Mass marketing techniques of yesteryear, powered largely by demographic data, very often missed their intended marks. The strategy was more like a scattershot approach to acquiring new customers. While effective at delivering messages to the masses, it failed miserably as a dual-directional mode of communication with consumers, making campaign effectiveness guesswork as marketers had difficulty tracking campaign results. That's not to say that demographic data was not valuable. It provides marketers much needed details such as age, sex and location of consumers, but it could not provide marketers with two critical marketing components: consumer behavior and attitudinal information.
To help companies improve customer relationships and profitability, in the late 1990s, experts and analysts pounded pulpits across the country to introduce the one-to-one marketing concept. One-to-one marketing is simply the ability to understand and treat each customer as an individual by tailoring products and solutions to each customer's needs.
Proponents of one-to-one marketing boasted of its ability to improve customer satisfaction levels, and they maintained customers would show their appreciation by rewarding vendors with more business.
But tailoring products and services to individual customers became too costly for many businesses. And marketers quickly retreated. However, many have new marching orders, which merge the simplicity and economic efficiencies of mass marketing with the accuracy of one-to-one marketing and incorporate the final piece of the marketing equation attitudinal data. Call it, one-to-some, one-to-few, life stage or life style marketing, the goal is the same: bring the right message to the right customer at the right time.
We asked respected experts and consultants, such as Barton Goldenberg, Robert Nascenzi, Martha Rogers, J. Walker Smith and others to share their views on the use of demographics and behavioral and attitudinal data in marketing. In the following pages, you'll find sound bites from our interviews on some of the most debated issues in marketing today.
Title: Product Manager, MapInfo
Focus Area: Geodemographics
Previous Employment: Urban Planning Consultancy, Compusearch, MapInfo
Education: BA in Geography from the University of Guelph, Ontario, Canada
Facing New Challenges
IN THE HYPER-COMPETITIVE MARketing landscape of the new millennium, marketing success and the need for clear differentiation has presented challenges like never before. Successful marketing efforts that boost revenue and patronage must be intensely focused and driven by accurate, timely and meaningful information about the marketplace and consumers. Demographics are an essential element in the mix of understanding target audiences past, present and future and play a defining role in the shaping of campaigns and messages that hit the mark and optimal real estate site selection. It is mission critical to gain an in-depth understanding of all facets of your business to best understand your markets, site opportunities and customers. Demographic products, including geodemographic segmentation, are the key to insightful market decisions on customers, prospects and markets, a thorough command of market and product potentials, clear assessments of site location opportunities and adding broad vision to target marketing campaigns.
If actual customer transactional data is available and used, then there is naturally a high degree of relevance and utility in one-to-one marketing applications that make use of such resources. The utility of customer data is paramount in cross-sell, up-sell and brand association opportunities. One-to-one that takes advantage of customer data can be very effective. I had a gas company mail me a postcard offer with an image and write-up about the car I had just purchased. This campaign was relevant and effectively hit the mark with me. On the other hand, when national consumer lists are utilized for this type of campaign, the data is not necessarily up-to-date, complete, relevant, specific or as detailed as recent point-of-sales records and risk missing the mark and intended audience.
One-to-one marketing can certainly be effective given adequate circumstances and data. However, the utility of geodemographics in providing a snapshot of who your customers are, where they come from and what else they like to do (read, watch, purchase) is a proven effective approach to get a handle on how to conduct meaningful marketing dialogue with the market and customers.
I believe there is a trend in the marketplace to conduct more one-to-one campaigns, as partly evidenced and fuelled by the increased availability of one-to-one information products, list companies/data mining companies and techniques. However, upon the realization that one-to-one requires a broader geodemographic influence to achieve success and response, we believe that the market will trend toward a mix of geographic and household direct mail for the accuracy of high-level geodemographics and the perceived precision of household level data. Our customers use geodemographic information products to narrow down their efforts to target neighborhoods, and then move to direct mail [addressed and unaddressed] to reach targets within select neighborhoods.
MapInfo has extensive experience in both the one-to-one and one-to-some applications of target marketing. We do recognize and are closely watching the trend in the marketplace to move to, or experiment with general household level information in target marketing. We have found that the one-to-one approach can be successful given the appropriate circumstances and accurate, relevant data ingredients. However, we do suggest and caution our clients in the use of national generic household level one-to-one products for several reasons. We suggest caution where the input or source data may not be complete and/or is subject to inaccuracies in reporting. We have found over the years, that there is no better data to predict consumer and market behavior than that of a robust customer data set, captured via loyalty program or transaction-based processes. In most cases, these data sets are not commercially available and may only be employed by the owner [or commissioned consultant such as MapInfo-Thompson] of the data in any analysis and marketing effort. Marketing systems and campaigns built in this manner have proved to be effective.
