How To Do Customer Segmentation Right

Many segmentation efforts today are an exercise in futility because companies are basing their segments on inappropriate criteria.

RBC Royal Bank is one of the few companies that actually gets it. Here's how the bank has grown its market share by courting specific customer segments, such as retired snowbirds and future doctors.


  • How understanding your customers' needs can boost profits
  • How to segment customers according to their necessities at each life stage
  • Why segmenting by age or revenue alone is ineffective

If banks could choose their customers the way kids choose sides on the playground, customers in the 18-to-35 age bracket would be picked last. With their relatively small incomes, low account balances and large student loan debts, young customers aren't exactly the sort over whom the average bank salivates.

At RBC Royal Bank, however, executives recognized that some of those impecunious young customers might eventually turn into wealthy, profitable customers. So RBC analysts pored through the bank's data on its young customers looking for subsegments with a strong potential for rapid income growth. Their analysis identified medical school and dental school students and interns as a group with a high potential to turn into profitable customers. So in 2004 the bank put together a program to address the financial needs of credit-strapped young medical professionals, including help with student loans, loans for medical equipment for new practices and initial mortgages for their first offices. Within a year, RBC's market share among customers in this subsegment has shot up from 2 percent to 18 percent, and the revenue per client is now 3.7 times that of the average customer. Martin Lippert, vice chairman and CIO at RBC Financial Group, says the bank's willingness to help these young professionals get started will likely be rewarded with a lower attrition rate down the road.

"We may have customers we're not making money on, but we look at that as more our problem than the customer's," Lippert says. "Our opportunity lies in finding what the needs of the customer might be so we can offer them additional products and get them to a point where we're making some return."

While lots of companies claim they're customer-centric, RBC is one of just a handful of organizations that segment customers based on customer needs, not their own. And by focusing its operations on addressing those needs, RBC has grown its market capitalization from $US18 billion almost six years ago to close to $US50 billion today.

So far, few companies are as sophisticated at segmenting customers as RBC. Many don't do any customer segmentation at all, and those that do typically don't reap much value from the exercise because they segment on the wrong criteria. Precise, needs-based customer segmentation is time-consuming and difficult, and very much in its infancy. But it's worth doing because it enables cost-effective targeting of customers with product and service offerings that match their needs. That kind of precise targeting obviates spending a bundle on largely ineffective mass mailings - and alienating customers with irrelevant offers. It's the quintessential win-win: Customers get what they want and subsequently buy more; companies waste less money and increase sales and profits.

"The more you're able to do for the customer, the more likely she is to pay attention to the next offer," says Martha Rogers, co-author of Return on Customer and co-founder of Peppers & Rogers Group. "Yet companies are not doing nearly so much as they can."

The Wrong Way to Segment Customers

Many segmentation efforts today are an exercise in futility because companies are basing their segments on inappropriate criteria, says Larry Selden, co-author of Angel Customers & Demon Customers and professor emeritus of finance and economics at Columbia Business School. As a result, organizations often wind up with segments that drain resources, instead of with segments that lead to more effective ways of running the business or meeting customers' needs. For convenience, companies that are organized along product lines often segment customers by the products they buy. This approach, however, risks alienating customers in two ways: Customers who happen to be in more than one segment get bombarded with multiple uncoordinated offers. And big spenders in one product category who start buying in a second category are justifiably miffed when they're treated as strangers.

Segmenting by demographics is also quite common, but it's generally not useful unless customer needs happen to align neatly with demographic characteristics. Lego customers' needs, for example, do tend to shift with age. Preschoolers, after all, play very differently from kids who are between 5 and 14 years old (what Lego calls the in-school segment). And the 30,000-plus adult fans of Legos tend to be hobbyists with a completely different mind-set altogether. Yet considering age alone isn't sufficient; Lego also looks at what users do with their bricks. In-school kids who focus on building when they play will likely want plain bricks, but those who focus on role-playing usually gravitate toward themed sets. Cases in which demographics alone are an indicator of a common need are generally rare.

