Segmentation is a fairly basic concept. Simply put, it’s “separating customers based on similar characteristics,” explains Nick Necsulescu, senior manager of customer segmentation for TD Insurance. But it’s when marketers break down this definition into micro-concepts that segmentation becomes more complex.
According to Necsulescu, there are three ways marketers can most effectively segment their customers: based on behavior, needs, and profitability. And each way requires different forms of data. Behavioral segmentation is based on interactions and database information (like purchase history or demographic data), Necsulescu says, while needs-based segmentation can be determined by survey or lifestyle data. Marketers can also segment based on low, medium, or high profitability, he adds, which can be determined through methods like customer lifetime value analysis.
All three methods are valid on their own, but marketers derive the most value when they leverage all three, Necsulescu explains. Here’s a progression of how three brands leverage various levels of segmentation incorporating an increasing number of those attributes to drive relevancy and, ultimately, ROI.
Segmentation based (mostly) on behavior
When it comes to email segmentation, Johnston & Murphy (J&M) is just getting its foot in the door. Indeed, the premium footwear and apparel company divvies up its products into categories and then relies on customers’ behavioral data to target patrons.
Back in the 1970s and ’80s, J&M was one of the go-to places to buy business shoes, says Heather Marsh, VP of e-commerce for the brand. “You go to Brooks Brothers and get your suit; you’d come to us to get your shoes,” she says.
But as time went on, some men wore business suits less frequently; others stopped wearing them altogether. So, by 1998 J&M expanded its offerings to include apparel. The retailer also debuted its first women’s collection in 2008. These changes caused the retailer’s marketing to evolve. Instead of just showing footwear in its direct mail or email creative, for instance, J&M began featuring full head-to-toe outfits for men and for women, Marsh explains.
As J&M’s products and marketing progressed, its use of segmentation followed suit.
For the retailer, segmentation begins at the product level. J&M assigns its footwear and apparel to one of three groups: end-use, which includes business casual and weekend wear; dress products, which fall into the brand’s more classic and formal look; and style, which includes more fashion-forward and edgy items, Marsh explains. The brand then looks at customers’ past purchase information, such as which categories they bought from, how often, and how much they spent.
Based on this information, J&M then tailors its emails to where customers are in their lifecycle.
For instance, when a new customer signs up for J&M’s emails, the brand sends him a welcomes series, including a thank-you message for subscribing, as well as some educational materials that tell the customer about the brand and its offerings, says Augie MacCurrach, founder and COO of Customer Portfolios—one of J&M’s marketing platform providers.
Once a newcomer makes a purchase, J&M focuses on driving repeat business. This is especially important, MacCurrach notes, given that more than 65% of J&M’s customers were one-time buyers when Customer Portfolios started working with the brand about a decade ago.
J&M encourages repeat purchases through repurchase and cross-sell email series, which leverage product recommendations and predictive modeling based on previous purchases. However, instead of sending three emails—one for each product category—to everyone, J&M suppresses customers from the list who may not find the email relevant based on their previous purchases. So, if J&M promotes shoes from the style category, the retailer will remove subscribers who purchase apparel from the end-use category.
J&M also sends reactivation series with special offers to lapsed customers to try to win them back.
Finally, the retailer occasionally sends surprise-and-delight emails to customers who spend the most, such as $50 gift cards to people who spend $1,000, MacCurrach says. These emails, he adds, can encourage customers to explore other product categories.
When J&M did send separate emails for each product category, these modeled emails generated a 150 to 200% lift in conversion compared to J&M’s generic emails, which produce “very low” conversion rates. However, while the modeled emails’ conversion rates were higher, their derived revenue was lower because J&M was reaching such a niche audience. “A broader message to a broader audience is driving more revenue than a very specific message to a small population,” she explains.
J&M also experiments with different forms of customer data. For instance, Marsh says that the retailer created a preference center to better determine which product categories customers want to receive communications for; however, most customers opt to receive information about all of them. It also tried incorporating more customer data—such as demographic and psychographic information—yet, there was too much category crossover to form any definitive segments.
