There is no doubt that the multitude of new channels and people’s ever-decreasing attention span make it a challenge for insurance companies to connect with their audience and get them to engage with their brand. As marketers, we must continually adapt to the new realities in the marketplace.
Science can help overcome these issues in two important ways. First, we can leverage science to better understand our audience, and then we can use it to move them to engage. The first step is achieved with data science, and the second with behavioral science.
Consider these statistics from LIMRA: Sales of individual life insurance policies have declined more than 40 percent since the 1980s, and about 30 percent of U.S. households have no life insurance at all — up from 19 percent in the 1980s.
These statistics should spur industry change, and disruption from an underwriting perspective is already underway. Data and technology play a role in changing the process such that client profiles, instead of client medical history, allow online sales to grow, particularly for younger clients. Improving the process of getting a policy means the customer experience is easier.
But use of data for underwriting is only effective after someone has made the decision to get insurance in the first place. What we need is something that helps get prospects “in the door".
Data science plays a key role in defining the audience so that outreach can be more effective. The economics around data are clear: Increased sales and efficient communication means costs decrease with the effective use of data in segmenting the audience — and then communicating with them at the right time, over the right channel.
The Role Of Data Science In Increasing Profitability
How do we take advantage of data science? Practically every analytics project should start by identifying the business value that can lead to revenue growth and increased profitability. It’s too expensive for marketers to ask everyone if they want to buy their insurance product. It’s always been true, even before the current level of robust data was available, that a dragnet is less effective than a targeted message.
The trick is to target the right group, which is where market segmentation comes in. Bundling like-minded consumers into the same groups increases the likelihood of engaging the audience. Thanks to the digital revolution, marketers can now comb reams of data to develop those segments. Algorithms can find correlations that elude human perception, which then provide information marketers can use to develop market segments.
Of course, this can be taken to the extreme. I once met with an insurance company client who told me they had built a model that revealed the ideal candidate as athletic, married, with a Ph.D. in biology, and living in suburban Boise, Idaho, with two kids and a dog. One person.
Obviously fantastic, but in this case, just have a salesperson call her. This is not the efficiency we expect from successful segmentation. It’s silly to cut so fine a segment/model that it’s doubtful you’ll ever see someone with those characteristics again.
For effective segmentation that includes a larger pool, we should start with a data audit. What information do you have about your audience, and how valid is it? This helps develop your append strategy.
Is there enough information about both current customers and prospects? If not, then we need to develop a strategy to append the additional data attributes about the audience to more effectively separate them into groups.
With the data set refined, we then have the ability to build regression models. The purpose of this modeling is to generate a deep understanding of how the customer base and the prospect base compare against the general population in several key areas of behavioral and demographic data.
By isolating not only the customer group but also the addressable market, the data can be cut by several different factors (such as gender, wealth index, media consumption, and charitable giving).
This lets us understand where these audiences show a propensity to engage with the brand, given their behavioral and demographic traits. This is statistically valid and predictive in nature, which allows us to expand to lookalikes with the highest propensity to respond.
The data scientists of the world think this is easy. “Anyone can do it!” Marketers, including myself, are less sanguine, but that’s why we employ those data scientists.
How To Cross-Validate & Confirm Your Findings
The next step in leveraging the data for these newly created segments is to test. The first test is to make sure you identify the segment in a reliable way. Run the analysis on half the data set, then run the same analysis on the other half.
Some segments showing up in the first run-through may be anomalies, and they can be discounted if they do not appear on the second runthrough. This cross-validation is important because we want the analysis to be confirmatory, not misleading.
From there, the testing should be in market. If we feel we have a valid segmentation strategy, let’s take it for a test drive. This is not with the entirety of the universe at the expense of our whole marketing budget, but with a subset to validate the results of our model.
A pilot test can not only validate the model, but it also help us “fail fast, but small” — yielding positive learning, even if it isn’t the ideal segmentation to start.
Even as a believer in data science as the driver for improved segmentation, I don’t think we should treat data and technology as a black box, throwing data into the box and automatically trusting the answers that come out. Someone on the team — a human — still has to understand the forces that shape a segment. Not knowing human behavior can leave a marketer in the dark.
This might cast me as a Luddite, but I believe that currently, the human direction of the models that are built shapes the best segmentation.
That said, today, with artificial intelligence, machine learning algorithms, natural language processing, and a host of other technologies, human intervention in creating effective segmentation may be as short-lived as the 16-bit microprocessor.