A lot of companies segment customers by revenue, intuitively assuming that revenue is a good indicator of profit. But, Selden argues, that's hardly ever the case. He maintains that an effective segmentation strategy should begin with a profitability analysis, divvying customers into 10 deciles ranging from most to least profitable. When he segmented one major retailer's customers by revenue, some that had the largest revenue generated among the lowest gross profits. And to get a true picture of profitability, banks need to think about the amount of capital they must allocate to high-risk customers.

It's not a given that all or even most customers within a certain profitability decile are necessarily alike. Even so, understanding which customers are profitable and which aren't is a good starting point. The trick is to delve into each profitability segment to look for hints of possible subsegments, that is, customers whose behaviour patterns or other shared characteristics suggest they might have common unmet needs. Once RBC identified the shared unmet needs of young medical professionals, it was able to put together targeted offers to meet those needs and increase the profitability of that subsegment. "The goal for segmentation is to put customers in homogeneous groups based on common needs and wants that you can act on with a common solution," says Selden. "So who are all the people you can go at with a common offer that will make you a boatload of money?"

Page Break

The Royal Bank Way

Determining your customers' needs is not a onetime exercise. Although this means you can never be done with the process of needs identification, the good news is that you don't have to be perfect on the first go-round. Effective segmentation is an exercise in fine-tuning. For instance, RBC started back in 1992 with just three customer segments: high, medium or lower profitability. Over time, RBC's segmentation process has become much more sophisticated. Today the bank has more than 80 customer models in its data warehouse, and each month it scores all of its eligible customers on all relevant strategic and tactical models. (Someone who already has a line of credit at RBC, for example, would not be scored against a model that predicts the likelihood of acquiring a line of credit. And customers who have opted out of having RBC use their information for promotional purposes aren't scored at all.) Strategic models - including profitability, life stage, potential, defection risk, client commitment or loyalty and overall risk - help the bank home in on customers' needs and priorities. Tactical models - such as propensity to buy, the likelihood of a customer cancelling a product or service, and the degree to which a customer uses the products he's acquired - are used to identify revenue opportunities and generate lead lists for employees who deal directly with customers. Scoring customers monthly on the 80-plus models is helping RBC generate more than 13 million targeted leads each month. Roughly 6 million leads are used by salespeople to reach out to customers. The other 7 million are called "sales opportunities" to be used when the customer initiates contact with RBC. That means customer service reps and branch employees have targeted product pitches to offer customers after they've handled the customer request.

The three most critical models that drive RBC's business are profit potential, current profitability and life stage, says Gaetane Lefebvre, vice president of client knowledge and insights at RBC Financial Group. "If you have these three things, you can fundamentally manage your business very well," she says. "Those are the ones we've been using as a proof of concept since as early as 1996." RBC's method of projecting potential for each customer is so proprietary that Lefebvre won't even discuss it, aside from saying that it is not simply a lifetime value calculation (see "The Lifetime Value Equation", above ). "Profitability is extraordinarily important and it's where you want to start," says Lefebvre. "But it's not very informative for understanding client needs." For that, RBC relies heavily on its life-stage model. Using this model (supplemented by focus groups, surveys and third-party research) divides individual clients into five strategic life-stage segments:

1. Youth: These clients are younger than 18.

2. Getting Started: These clients, generally between 18 and 35, are going through first experiences: graduation, first credit card, first car, first loan, marriage, first child.

3. Builders: These clients, usually between 35 and 50, are in their peak earning years. Typically they borrow more than they invest, as they build families and careers. With many expenses, their primary goal is to manage their debt load effectively.

4. Accumulators: Typically between 50 and 60, these clients are worried about saving for retirement and investing wisely. They want to know if they've saved enough to retire, if they'll have to change their lifestyle when they retire and if they'll need to work to supplement their retirement income.

