Fall 2016
Insights and Applications from Experts Over Four Decades
ABOUT THIS PUBLICATION
The Analytics Journal focuses on how marketers and business leaders use data to reveal, target, engage and retain their customers. This edition is devoted to a retrospective of the concepts, applications and case studies that have helped shape the marketing analytics industry and that continue to play important roles in advancing data-driven decision making. The Journal features articles that have appeared previously in Canadian Direct Marketing News (Direct Marketing) and other industry publications over the past four decades. We welcome your comments. --The Editors
An Environics Analytics Publication Produced by Lloydmedia Inc.
302-137 Main Street North, Markham ON, L3P 1Y2 • 905.201.6600 President - Steve Lloyd • Editor - Sarah O’Connor • Creative Director - Jennifer O’Neill
Environics Analytics 33 Bloor Street East, Suite 400, Toronto, ON M4W 3H1 • Phone Toll Free: 1-888-339-3304 environicsanalytics.com • inquiries@environicsanalytics.com President and CEO - Jan Kestle • Editor - Michael J. Weiss © 2016 Environics Analytics
3
President’s letter In association with Lloydmedia Inc., I am pleased to present this issue of The Analytics Journal, a compendium of articles about database marketing, customer insights and analytics written by industry leaders over the past four decades. While we are all acutely aware of the profound changes that have marked consumer behaviour and the proliferation of media choices over the years, we should also acknowledge the best practices in analytics that remain true today: • Gather as much data as you can—a variety of types from reputable sources—to understand consumers and markets. Today we just have more choices than ever. • Clean, integrate and enhance your data to create a comprehensive view of the customer. That goal remains the same even though with so much data available, it’s still not widely achieved. • Incorporate geographic location data in your analytics. It’s been the secret sauce in database marketing all along simply because everything has to happen somewhere. While these articles from the industry’s archive present principles and methodologies first articulated many years ago, it would be wrong to think there has been no progress in the discipline. Rather, this retrospective collection highlights the history of best practices and related tools that can be applied in our new age of data-driven marketing, leveraging Big Data and using data and analytics to make our organizations more competitive and more successful. Indeed, the tried and true techniques found in these articles can be updated, enhanced and combined with new methods and new types of data. As the saying goes, sometimes what’s old is new again. So enjoy and learn from the past while you forge a new path on your analytics journey. We thank the Lloydmedia team for joining us in this effort, we salute the authors who have over the decades contributed to our industry in Canada, and we particularly honour the memory of Bruce Carroll and Robin Page—both trailblazers in our industry. A small sampling of their thought leadership appears in this journal. The compendium concludes with a current article by Richard Webber, a 40-year veteran of the data-driven marketing industry who offers his unique perspective on where this journey has led us to today. It is my hope that his thoughts, as well as those of all the authors presented here, will inspire everyone. Regards,
Jan Kestle President and CEO Environics Analytics THE ANALYTICS JOURNAL ❯ Fall 2016
4
Contents 3
President’s letter
6
PRIZM: An introduction
By Robin Page
10
Linking today’s direct marketing to new technology
By Paul R. Lacey
13
Turning data into assets
By R. Bruce Carroll
How to make your data clean, smart and accessible to more people
18
The changing world of target marketing
By Jan Kestle
20
Alphabet card deck strikes tourism chord, ECHO gold
By Susan Maclean
23
Environics Analytics launches ENVISION, a web-based micromarketing tool that provides ease of use
An interview with Gary Wood
Fall 2016 ❮ THE ANALYTICS JOURNAL
5
26
When unaddressed mail produces upmarket results
Holt Renfrew finds surprising success with unaddressed catalogues
By Michele Sexsmith
29
Geodemography Q&A with Jan Kestle
32
Direct mail devotees in the digital age
Just because they’ve gone digital doesn’t mean consumers don’t like direct mail
By Jan Kestle and Michael J. Weiss
35
On becoming an analytics organization
By Catherine Pearson
38
Big Data analytics: Isn’t it still just analytics?