Title: Vice President, Gartner Inc.
Focus Areas: CRM business strategies and overall CRM vision
Education: MBA from the University of Chicago
A LOT OF MARKETERS THINK they're doing one-to-one, but they're really doing one-to-some. One-to-one is very theoretical. While firms are talking about -to-one they do instead common groupings that are identifiable, predictable and something they can implement against. And, as a result, they have a larger segment that they're marketing to.
They think they're trying to identify this is what we know about Scott Nelson and this is what we know he'll respond to. These are the creative elements we'll use to be most predictive. The reality is they don't know that much about him. They may have behavioral and transactional information, but it's pretty much a rearview mirror. They have historical data, but they don't know what's predictive. If you go to a wine Web site and look at domestic Merlots, the next time you visit the site it may display information about that. But if I'm looking for imported sparkling white wine, all the info you have about Merlots is useless. I could be shopping for someone else, trying something new or a friend could be using my account. So what they have is useless.
Key segments will want to pursue one-to-one. It's analogous to the way banks treat their customers. Highly profitable customers get personal attention. We'll continue to see that, but for the other segments it's not worth it to have the same deep level marketing strategy for them.
We're seeing more one-to-one marketing emerging in the wireless space for churn analysis. Retailers are starting to pursue it, but they're lacking in detailed information to do effective one-to-one, so they're more prone to use demographic information. One-to-some works best in any area that has a large mass market that needs a way to break it up, but not to the level of one-to-one.
Title: Founder and president of ISM Inc.
Focus Areas: CRM, sales and marketing strategies
Previous employment: Senior and management positions at the U.S. Department of State and Monsanto Europe S.A.
Education: BSc (Economics) with honors, from the Wharton School, University of Pennsylvania and a MSc (Economics) from the London School of Economics
The Way the World Works
DEMOGRAPHICS IS STILL AN important piece, but by no means sufficient for effective marketing. You need demographic information, behavioral and attitudinal information. Unless all three are looked at, you will not get a good view of the customer.
One of my customers, AAA Mid-Atlantic, has more than 3 million customer numbers. They need to know who their customers are, why and so forth. Their work was done on demographic and behavioral information, only. I helped segment the market by demographics, plus behavioral [what you purchased in the past], plus attitudinal [why you do what you do]. Then we brought all three together into a common data structure.
With attitudinal data you can break down the groups into smaller pieces to get a group of similar types of people with common values. For example, 52 percent of all Americans inside AAA's Mid-Atlantic region think of Disney World as a fun place, but 48 percent see kids as an obligation. So they were somewhat insulted by the [Disney] brochures we were sending out. There's a good example of how you overlay demographic data with attitudinal and behavioral data. It's shockingly accurate and that's the thing that many people don't get. Unless you have all three you cannot get a complete picture of your customer.
I never bought into the one-to-one concept. I bought into the value of segmenting markets and creating unique products and services for groups of customers that make up those segments. One-to-one can cost a lot of money, especially when dealing with fickle customers. A customer that's not happy can quickly add to the cost and you can quickly lose money. The concept is good, but one-to-one has its restrictions and limits. That's why segments of groups make a whole lot of sense.
The bottom line is people like to be treated differently, they like to feel as though they are important and unique and that's just the way the world works. The concept of segmentation is a sound, proven concept. The problem is that 80 percent of marketers talk about segmentation, but less than 20 percent of all marketers truly understand the value of segmentation and do it well. People segment too broadly. People don't take the time to challenge the segmentation and say Do we have the best segmentation? Even if they have the best segmentation, they need to take the time to listen to customers. Marketing is half art and half science and the art is not appreciated as much as it should be.
Title: VP of customer insight, Experian Marketing Services
Focus Areas: We're about helping clients optimize their interests through programs that help them to be seen by their customers as smarter and more valuable.
Previous employment: Hoffman-LaRoche, InfoUSA
Education: MA in statistics, BSc and BA in demography from Montclair State University, Montclair, N.J.