5. Preservers: The primary needs of these clients, who are usually older than 60, are to maximize retirement income and maintain the lifestyle they desire. They typically manage multiple income sources and are starting to do estate planning.

Lefebvre and her team overlay these life-stage segments with other strategic models such as profitability, potential, client credit risk and client vulnerability (risk of leaving the organization) onto the bank's objectives: retaining profitable customers, growing customers with potential, managing and controlling customers with higher credit risk profiles and optimizing the costs of less profitable customers. By doing so, they can identify opportunities to make a difference in the market, she says. Once the bank has identified a high-potential opportunity, it models the opportunity to see how much it might grow the business. Then RBC fine-tunes and validates the offer with 100 to 200 customers in focus groups or client interviews. If the offering is complex or demands significant investments of bank resources, RBC will often verify the results through further qualitative research or a pilot, testing different offers and creative among thousands of customers before rolling out the optimal version on a larger scale.

Targeting the Snowbirds

Once you've identified a group of customers who appear to have common needs, you have to determine if you can profitably offer a value proposition to meet those needs. It's all about finding the ideal middle ground between segments of one (it would be too expensive to address customers' needs individually) and segments that are so large and heterogeneous that you can't tailor offerings to customers' needs. Sometimes it's not worth subsegmenting your customers. (Selden observes that Wal-Mart essentially has one segment of 200 million, based on the assumption that everyone just wants low prices, period.)

"Normally, companies start with a relatively small number of segments, which are fairly coarse," says Selden. "But if you have 40 million customers and five segments with 8 million customers per segment, the chances of those being highly homogeneous are very limited." The goal, then, should be to evolve those segments into more precise subsegments that allow you to deliver more targeted value propositions.

"Subsegmenting," Selden says, "is where the gold is."

Defining a useful subsegment generally involves doing a deep dive into the most profitable end of the segment to tease out distinct behaviour patterns. On delving deeper into RBC's preservers segment, Lippert says, RBC noticed a subsegment of people who spent a lot of time out of the country in certain months. Many of these were snowbirds escaping to Florida to avoid the harsh Canadian winters. Because RBC has branches in the United States, the bank quickly realized that snowbirds represented a sweet spot of untapped potential for the bank.

To address these customers' unmet needs, RBC put together a "snowbird package" that included travel health insurance, easy access to Canadian funds, online consolidated account review, real-time transfers, the ability to leverage a Canadian credit history to secure mortgages in the US and a toll-free number for cross-border banking questions. RBC also began introducing customers in the snowbird subsegment to personal bankers in the US, making it clear that the institution knows its customers and understands the importance of their business.

Page Break

Catering to snowbirds has resulted in a higher than average number of products per client in that subsegment. And for RBC, that translated to a 250 percent increase in net income per client before taxes. Equally stunning is the 45 percent decrease in the defection rate among the snowbirds. Because acquiring a new customer costs RBC a projected five to 10 times more than holding on to an existing customer, Lippert says that a reduction in the defection rate adds significantly to the bank's overall P&L. The package has been so successful that the bank now offers a similar RBC Access USA package to other groups of customers, such as students and executives, who spend a lot of time in the US.

Although the snowbird subsegment turned out to be highly profitable, unearthing other subsegments may reveal that they are financial drains on the institution. When RBC uncovers unprofitable behaviour patterns, it looks for ways to more efficiently address those customers' needs. For example, the least profitable group within the preservers segment turned out to contain a number of Canadian retirees whose fixed incomes plummeted in value when interest rates fell and stock returns diminished. Because of the poor return on their investments, they were highly dissatisfied - as well as frustrated with the limited advice offered by financial institutions. Although unprofitable, they were valuable customers who carried twice as many products and services as the average preserver. It turned out they were keeping large balances in short-term guaranteed investment certificates (GICs, the Canadian equivalent of a fixed-term deposit) that they were rolling over instead of cashing in. And because they were good at negotiating higher rates when they rolled over their GICs, they were that much more unprofitable for RBC. So the bank developed a group of cash-flow model portfolios to offer these customers. The portfolios, which typically include mutual funds as well as GICs and vary by level of risk and expected return on investment, deliver a better return as well as tax breaks. The customers are happy because they make more money and get to keep more of what they've invested. And within two years, RBC has generated 21,000 new retirement-income plans and achieved a net growth of $US1 billion in account balances.