Tactics must evolve but the mission hasn’t changed
By Richard Boire
41
The value matrix
By Stephen Shaw
45
Why geodemographics is relevant in an age of Big Data
By Richard Webber
THE ANALYTICS JOURNAL ❯ Fall 2016
1982 Originally published in 1982 in PRIZM: Geo-Demographic Market Segmentation & Targeting by Claritas Corporation.
PRIZM: An introduction
By Robin Page
T
he goals of marketing and marketing research are swiftly changing. Over the past three decades, sales objectives could be met by exploiting growth, riding the crest of a spectacular climb in our national product and its consuming population. Broad distribution of mass-appeal products appeared to be the winning formula for exploiting this seemingly limitless uptrend. Advertising and promotion campaigns were geared to national audiences. Media were selected for “reach” rather than efficiency. Today, however, the marketer faces a new reality. Inflation has cut deeply into consumer confidence and effective buying power. Recession has tightened price elasticity. Environmental and political issues have made us critically aware of values hitherto ignored in the buying decision. Population has reached zero growth. The “good life” seems remote. The sobering situation plainly means an end to riding the “growth gravy train.” The pie isn’t getting uniformly bigger anymore. Marketers must now fight for their slice. Future sales growth and profits must be engineered by identifying high-potential market segments and by rifling desired products and selling propositions to these segments with maximum efficiency. The traditional techniques of marketing research were never equipped for such a task. Sample surveys were classically designed to measure national sales trends and consumption patterns but never to discover the actual, local communities and neighborhoods where high-potential consumers live and shop, let alone predict their purchase behavior. Instead, all findings were reduced to standard univariate demographic scales (e.g. age, sex, income, etc.), which prove to be, by far, the weakest links in the marketing information chain. With univariate demography, market segmentation is usually approached by “cross tabbing,” a process which can quickly reduce a survey sample to statistically unstable fragments. For example, the marketer who fields a typical 2,000-respondent survey, and asks for five breaks on age, two on sex, five on income, two on race, four on family size, and three Fall 2016 ❮ THE ANALYTICS JOURNAL
7
on child status, must be extremely cautious with cross tabs. He must not forget that the possible combinations are 5 x 2 x 5 x 2 x 4 x 3, or 1,200, over half the size of his total sample! He may also face problems in translation. A pinpoint cross-tab segment could easily leave him with a primary target representing under 10% of the consuming population… and what of the other 90%? What of his secondary market? To obtain a valid, realistic marketing target, he must compromise. Moreover, with few exceptions, mass media have not been susceptible to pinpoint cross tabbing. Measurement standards vary widely, buying practices interfere, and, where there is a choice, that choice may be limited to a few breaks on age and sex. This dilemma—the need to define and reach mass consumer markets from cross-tab segments—explains why several thousand brands have shared the same, diffuse, tiresome target of “Women, 18-49, w/Above Average HH Incomes,” why some $60 billion in shotgun media have been aimed at this mythical bird over the past two decades and what old John Wanamaker meant when he said, “I know half the money I spend on advertising is wasted, but I’ve never been able to find out which half.” It also explains why contemporary marketers, in their genuine alarm over across-the-board inflation, have been seeking advanced targeting methods. They demand hard answers. Who are my prime targets? Where do these people live and shop? What marketing mix will win them over? How do I reach them most efficiently, with the greatest possible return on my marketing investment? How can I reduce waste, and build my bottom line? PRIZM was created to answer these questions. Its entire output is decision data, and it works. It works because ZIP Codes in the U.S. are nothing less than perfect five-digit computer links, links which are today connected to virtually all forms of marketing data. ZIP’s link… Every consumer household, and the 100% U.S. Census measures of household demography which describe them, with… Every marketing research respondent, whether for products, advertising, media audiences, motives, or opinions, with… Every consumer purchase record, from credit cards to warranty cards, from audited circulations to auto registrations, with… Every retail dealer, or service outlet, and their immediate neighborhood trading areas, with… Any defined geographic area, be it a city, country, TV market, SMSA, state, region, sales district, or franchise territory. This linkage clearly predicted a powerful software system. All it lacked was method. The key to utilizing this wealth of data—forging it into a marketer’s tool—was the creation of statistically homogeneous ZIP-Market Clusters by Claritas Corporation. Briefly: Over 1,000 U.S. Census demographic measures were prepared for each of the nation’s 35,600 residential ZIP-Code neighborhoods—a vital statistical database. Every ZIP-Code neighborhood was then computer analyzed, classified and assigned to one of 40 homogeneous groups, called ZIP-Market Clusters. Once created, each ZIP-Market Cluster was fully documented to present a holistic demographic portrait of a distinct American neighborhood lifestyle. These 40 lifestyle Clusters were then completely cross-referenced, via their assigned ZIP Codes, to all standard market definitions (countries, SMSA’s, TV markets, states, etc.). THE ANALYTICS JOURNAL ❯ Fall 2016