Making the Grade
DEMOGRAPHICS IS ABSOLUTEly essential. Companies that have been the most successful in building strong, lasting, profitable relationships with customers have used demographics as a key to that success. Transactions can tell you what someone bought, but they're not going to help you predict what people are going to do. Behavioral and lifestyles data can tell you what someone's attitudes are. But demographics are what give you the basic information and understanding of who your customers are. This is particularly critical in an era of depressed response rates and the clutter that consumers feel. Demographics can identify what their need states are, whether they're moving into a need state that will make them valuable to you as a customer, and then you add to that dimension those other data points that enable you to have relevance to that customer. So, demographics are very helpful in being predictive, because with them, you can anticipate where your customers' needs are evolving to.
One-to-one is a powerful part of the puzzle, because its promise is that you can manage relevancy and reach for each of your customers. But the reality today, although the push is toward one-to-one, is more like one-to-few, because you're able to manage more information, more data into your models, but you still need to scale efforts, and can't do that across millions and millions of customers. We've measured return on investment on various market intelligence sources, and we typically see a 30 percent to 40 percent lift on performance when we apply demographic segmentation models to the customer transaction and other behavioral data sets that are available.
Historically, we've been a one-channel, maybe two-channel world, because of the limitations we've had on information about our customers But now, as the question becomes, How do I manage complexity of information points powerfully? we're seeing the scenario of experimental design make its way into the direct marketing ecosystem, and just as there are multiple information and data points about each customer, you're seeing multiple channels being tested to reach them.
We're seeing a great demand for the skills involved in managing the complexity of information, and it's a priority, but companies have their everyday business to conduct, so it's not going to happen all at once. Demand is increasing among companies for marketing intelligence and customer insight as companies focus on the core businesses that will make them grow. It's a question of migrating those imperatives into building the models balanced so that you've got transactional data as one dimension, demographic data as a second dimension and psychographics as a third dimension of insight. As far as apportioning resources, transactional data is part of the operational infrastructure of most companies, so it's a question of adding demographic and psychographic insight resources to pull out the greatest value from the intelligence. Some are more advanced than others, but I would say that as an industry, we're at a grade C or C+, with a lot of work to do and a lot of opportunity to get the balances right.
Title: Industry analyst at Yankee Group
Focus Areas: CRM, sales and marketing effectiveness
Previous employment: Consultant at Blanc & Otis; Product marketing management for Computer Corp. of America. Marketing management at Datamirror
Education: MBA Simmons College, Simmons Graduate School of Management
The Fine Line
DEMOGRAPHIC DATA IS ONE piece of the puzzle. You can't get rid of demographic data. You can't get away from demographic data because there are so many benefits to having regional information. Plus, no one is really doing effective one-to-one messaging. You really need clean customer data and clean real-time data, like what happened in the last hour and day and it has to be segmented properly.
We've been talking about one-to-one for many years, but in reality they're doing one-to-some. We've gone beyond mass marketing, but we're not at one-to-one. It's a great mission, but to really be one-to- one, providing customized products and services, do we really need to customize to the nth degree? Is it really necessary?
You have to find that fine line that helps improve customer satisfaction and profitability. If you're doing one-to-one, you're maximizing customer satisfaction, but hindering profitability, so you really need to balance the two. The consumer doesn't need to give up all that information for successful one-to-one. So if marketers can deliver the right message at the right time the customers are happy and that can happen in one-to-some marketing. While I'm unique I still fit in an affinity group.
Title: President and CEO of Cohorts
Focus Areas: Database-marketing and consumer research
Previous employment: RL Polk & Co., National Demographics and Lifestyles
Education: MBA and Undergraduate degree in economics from the University of Denver.
The Sweet Spot
I SEE IT AS ACTUALLY NOT A world of two types of marketing, but a world of three types of marketing. Historically, we've had mass marketing. That was one message developed by a manufacturer, pushed out through mass media to everybody. Everyone gets the exact same message. In the early '90s, one-to-one marketing caught the attention of the marketing community, particularly the database marketing community. It put forth the promise of getting exactly the right message to exactly the right household and differentiating that message infinitely through the magic of database marketing. It really allows you to have more effective marketing and more relevance. The drawback of one-to-one is it's very expensive and very complex to implement. Marketing departments, even database marketing departments, aren't geared to send out a million different messages to a million different households even with e-mail and direct mail. Where mass marketing had the efficiencies and ability to deliver one message cheaply to every household, one-to-one offered the promise of delivering relevant messages uniquely to each unique household directly. I see one-to-some fitting distinctly in the middle of those and capturing the benefits of both. It allows you to have the relevance of one-to-one marketing because you're marketing to homogeneous groups of households and you can tailor the message and version of the offer. You're sending it to like household groups. It reduces the operational complexity dramatically because you're sending the same message to a whole bunch of the same households. Therefore it's lower cost than one-to-one marketing.
