing things a certain way. That means they're great at aggregating data to analyze risk, but they're not so good when it comes to using big data to improve front-end customer experience.
If you look at the websites for many insurance carriers, you'll see a lot of static sites, no A/B testing, and very little audience segmentation.
What are specific challenges that smaller insurance companies face in competing with larger ones? How can analytics address those challenges?
The first challenge we face is getting the amount of data that our competitors have.
Consider this: Nationwide can hire Peyton Manning for its commercials. State Farm can hire Aaron Rodgers. These guys are huge media presences and marketing presences. A consumer who knows very little about insurance but wants to insure their business will think first about the company they saw advertising during Monday Night Football. That means the major carriers are up to their gills in customer data.
As a start-up, we have to compete against these giant brands for market share and data share.
How do we do it? There are advantages to being a small, technology-first company. We can quickly customize our communications to customers. We can A/B test ad copy and website landing pages, changing these the second we find a winner.
Basically, we have to use our agility to respond to the market and make sure that when customers come to our door, we don't lose them.
Can you give some examples of how analytics is used in the insurance industry?
The entire insurance industry is built on analytics. Statisticians use models to predict risks and lifetime value of customers. These customers are often with you for years, so your models have to be precise.
In digital marketing, analytics is everything. At any given moment, we're testing display advertising, search engine marketing (SEM) copy, affiliate marketing campaigns, email campaigns, and various designs of our homepage. We don't launch any campaign unless we can measure it.
We've also been able to use our data to drive innovation in the insurance marketplace. We're an insurance broker, which means we aggregate insurance options for business owners and sell them insurance policies that carriers provide. We're the middleman, which gives us an interesting position in terms of access to data.
If we work with carrier X and see that it isn't selling many policies to food truck owners, we can go back to the carrier and say, "If you change your insurance requirements here, we think you can sell a lot more policies."
To understand what this means, imagine that Kayak.com went to United Airlines and said, "You should schedule more flights to Miami this February." That's what we're doing. We're identifying areas that are underserved and convincing carriers to expand their markets.
How does your company, Insureon, use data and analytics for marketing?
We follow a test-and-learn model. We want to test a lot of marketing campaigns, find what works, and scale up quickly. To do that, you need an agile team. Honestly, that's my favorite part of working at Insureon.
We're a small, tight-knit group. Our team is a mixture of creative people who do the design and writing for marketing projects, analysts like me who design and track the test components of campaigns, and developers who build fast, customer-focused websites.
We sit in a room together, figure out the scope of a project, and get to work.
Everyone is flexible enough and good enough in their field that they can handle any project that gets tossed on their plate. In the last 12 months, we launched a major TV campaign that aired during the Final Four, ran a direct mail test, and A/B tested SEM copy for 400,000 different keywords.
Can you explain how technologies such as search engine marketing (SEM) and search engine optimization (SEO) are used to measure website performance? How do they affect your marketing efforts as a whole?
There's a lot to say here.
SEM is paid search advertising -- the ads that appear above your Google search results. SEM testing is probably the easiest to measure and execute. Everything can be tracked. Costs are controlled on a per-click level. Copy can be customized based on the user's geographic location. We can track how far users get in our funnel, their conversion rates, and our return on investment. SEM is the closest thing we have to testing in a closed system, and it's relatively easy to figure out if your campaign is successful.
SEO is the art to SEM's science. SEO is all about how high your pages appear organically in search results, and there's a lot that goes into that. The way your site is coded, the number of links you have, and the quality of the information you provide are the three pillars of SEO. What makes it tricky is that you're competing with all kinds of people for very limited real estate.
Say we're trying to rank for "general liability insurance." Well, Wikipedia is just as much a competitor as State Farm. We're competing against anyone and everyone who's trying to provide information about business insurance, and the competition is brutal. Click-through rates plummet as you go down the search results. The top two or three pages get the lion's share of the clicks; the bottom five get less than four percent combined.
Measuring SEO is challenging, to say the least. If you want to rank for a specific keyword, your organization has to accept that it's not going to get there overnight. Ascending the SEO rankings means committing to a long-term strategy. That strategy needs to include writing better content, coding better websites, and keeping a close watch on your competitors.
What about A/B testing?
A/B testing is fairly straightforward. Say I have a website with the call to action "apply for insurance." I can put those words in a blue box or a red box. I test both colors and see which wins. That's a basic form of testing, but the endgame is really multivariate testing.
With multivariate testing, we can say, "Ah, we have a visitor on our site at midnight. They generally convert better with the blue button." We're taking two factors -- time and design of the page -- and testing these together. That sounds simple, but it isn't.
Here's a more complex version of that kind of testing. We take a media market and flood it with all kinds of marketing. For example, we pick out Phoenix and increase the amount of radio, TV, and online display advertising in that area. Then we see if this has "lift." Does this strategy increase our conversion rate? Then we take away one of the components and see if the effect lasts. We adjust the levels of TV, display, and radio to find the optimal balance.
I bring up multivariate testing for a reason. When looking at the performance of any one particular marketing campaign (SEO, SEM, or website A/B testing), there's always an underlying question: does this campaign contribute to our branding in ways that aren't directly measurable? In concert, campaigns can actually increase the effectiveness of your whole marketing strategy. Two campaigns together might work better than each campaign alone. A user who saw your SEM ad a month ago may be more likely to click on your SEO result and apply for insurance. How do you measure that?
That's the impossible and fascinating task of the digital marketing professional. You're tracking individual efforts, but you're also trying to find ways to measure the complex interactions of these campaigns.