8
The finished Cluster System was then linked (by consumer-respondent ZIP Codes) to several of the nation’s leading syndications of product and media data (e.g. SMRB, R.L. Polk, Arbitron, NFO, NPD, etc.), and to hundreds of client-supplied surveys and customer files. The end result was PRIZM. PRIZM (which stands for Potential Rating Index of Zip Markets) is “Geo-Demography,” a wholly new perspective on market segmentation, which permits all forms of ZIP-identified marketing data to be linked, correlated, profiled, and evaluated across 40 homogeneous neighborhood Clusters, in addition to traditional, univariate demographic breaks. The system is complete—tested, proven, and double-edged. The Clusters produce highly discriminative profiles of behavior towards products, services, media and promotions, and are documented in vivid demographic detail. Therefore, by definition: PRIZM is a 40-point, lifestyle segmentation system strong enough to explain, and accurately predict, consumer behavior. And because we know the exact location of every ZIP neighborhood within Clusters, we can project this consumer behavior back into local market geographies, and maps, thus, by construction, PRIZM is also a versatile and efficient targeting tool, geo-coded to actual consumer neighborhoods, and the retail outlets which serve them. These are ‘living’ data, clear reflections of life, and of the evident link between social structure and consumer behavior. These are ‘action’ data, which permit the marketer, at long last, to trade in his shotgun for a rifle, knowing with clarity WHO his prime targets are, WHERE these targets live, and HOW to reach them most efficiently.
How PRIZM works PRIZM begins where it should, with the casual relationship between human behavior and community structure—a branch of sociology sometimes called “human ecology.” Fundamental to this science is the observation that human beings, obedient to their cultural origins and natures, are predominantly tribal, territorial, and socially hierarchical. Indeed, this behavior is so evident, so pervasive, as to seem routine. At each phase of the adult life cycle, each rung on the social ladder, we strive to emulate, dwell amongst, and be accepted by our perceived peers. We have behaved in like manner throughout recorded history. We do so today, and will predictably continue. The first principal of human ecology is, therefore, that people with similar cultural backgrounds, circumstances, and perspectives will ‘cluster’, i.e., will group together to form relatively homogenous communities. This is a global phenomenon, and is nowhere better illustrated than by the fascinating diversity of neighborhood lifestyles in America. The second principle of human ecology is that once established, the demographic character of a neighborhood tends to persist over time, i.e., is inherent and self-perpetuating. This is largely due to fixed neighborhood assets, those generally classified under the purview of “real estate,” such as housing stock and land values, the commercial infrastructure, the adjacent job pool and transportation network, the school system, zoning laws, and other factors which are stable over decades. But also compelling is social estate—the economic opportunities, the factors of race, language and cultural origin, the stratified lifestyles and ambience, which support and Fall 2016 ❮ THE ANALYTICS JOURNAL
9
advertise neighborhood status and appeal. A common misconception holds that a highly mobile population must, in a decade, contradict the stability of neighborhoods. However, while Americans are among the world’s more mobile people, demographers have found that the inherent lifestyles of the vast majority of our neighborhoods change rather slowly— rarely in less than a generation, usually longer. The correct picture of mobility is one in which people move through neighborhoods in search of compatible lifestyles. Young couples give up the excitement of downtown apartment living, and move to the suburbs to bear children. As family size/incomes increase, they move up to larger homes in more exclusive surroundings. Divorces take their toll, sending fractured families to neighborhoods better suited to fresh starts. Many return to the city once their nest is empty. Others pull up roots, and move to retirement communities in mild southern climes. Predictably, as such homes are vacated, they are filled by new families seeking the very lifestyles abandoned by former tenants. And so the neighborhood abides, and keeps its character. Given the above—the natural formation and enduring character of homogeneous neighborhoods in an otherwise heterogeneous society—the third principle of human ecology follows logically. Simply stated, the demographic measures which accurately describe our place of residence will provide an equally valid statistical mirror of our ‘place’ in society. This is an equation of utmost significance to domestic marketers. For in America, ‘place’ in society is neither dictated by authority, nor handed down by blood and title. Instead, it is earned by achievement, and demonstrated by consumption. This is our cultural paradigm, or “mind set”—our motive force—and it has created the world’s highest standard of living. We are under five per cent of the world’s people, consuming almost 20% of its total material resources. We are a nation of consumers, and our entire conjoint system of attitudes, beliefs, desires, and expectations hinges on this point. And it is on this point that human ecology turns from social science to marketing science—to the casual relationship between community structure and consumer behavior.