Demographic data can be utilized in all three types of marketing and usually is. In mass marketing demographic data is used to make decisions about what television shows to advertise on and what magazines to buy ads in. You're doing all that based on composite profiles of the readership or viewership. There's other data that needs to be incorporated. In the direct marketing space it's RFM: How recently did customers purchase from you, how [frequently] do they buy from you and how much [money] do they spend? Those are not demographic data elements, but they're all important to a database marketer.
For example, one of our clients is a cable television company. What they had been doing in the past was sending out direct mail campaign where every household in their service area received the same message. In a way that's mass database marketing. That's one message going to many. By using a combination of demographic and media consumption data like what cable channels households are more or less likely to watch, they were able to develop five broad groups of households. From that they were able to send five more relevant messages. They were able to change the messaging. They emphasized certain channels to certain households and other channels to other households based on the demo characteristics and cable viewing patterns to make the message more relevant. It produced a 20 percent increase in response rates. This cable marketer was a national marketer, which meant this campaign was sent to roughly 7 million households and it resulted in an ROI of 2,200 percent.
One of the things that's driven me throughout my career has been a desire to see the consumer knowledge gained through the market research discipline applied more effectively in the technologies that are used in the database-marketing environment. What I'm excited about is bringing those two disciplines together in one-to-some marketing.
Customers and Strangers
Title: Founding partner, Peppers & Rogers; adjunct professor at Duke University
Focus Areas: We're trying to help our companies optimize the value they create for their customers and the value their customers create for them. We like to help customers teach companies how to make money and grow.
Age: Of enlightenment, and of reason
Education: PhD, University of Tennessee, Communications/Marketing
DEMOGRAPHICS ARE, IN MY MIND, A DESCRIPTION OF THE FUNDAMENTALS OF who we are and how we fit into the society in which we live. So I can't imagine anybody living without it, unless you're a hermit. Since people are not hermits, they are described and understood in any number of ways. Demographics is one of those ways, and an important one. The only caveat is that we have to make sure that we don't rely entirely on demographics.
Increasingly, our understanding of each customer will help us decide how to group those customers. Eventually, what we know about all of them will be superseded by what we know about each of them.
If customers were all alike, then all we would do to compete as a company would be to try to keep our costs down as low as we can it would all be about commoditization and operational efficiency. Only because our customers are different that we're able to meet their needs in terms of better products, that we're able to have better services for them.
What do I know about a customer that will help me be more valuable to that customer next? I'm not going to just determine that by what ZIP code they live in. I'm going to determine it by what I know about them otherwise, some of which is demographic, some behavioral, some we can only get from customers, some we can only get from third party data. If we're assuming that demographic information is available everywhere, what that gives me is a necessary floor of information. It's that balance we need to strike, between what do we need to know because everybody can know that about our customer? And what do we need to know that nobody else will know about our customers?
A customer used to refer only to somebody who paid us money. Then there were other people who could pay us money: prospects. Now, our customers are anybody that we can identify. If I've got your name and address, and I can figure out anything about you, whether you'll be valuable to me, then you're a customer of mine. I haven't sold you anything yet, so I have a 0 percent share of your business. But you're still a customer if I've identified you.
Other than customers, who exists on the planet? Strangers. So it's not so much that there are customers and prospects, as customers and strangers. And some of the customers we have are measurable greater than 0 share, and some of them we have a 0 percent share of customer, and our goal with every one of them is to increase the share of business. Now, acquisition has a two-pronged effect. One is to go into that group of customers that group of people we've identified who haven't bought anything yet and increase our share of their business. And the other function of acquiring customers is to get total strangers to raise their hands and identify themselves.
In a world in which we can measure a lot of different data points about each of our millions of customers, any one of those measures can be less accurate, and yet the entire numbers start coming out to something that's more believable. So it takes a lot more of the risk of analysis out if I'm looking at more data points about every customer. Some of them imply future value. Some of them imply needs, and my ability to meet those needs implies future value. It becomes very important that we are able to look at many data points across millions, maybe tens of millions of customers, and we may be able to come up with a number that is more accurate and more predictive of future value.
Title: President and CEO of Claritas
Focus Areas: customer acquisition, cross-sell and retention and site planning services
Previous employment: National Decision Systems, Innovative Systems Inc., Bank of Boston, Chase Econometrics
Education: BA in Economics, Boston College, MBA from the Wharton School, University of Pennsylvania
Identifying the Problem
DEMOGRAPHIC DATA IS critical. Behavior and purchase data only tell you what happened. Not what could happen. Demographics let you size the market to tell you the opportunity. Info such as age, income, homeownership, etc., all have a significant impact on purchase behavior, media preferences, lifestyles. Can a company survive without it, they could probably survive. Can they be successful and thrive? No.