“A customer who only buys classic footwear that’s business-oriented isn’t necessarily what you might think [of in terms of] an older guy in the finance profession,” Marsh explains.
Still, segmenting based on product category and customer behavior creates a few challenges. For one, there are discrepancies between marketing and the brand’s merchandising and product development teams in terms of what category an item should fall into, Marsh notes. “We spend a lot of internal time talking about those buckets and how products should be classified,” she says, “and we almost get tripped up on our internal viewpoints of these buckets.”
Customers can also view these product categories differently. Because “style” is so subjective, for instance, J&M may consider a particular blazer fashion-forward while a consumer might not think it’s edgy at all, she adds.
To help solve these dilemmas, J&M is embarking on a “segmentation 2.0” mission. Its goals, according to Marsh, are to create categories that are less subjective (e.g. “style”) and more consumer-facing. She aims to achieve these objectives by working with product development and merchandising to create the segments—instead of having marketing head the categorization solo. She plans to start developing this new segmentation model this year and act on it next.
Segmentation based on behavior and needs
While J&M primarily focuses on customer behavior for its segmentation, Baxter Credit Union (BCU) aligns its customers’ behaviors with their needs. Taking this approach is not without challenges.
The first is determining what data to use. About seven years ago BCU aggregated all of its siloed data into one data warehouse. This, in turn, created a shift: Instead of wondering how they could obtain more meaningful customer data, BCU’s marketers questioned where they should begin and which data they should ignore, says John Sahagian, BCU’s VP of marketing.
The second challenge is defining segmentation as it pertains to BCU’s business. Before having a centralized database, the credit union didn’t do much segmentation. According to Sahagian, it relied on basic slicing and dicing of data—such as whether customers had checking accounts. However, after creating a data warehouse, BCU began data mining and built a marketing automation engine to enhance member outreach and engagement in May 2013. Now, the credit union can align its product information with lifecycle-, behavioral-, and needs-based data to deliver a more personalized experience.
Still, some disparity remains between how BCU’s senior managers and marketers viewed segmentation in the past and how they’re trying to expand this definition in the present and future. “We have a lot of people who are approaching this from more traditional segmentation: What are the four buckets that we’re going to build large campaigns around?” Sahagian says. “Today the availability of intelligence about your customers and the tools to leverage that data have become so much better that you can do a [better] job of reaching out to specific individuals and understanding their needs.”
Understanding existing customers’ needs is important, too, because, as a credit union, BCU legally can’t sign up “anybody and everybody on the street,” Sahagian notes.
BCU used its understanding of customers’ financial needs to launch a campaign promoting its new PowerPlus Checking product this past March that targeted 14 different segments.
Before introducing PowerPlus Checking, BCU had two different products for its checking accounts: an ATM reimbursement product and an interest product. If customers made a certain number of transactions within a month, BCU would let those customers choose between the products, which would result in them either receiving a 2% APY on their balances up to $25,000 or up to $10 in reimbursements for non-BCU ATM fees, Sahagian says. However, with PowerPlus Checking, BCU brought these two rewards together so that customers could have both.
To show customers how PowerPlus Checking could serve their needs and lifestyles, BCU broke the campaign down into four key segments: notify, activation, acquisition, and upgrade. Then the credit union broke these groups down into 14 micro-segments by looking at customers’ language, activity level, and age of their account. So, if a customer already had one of the two checking products and was actively using it, BCU would automatically convert that customer to the new account and notify them of the “good news,” Sahagian says. However, if the customer wasn’t using his existing account frequently (or at all), BCU would send him an activation communication encouraging him to start using the new checking product and taking advantage of its new perks, he adds. As for customers who didn’t already have one of the two checking products, he says, BCU would take the opportunity to try to get them to open a new account through an acquisition communication. Finally, BCU targeted customers who had checking products that didn’t offer any rewards, but they met the activity requirements to earn the ones PowerPlus Checking offered. According to Sahagian, BCU sent these customers a communication encouraging them to upgrade their account and stop “leaving financial rewards on the table.” Customers were contacted via direct mail, email, or outbound call for these segmented communications.