Update Robin Page enjoyed a long and illustrious career in the geodemographic industry. A former Coca-Cola adman from Atlanta, he is considered by many the godfather of lifestyle naming in segmentation systems. During stints at Claritas Corporation, Compusearch and Acxiom, he gave the world such enduring segment names as Blue Blood Estates, Bohemian Mix, Money & Brains and—in Canada—Brie & Chablis and Mortgaged in Suburbia. For the first PRIZM system, Kids & Cul-de-Sacs was among his classic creations; for US MOSAIC, Country Clubbers perfectly captured the life of upscale duffers; and for PSYTE USA, he christened a segment dominated by legions of babies as Gerberville. At one point in his career, he’d named the lifestyle types of the three largest U.S. segmentation systems, the equivalent of one person naming every car model from the Big Three automakers. Robin Page died in 2004.
THE ANALYTICS JOURNAL ❯ Fall 2016
1989 Originally published in the March 1989 edition of Canadian Direct Marketing News magazine.
Linking today’s direct marketing to new technology By Paul R. Lacey
T
he new economy is an information economy, with the demand for information being virtually limitless. The largest, most profitable and fastest-growing area of the information business is marketing oriented information. New breakthroughs occurring in this area are technical breakthroughs: they are mostly in market data-collection, improved message transmission and evaluation of marketing efforts. And behind the scenes, Canada’s low-profile postal code facilitates these marketing innovations with modern computer technology. Today “targeting” and “accountability” are constantly in the vocabulary of major marketing and advertising companies. Canadian companies are estimated to spend over $5 billion each year (out of $8 billion in 1988) on mass market advertising, that, by its very nature, cannot meet either of these requirements particularly well. Mass advertising offers little targeting ability to advertisers who want to reach individual businesses or consumers: it provides even less accountability for the money spent. The overall effect of this estimated $5 billion annual expenditure is measured simply in terms of audience “awareness” and “recall.” On the other hand, there is direct marketing. Everything about direct marketing is directly accountable, directly measurable, directly related to hard numbers on a corporate bottom line. Instead of just creating ideas, the direct marketer can actually influence behaviour, and do that in an observable way. In the area of behaviour, direct marketers are learning that beyond targeting, there is also the capability to predict buying behaviour. But whether it is to target, to predict buying behaviour, or simply to provide good old bottom-line accountability for sales, direct marketing is now being adopted by many Canadian companies. Canada Post supports direct marketing now only as the message carrier but increasingly as the provider of vital postal code information which accurately targets messages.
Fall 2016 ❮ THE ANALYTICS JOURNAL
11
Integrating the computer database “Databased direct marketing” is expanding rapidly with the involvement of many major support industry sectors, including direct response advertising agencies, computer service bureau and mailing list suppliers. Databased direct marketing immediately repositions these companies and support functions as part of a corporation’s marketing strategy, rather than as promotional after-thoughts. Databases are bodies of information organized and made accessible so many people can use the information for many purposes. That means that a simple mailing list can be used as a database. Companies therefore may have been doing “databased marketing” in a way, for many years, and didn’t know it until the word got trendy. Databased direct marketing is, however, more than just a mailing file. In essence, it is information purchased or collected that can be analyzed to achieve maximum results using direct response advertising to elicit orders, sales leads, donations, or to build retail traffic.