Demographics are the foundation for predicting consumer behavior. One-to-one tends to be impractical in certain applications, like in acquiring new customers, or it's generally not efficient when it's an infrequent purchase such as autos, appliances, insurance. Even when cross-selling, if you know that product X buyers also buy product Y, that's very useful, and it's probably a good indicator that you can sell product Y to people who've bought product X. Demographics can provide additional insight into why certain product X buyers didn't buy product Y. Maybe some of those who didn't buy product Y are younger, or don't have kids, or are less affluent, or whatever, there's some characteristic as to why they didn't buy, and why others do.
While in theory, you could handcraft a message under the one-to-one approach to each individual customer and get a better response, it's generally not cost-effective. It's much more effective to come up with three, four or five different messages or offers and then send them to segments or groups of customers.
What we see is that one-to-one works the best, and is used the most at point of contact, call centers or on the Web. Netflex and Amazon are good examples. But one-to-many is better for just about all other applications. Even in one-to-one, I would argue that knowing the demographics is very helpful. The more frequent the purchase, the better one-to-one is, because you're capturing much more behavioral information.
With insurance, it's an acquisition game, and one-to-one isn't that practical, because how often do you buy life insurance? They're trying to acquire customers and they don't know all that information about you, so they have to use something like demographics. Whether you own a dog or a cat doesn't mean anything, it's whether you have children and need income protection, or whether you're buying life insurance because of tax implications for a trust, it's demographically driven. It's driven by age, income, presence of children, etc.
Even when you go to fast-moving consumer goods, demographics are key drivers. You're not buying diapers if you don't have kids. Age and presence of children are very strong drivers in the fast-moving consumer goods market, as is urban-icity.
The biggest challenge we have is educating people on how segmentation can be used to link all of the different aspects of their marketing programs. One of the beauties of segmentation is to link all the different programs together and have a common denominator to link across Not only can you do a direct mail campaign, you can look at their media preferences, you can look at category management applications, you can look at whether they buy through a catalog, online or in a physical store. It gives you the ability to link together all your different marketing activities with a common denominator on how to describe the customer you're marketing to, if you're using it for site analysis, why aren't you also using it for direct targeting, for the merchandising, or the menu selection, or the advertising for that store, the free-standing inserts for that store? You're using the common links, and you're also being consistent in your message. That's our challenge.
J. Walker Smith
Title: President of Yankelovich Partners
Focus Areas: Marketing services firm, specializing in database solutions and lifestyle trends.
Previous employment: Director of research at Dow Brands, S.A.
Education: Doctorate in mass communications research from the University of North Carolina in Chapel Hill.
THE BIGGEST SINGLE factor that we have seen emerging is the breakdown of demographics as useful proxies for attitudes. Marketers should be marketing on the basis of attitudes, but they run programs based on demographics. It's not the best thing to do, but we've grown accustomed to using demographics to get what we want, which is information about attitudes. For example, 42 percent of people in a recent census study refused to check a box for race because they didn't think any categorized them. To say there are black, white and Hispanic attitudes doesn't capture the marketplace where culture is shared, so it's harder for demographics to operate as a good proxy for attitudes. That's not to say that life stage information is not important. As life stages change your needs change. But there is a better way to collect information, investing in the development of systems that can link attitudes to individual names and addresses. It is a strategic focus of Yankelovich, we work with clients to score attitudinal information of people.
Most CRM systems are not set up to gather attitudinal information. That's one of the things we talk to our clients about. There hasn't been a CRM system that hasn't been built on the basis of demographics, categorical information and transactional history. If you use behavioral insight as attitudinal information, you go on treacherous grounds it adds to the clutter problem. Unfortunately, marketing campaigns go out to people who look the same, but [these campaigns] are irrelevant to many because there's a gap in knowledge. The gap that CRM systems have is a lack of attitudinal information. CRM systems could be used to fill in that gap but by and large that's not how they're being used. So it's no surprise that we're seeing productivity issues with all forms of marketing, CRM included. Increasingly, marketing isn't as relevant as consumers would like it to be and consumers are getting better at shutting off things that are viewed as clutter. The theory of CRM is if you know a lot about your customer, you'll be able to target market, but it's not the practice. Consumers are still looking for better precision and relevance from CRM marketers and marketers can't close that gap because they're missing a major piece of data: attitudes associated with addresses.'