The Value of All That Slicing and Dicing

Companies that get the most from their segmentation strategies don't just pay lip service to the importance of segmentation, they organize their operations around addressing customer needs. At RBC, a senior manager with P&L responsibility manages each segment. In addition, RBC offers very specific, actionable data to customer-facing employees.

"Different people can interpret data in a different manner," says Lefebvre. "We didn't want to provide data that anyone would have to interpret. So we give them something very specific, such as: 'For this client, offer a pre-approved line of credit', or 'Call this client about a registered investment [The Canadian equivalent of the employer superannuation policy]'." So whether a customer dials in to the call centre, visits a branch or talks to a manager, she will be given the same offer because the employee who interacts with her will be prompted by the CRM system to do so. "Delivering data on a real-time basis to reps when they are engaging the client has a tremendous lift," says Lippert. "Some organizations are pushing home equity lines this month, credit cards next month. Our folks are asking clients about particular products that we have intelligence are the ones that those clients are likely to purchase."

By segmenting its customers to offer them targeted, relevant offers, RBC's personal and commercial division reached its goal of increasing revenue by $US1 billion. Since October 2003, client defection has also decreased from 8.4 percent to 6.2 percent, while the number of high-value clients has increased from 17.1 percent to 19.1 percent of the client base today. RBC also boasts a return on equity of nearly 25 percent.

Page Break

"We have a culture that recognizes that what's in the information vault is as critical to us as what's in the money vault," says Lippert. "It's important to understand that [segmentation is] not typically a year-one, year-two payback kind of investment. It's something that gets better with time and gets better with the organization's ability to understand the data."

Lefebvre says that companies just starting down the customer segmentation path should view it as an evolutionary process and just jump in and begin. "Don't wait for everything to be perfect," she says, "and don't wait for the next piece of data before you do things. Often you can improve the business successfully in a rudimentary way. Ten years ago we were not as granular.

"Segmentation," she says, "is a journey."

SIDEBAR: The Lifetime Value Equation

In financial services, calculating the lifetime value of customers is something of a misnomer, because few if any banks attempt to project customer value beyond five years. The calculation typically involves looking at a customer's age, tenure, and number of products and services used, as well as his propensity to acquire additional products and services minus the risk that he'll defect on his current products and services. The value of the customer's projected portfolio can then be calculated using typical profitability figures for each product and service.

SIDEBAR: Tips for Successful Segmentation

1.Put each customer in only one segment. Otherwise, customers can get bombarded with multiple, uncoordinated offers.

2.Be channel-neutral. Customers should get the same offers no matter which channel they use. So custom product recommendations should be available to all customer-facing employees and delivered to customers who go online.

3.Give customer-facing employees specific, action-oriented intelligence. Don't give them data that's open to interpretation. Tell them exactly which offer is most appropriate for each customer.

4.At the outset of customer segmentation, give the sales force only the best leads to ensure a very high success rate. Once the sales force sees the value of segment-driven sales recommendations, you can expand the lead list to include more customers, giving a "propensity to buy" score for each so that reps understand what level of receptivity to expect.

5.Give a senior manager P&L responsibility for each segment.

6.Put senior management in charge of driving segmentation. As RBC Financial Group vice chairman and CIO Martin Lippert notes, companies may miss part of the customer perspective if only a single line of business is pushing segmentation. Also, segmentation is less prone to budgetary constraints if it's funded by the enterprise instead of by a single group.

7.Start small, then evolve. First divide customers into a few coarse segments, then gradually break them into smaller, more precise subsegments. But don't wait until you've got it all perfect. Just begin.