New methods, new opportunities Increasing interest in direct marketing is forcing the evolution of two groups of advertising agencies. First, the traditional agencies, which make their revenue from commissions on media sales. They cannot target their advertising messages particularly well, since they use the shotgun approach. The second group of advertising agencies are direct response oriented. These have true targeting capability—they use the rifle approach. Direct response agencies can do direct mail production, some with their own affiliated printing operations. They can do mailing list processing, and many have in-house computer service divisions. With their skyrocketing billings, it’s not at all surprising that there are now more traditional agencies entering the direct marketing field. Today, the personal computer is in common use as a business tool, and can be readily applied to develop mailing lists. With something like half the lists now in use devoted to direct mail campaigns of fewer than 50,000 names, many mailers simply use a desktop microcomputer. The next five years will see more and more direct marketing software, as the computer, including the personal use computer, becomes more than a data manipulator and data management device. It will choose, judge, and “reason” its way among options and alternatives, based upon “experience” and interactive learning. Direct marketing will become a high-paying field, particularly for computer “techies.” Closely related to that, general advertising agencies will continue to open or purchase direct marketing specialty agencies. The bigger direct response agencies—those with no actual information-management capability—will expand and start to buy sophisticated computerized service companies. They will do this to fulfill the new business opportunities that “databased marketing” provides. Marketing is big business. In the next five years we are going to discover that “databased direct marketing” will be bigger, with Canada’s postal code system playing a crucial and even more visible role.
The humble postal code Databased marketing works because it’s efficient and reduces the advertiser’s margin of THE ANALYTICS JOURNAL ❯ Fall 2016
12
uncertainty. It’s better to have an idea of where potential customers live so effective selection can target those most likely to respond to each advertisement. Probably the most critical factor contributing to the success of a direct mail campaign is the mailing list. An integral part of such a list is the postal code. The Canadian postal code’s unique ability to isolate small geographical areas (as small as one face of a city block) can be married with census data to provide market statistics. Jointly, they can identify things like average age, income, occupation, how many children, what type of dwelling, how many bathrooms and so on. This data can then be matched with Statistics Canada consumer expenditure surveys, which tell how different people spend their disposable income. Firms that market products through direct mail maintain lists of their own customers and prospects that are easily identified by the postal code. Each month Canada Post supplies such firms a computerized list of all valid postal codes, and offers incentives to encourage their use in preparing mailings. Thus it is a relatively simple (and economic!) matter to match and select only those postal codes that have the market characteristics desired.
Selecting the target For example, M5G 3H6 could contain a cluster of urban young singles, who earn significant incomes and live in high rises. When direct marketers add their own records or purchased lists, (such as magazine subscriptions lists) to their file, this postal code area could be recognized as one where expensive stereo equipment is marketable. This knowledge also allows many marketers to pinpoint optimum locations for dealer outlets, franchises, and sales territories. An important part of any marketer’s success is product design and selection. Market research can be used to ensure that a product meets the demands of its intended market. New products are then selected to attract the interest of specific market segments. That’s part of the reason we have all become familiar with the overused expressions like “baby boomers,” “empty nesters,” and “young marrieds.” If a specific segment likes expensive stereo equipment, then the latest $600 disc player will be attractive to the high income, single baby boomer who may live in postal code areas such as M5G 3H6. Direct marketing provides businesses with an efficient way to market their products. Obviously, if a product has a proven attraction to young singles it would be very inefficient to advertise to people who are retired. Traditional mass marketing approaches offer broad spectrum audiences; usually little is known about these audiences, beyond research on the type of consumer likely to listen to radio or watch television. Businesses must direct their advertising to only those who are seriously inclined to buy. Direct marketing, supported by modern computer technology and the Canadian postal code, offers select, targetable, measurable and homogeneous markets. Paul A. Lacey was Manager, Product Planning/Addressed Admail, with Canada Post Corp., in Ottawa.