Why Behavioral Science Makes Messages More Effective
Once we are convinced we have good segmentation — or a model to identify who within the audience we should communicate with — it’s time to think about how we’re going to communicate with them. We want to talk about benefits, not features, and focus on what our audience is interested in.
That’s just the baseline for engagement, but we also need to leverage behavioral science as a means to prompt action. Simply put, behavioral science uses sociology and psychology to understand and drive human action.
The premise, well outlined in Gerald Zaltman’s book How Customers Think, is that humans make 95 percent of their decisions in the subconscious. The decisions people make are ingrained in them and, in many cases, automatic. Behavioral science leverages these subconscious biases and moves people to action.
Once again, it’s important to understand the economic reasons to employ science in marketing. We know audiences have a decreased attention span and many more channels to choose from when it comes to getting information.
Marketing campaigns need to cut through and engage people. When they do, the results speak for themselves in increased revenue. Behavioral science is a tool that differentiates the message and creative in a way that enhances engagement.
Start With Cognitive Fluency, Hyperbolic Discounting & Storytelling
There are countless behavioral science principles, but for the purposes of this article, I’ll cover three that are important to marketing insurance.
The first is cognitive fluency. The idea is, the easier we make something to read or ingest, the more likely people are to believe it and act upon it.
Many studies have been done to show that color, white space, language, and the creative organization of marketing has a huge impact on engagement. Humans see something complex and wonder what’s being hidden from them. Conversely, when they see something that’s easy to read, clear, and transparent, they believe it to be more true.
Insurance marketers must do all they can to take their complex topic and make it less daunting for our customers and prospects. Using common language, easy to grasp examples, and simple graphics to illustrate main points are just some of the ways to inject cognitive fluency into communications.
Another strategy, particularly for younger audiences, is to overcome hyperbolic discounting. Essentially, this is people’s tendency to choose a smaller reward they can have now, instead of a larger reward they would have to wait for later.
We are hardwired for instant gratification. This is one of the biggest obstacles to the sale of insurance of any kind.
Particularly for Millennials, to overcome hyperbolic discounting, companies must craft their message in a way that helps place the audience in the seat of their future life. The key is to get them to imagine themselves in the future.
In our own proprietary research, we found that not only can we influence the purchase of insurance, but we can also influence the amount purchased by leveraging this behavioral science principle. When combined with the emotional association of Calm, it improved purchase intent even more.
Emotion plays a big role in moving the audience. Most of us think we’re quite logical, but in fact, we make 87 percent of our decisions emotionally — and then we justify those decisions with logic after the fact. Combining behavioral science with emotional associations provides a one-two punch for engagement.
Storytelling is the third behavioral science principle that can move the needle on response. This is a way to get people to relate, relive, bring their own emotions, and, most importantly, engage.
Neuroscience gives us a clue as to how. FMRI tests that observe the brain while a person watches a story found that the brain doesn’t look like a spectator, but more like a participant in the action. In watching a scene that’s sad, for instance, the viewers’ brains also look sad.
Studies show that our minds are simply flitting all over the place all the time. So how do we pin down the wandering mind? We can do this by telling stories — engaging, relatable stories not about us, but about what matters to the reader or viewer.
In normal life, we spin about 100 daydreams per waking hour. But when immersed in a good story, when we watch a show or read a novel, for example, we experience concentration (the opposite of distraction). Storytelling and engagement go hand in hand.
For me, personally, the enjoyment of listening to stories has led me to tell them, just for the fun of telling them. I recently told a story about my siblings at StorySLAM, an open-mic storytelling competition hosted by The Moth.
The audience seemed to enjoy it so much that I’m headed to the GrandSLAM soon. I can tell you firsthand that in both telling and listening, the emotional connection cannot be ignored.
Science Is The Key To Engagement
In the end, we all want to connect with our audience and have them engage with our brand. This can’t be done without understanding them first, and data science gives that insight. It can’t be done without driving interest from that audience, and behavioral science gives us that edge to drive engagement.
We have seen, time and again, the efficacy of employing data science for audience definition and the science of human behavior for engagement, independent of channel.
Many marketers need to drive engagement with a limited budget, even though the target for growth is always high. Data and behavioral science are two keys to making that happen.
This piece was originally published by LIMRA in their MarketFacts Quarterly magazine. John Sisson, Chief Strategy Officer for Universal Wilde, will be speaking at the LIMRA Marketing Conference in Nashville, TN on June 2, 2017. Head over here to learn more about the event.