Because the campaign is still in market, BCU is waiting on results, Sahagian says; however, he plans to analyze retention rate and transactional activity for people who already had a checking account, as well as open and activation rates for customers who did not. In addition, he says that the credit union will send out surveys and look at Net Promoter Scores to better determine customer loyalty and perceived value.
BCU is aiming to extend its segmentation across all of its channels within the next five years. It’s started taking baby steps toward this initiative, which include rebuilding its website late last year. The next goal is to connect the website and the credit union’s other marketing channels in its CMS and Adobe Campaign to deliver relevant content no matter where customers are.
“If I present to you a credit card offer on the website this morning and you don’t engage with it or you engage with it but say, ‘That’s not what I’m interested in,’ when you stop into the branch later today, I want to make sure that the information the branch has on you is updated to recognize that,” Sahagian says. “Because if they come after you and offer that same thing that you said no to this morning, you’re going to say, ‘Well, you don’t know me as a customer.’”
Segmentation based on behavior, needs, and profitability
Customers who fall into the same segment don’t always bring the same level of profitability. For BP, this may mean that two customers who fill up at the same gas station and are members of the same rewards program are of different value: one may fill up with premium gas every week while the other purchases regular unleaded fuel once a month. That’s why the fuel company relies on behavioral-, needs-, and profitability-based segmentation to identify its highest-value customers and retain them through relevant messaging.
For BP, customer loyalty is all about repeat visits. “If you had four opportunities a month to fill up with fuel, we want you to come to us four out of four times,” says Jonathan Lee, brand communications manager for BP Fuels Communication. “That’s ultimate loyalty for us.”
However, industry challenges have made driving these repeat visits difficult. Consumers now own more fuel-efficient vehicles, which results in more infrequent visits to the pump, and gas prices can deter people from driving altogether, Lee said at the 2015 Adobe Summit. This volume drop, in turn, creates more competition for BP and a dire need to hold onto the customers that it already has (plus, pull in new ones from competitors), he added. BP entices its customers to come back, he said, through offers and its loyalty program, BP Driver Rewards—a cents-off-per-gallon incentive that launched in April 2013.
But launching a loyalty program proved to be a bumpy ride for BP. Three main CRM issues complicated the fuel company’s ability to learn about its reward members and measure the program’s effectiveness. First, BP had a disconnected view of its customers because it couldn’t link its attitudinal data (information from in-depth interviews, focus groups, and surveys) to its behavioral (online interactions) or transactional data. Second, BP didn’t have any place to put all of this data and make data-driven decisions. Finally, the company had limited measurement in place: While it could track simple metrics like clicks and impressions, it wasn’t able to see the big picture and determine whether its marketing was effective, Lee explained.
This disconnected view of the customer caused BP’s communications to suffer. When it came to email, for instance, the fuel company would only send messages based on transactions. So, if a member filled up her car with gas, which was logged via her rewards card, BP would send that customer an email acknowledging her pump visit and notifying her how close she was to her next reward, Lee said. If a member wasn’t redeeming rewards, BP would then send her some “nudges” encouraging her to do so or reminders to use them before they expire, he explained.
“It’s pretty basic,” Lee said. “It’s also pretty inflexible.”
So, BP outlined new goals for its CRM strategy. The fuel company wanted to have integrated campaigns that not only incorporated multichannel messaging, but also brought multiple data points back in. In addition, it wanted to leverage CRM to talk to segments in individualized ways and provide offers that were appropriate based on behaviors, Lee said.
By November 2013 BP had submitted an RFP for CRM solutions. The company decided to work with Epsilon and began building a CRM solution with the data-driven agency in July 2014 before launching the solution that December with Adobe Campaign, which it uses as an ESP and to store its segments.
For BP, segmentation begins at a basic transactional level. The fuel company looks at purchase behaviors to determine the VAP (value, attrition, risk, and potential) of each Driver Rewards member, as well as where he is in his lifecycle.