Fall 2016 ❮ THE ANALYTICS JOURNAL
1998 Originally published in the March 1998 edition of FYI, a newsletter published by Compusearch.
Turning data into assets How to make your data clean, smart and accessible to more people
By R. Bruce Carroll
W
hat is database marketing? Database marketing has become one of the important weapons in any marketer’s arsenal. At a minimum, database marketing requires collecting the names and addresses of your existing and potential customers, along with as much data as possible about their individual buying behavior. These data can be used to directly communicate with your customers, one at a time, in a timely and relevant way. Hence, the popular phrase: one-toone marketing. The term micromarketing is often used as a synonym for database marketing but it actually describes a whole series of activities that may or may not rely on a database of customers. For example, distributing unaddressed mail to households within one kilometre of stores where sales of a particular product index greater than 120 could be described as precision or micromarketing. But it is not database marketing. In any event, if there is some debate as to what it is, there is no debate as to why database marketing has become so important: You will understand your customers better. Here are some of the questions about your customers that database marketing enables you to answer: • What’s their buying behavior? Discover product affinities and gaps. Compare actual spending to potential. See who’s increasing (or decreasing) spending. • Who are your most loyal customers? Identify and rank customers by their value based on spending volumes, purchase behavior, order size and demographics. • What are the key customer segments? Identify the key customer groups by their demographic and behavioural characteristics.
THE ANALYTICS JOURNAL ❯ Fall 2016
14
• Who are your most profitable customers? Calculate the profitability of individual customers and identify the most valuable ones. You will be able to target your customers with greater precision than ever before. Database marketing allows you to: • Manage customer relationships. Set behavioural “triggers” that initiate an automatic marketing response to certain events or conditions. • Measure your marketing success. Maintain promotion and response histories to learn which customers responded and what marketing efforts worked best. • Create more relevant promotions. Tailor messages and strategies to specific customer segments, resulting in more relevant communications and higher response rates. • Predict customer behavior. Create models to predict who is most likely to respond— or leave. • Develop preferred customer programs. Reward and solidify customer loyalty and provide preferred treatment to your best customers. • Select the right location. Evaluate the revenue potential of new retail sites, and select products and services that will be successful in both our new and current sites.
File preparation The most basic step in any database marketing program is to ensure that your customer files are accurate, up to date and uniform. Regardless of the size of your database, you should be running your files through all the processes outlined below. Why bother? Well, what’s a customer’s name worth to you over the course of two, three, four or five years? There are six basic steps to file cleansing. You can use all of them or you can select any combination of the following individual options.
National Change of Address (NCOA) Keeping track of your customers’ current address is a key component of good file hygiene. Each month, Prospects updates its own database by incorporating addresses of people who have moved. Building on this in-house process, we have created the NCOA file, which contains changes of address going back to February 1990. Currently the file has over 4.5 million records and is the only file in Canada large enough to be used for NCOA purposes.
Merge/purge A staple of any customer-based marketing system is the identification and/or elimination of duplicate names/addresses within a single file or across multiple files. Prospects has purchased the best software on the market (Group One) for updating its own files (over 700,000 transactions monthly) and uses this same software on behalf of our clients.
Householding Using the same process as above, Prospects can identify all individuals living within a single household. By taking all the information appended to individual records, Prospects can then Fall 2016 ❮ THE ANALYTICS JOURNAL
15
aggregate the data up into a single household record. This is extremely important if you want to calculate the total purchasing power of an individual household.
Deliverability verification Exclusively ours, Deliverability Verification identifies which records on your customer file match the Prospects file. Since our file is updated monthly, any of your customer records that match our file are considered deliverable. By identifying deliverable records, you can reactivate inactive files. Those names cost a lot to get, so keep them!
Postal code verification and enhancement This process verifies customers’ existing postal codes and corrects those that are correctible. Why bother? Because Canada Post makes many postal code changes throughout the year. Last year, for example, the rural codes in Stoney Creek, Ontario were completely changed in keeping with Canada Post’s plan to re-code rural postal codes across the country. That means in any given year, entire towns on your file could become undeliverable. We recommend verifying postal codes at least annually.