For instance, BP considers Driver Rewards members as new for 60 days after they register. Although that period is too early to assign value to this segment, the fuel company then breaks it down into two micro-segments based on transactional information: new start (members who haven’t made a purchase in that 60 days) and new active (members who have made at least one purchase within that 60 days), Kathryn Zajac VP and GM of client services for Epsilon said during the Summit.
As with the new member segment, the active segment is then broken down into two smaller segments: single-visit members (those who have only made one purchase within the past 12 months) and multi-visit (those who have made two or more transactions within the past 12 months). BP also assigns risk and value levels to these segments based on their likelihood to purchase and estimated incremental value. The final segment is the inactive segment, which consists of members who haven’t made any transactions within the past two to 12 months.
“We’re reaching each of those segments with different messages and different offers and prioritizing frequency, cadence, and content based on that,” Zajac said.
But after defining these lifecycle segments, BP wanted to take its segmentation even further. So, it implemented Epsilon’s Niches tool, which accesses its data via a “clustering system of U.S. households,” Zajac explained, that maps them out across 26 different niches, or personas, based on five main factors: target income, household age, home ownership, presence of children, and economic data (like transaction and credit card data). The fuel company can also gain insights into members’ lifestyles, hobbies, and interests.
BP Rewards members fall into five key niches: Kiddie Kastles, Quiet Homebodies, Big Spender Parents, Value Focused, and Doing Well and Donating. Most of these niches, Zajac noted, consist of affluent consumers. Take the Kiddie Kastles, for instance. This niche is made up of middle-age parents who have older children (who may still be living at home), are white-collar, financially savvy, and enjoy outdoor activities. While this group makes up 9.4% of the U.S. population, it comprises 12% of BP’s Driver Rewards program. “This is important data to have as you’re building out the CRM strategy and what the marketing plan is going to be,” Zajac said, “because it can shape what the messaging should be [and] what the offers should be.”
In addition to BP’s first-party data and Epsilon’s data, the fuel company purchases third-party data from other sources, like automotive data from Polk (now part of HIS Automotive).
With a new CRM strategy in place, BP is working to develop a new email strategy—one that not only drives members to the pump, but also prevents customers from falling out of touch with the brand. The new email strategy is broken into four key focus areas: onboarding, engagement, retention, and opportunistic. When customers sign up for Driver Rewards, BP will send them a series of welcome emails to educate them about the program, such as by sending a new member checklist or instructions on how to use the card at the pump. “We need to do things with a little extra care in that beginning time period to help that behavior change take place,” BP’s Lee says.
Once BP detects a pattern of activity, or even a pattern of inactivity, it puts the new member into a segment. For instance, if a customer enrolls in the program and doesn’t visit the pump, BP will send him a reminder one week, two weeks, and three weeks after sign up. Or if a customer does visit the pump, BP will send him a “nudge” email on the day of the member’s next estimated fill up based on his behavior pattern. BP also sends a monthly newsletter to active members.
However, if a customer starts falling out of his regular activity pattern, BP will go into retention mode. For instance, it may send a fill-up reminder email, help them find a nearby station, or send an offer if the fuel company considers them to be high value.
As for the opportunistic emails, BP will send specific promotions to particular niches. For example, the company may try to upsell its BP Visa card to its rewards members that fall into the Kiddie Kastle or Doing Well niches.
“Now we have the ability to match all of that information together and to send a custom offer to a segment that might be more likely to respond, instead of blanketing the whole group,” Lee said.
Although it’s still early days for BP, Lee said the company is already seeing higher engagement with its emails. BP is also attributing how these emails contribute to the overall goal of driving people to the pump by looking at the time between recipients opening those emails and a transaction.
And it’s just the beginning. BP intends to build out its CRM strategy to other channels, like mobile. “We know that consumers want to interact with us,” Lee says. “We think now, with adding that extra level of a little more customized [and] a little more one-to-one, they’ll be taking those actions.”