Address Accuracy Since Canada Post implemented its address accuracy program three years ago, mailers now have to ensure that addresses are at least 90% accurate or else pay a postage premium when mailing bulk. Prospects can verify that existing addresses are valid, according to Canada Post standards, as well as correct those addresses that are correctable. For example, an address for one of your customers who lives in Preston, Ontario may appear to be accurate. But, since Canada Post is aiming to put all addresses into consistent formats, Cambridge is actually the correct city name. Address Accuracy not only improves the deliverability of your files, but also improves THE ANALYTICS JOURNAL â?Ż Fall 2016
16
matching within and across other files during the merge/purge process. For example, the unit or apartment number of an address should always appear in the same position, to create the uniformity needed to more accurately match addresses. Every time we run a file through the Address Accuracy program, Prospects also produces a statement of accuracy—or Software Evaluation and Recognition Program (SERP) report—which Canada Post requires to qualify a mailing for bulk postage rates.
Data enhancement The addition of external “overlay variables” to customer files is not as popular in Canada as it is in the United States or Europe, where virtually every file is enhanced to some degree with external data. This is difficult to understand because these data can add a lot of marketing intelligence and depth to any file. For instance, they can be used to help model distinct custom segments or clusters, improve cross selling to existing customers and target new prospects. In short, you can better identify who your customers are, what they enjoy doing and what their households look like by adding information to each individual record on your file.
Market analysis… MarketMath Right now there are approximately a dozen or so GIS mapping systems on the market. There could be 20 by the time you read this. Virtually all of these systems are light years beyond where such systems were even two or three years ago. Given the fact that they are all in Windows, you can do amazing things. Quite frankly, I have trouble telling them apart and, for anyone interested in buying a mapping system, I really think you can’t go far wrong choosing any one of them. Our staff uses about four in our own work. We chose ESRI’s ArcView GIS system as the map engine in MarketMath but that choice had more to do with some of the special technical advantages of its Avenue programming language than with any fundamental advantage in its basic mapping capabilities. Although, to be sure, it is one elegant system. About a year and a half ago, we reached the conclusion that although there were all these wonderful mapping systems, none of them delivered what might be described as clusterbased customer profiling, market analysis and basic marketing arithmetic. Furthermore, in an intensive survey of customer needs, we found that very few of our customers described themselves as GIS experts or experienced in operating GIS systems. Integrating complex customer databases with the necessary geographic and demographic data was not their mission in life, but simply the means to an end in terms of contributing to their company’s bottom line. They wanted to point to a particular database, to a market, to an application or to a map type and get what they want. They did not want to write SQL statements to some so-called “Remote Spatial Query Engine” or custom design a report or map. They wanted the 10 map formats that would serve most of their needs and they wanted to get at those 10 maps with a click. They wanted a Windows interface and a short learning curve. They wanted to be able to import and export data in a variety of files. They wanted to geocode customer records and create cluster profiles in their PC. Most of all they wanted desktop access to our extensive data files. Fall 2016 ❮ THE ANALYTICS JOURNAL
17
So… we built MarketMath And rather than trying to cobble a restrained relational database component into MarketMath for managing client files, we built a hot icon link between MarketMath and MarketLink. As a result, and as already discussed, marketers can subset customer records by geography, value, product or any overlay variable in MarketLink and then pass this selection to MarketMath for PSYTE cluster profiling, area or product ranking, market potential calculations and correlations to other products, etc., etc. There are several other advantages to MarketMath, which may not be readily apparent. When we launch a new database, we’re now able to cut a new CD and send it out to our MarketMath clients that week. When there is a bug, we can fix it immediately. Because we have control over the whole programming environment in MarketMath, we’re in a better position to constantly improve the product according to the needs of our clients. At this point you, our clients, are now the designers of this product! In MarketMath, literally dozens of analytical reports and maps are pre-programmed for you. No programming is required; you just point and click your way through the data. The analytical routines are geared to a variety of industry vertical markets and MarketMath “macros,” “wizards” and “mapping scripts” can easily be added by our client services team to fit your specific requirements.
The bottom line Successful database-driven marketing requires a new way of thinking: You have to start thinking of customer information as an asset. And you have to regard maximizing your return on this asset as one of your principal tasks.
Update A pioneer in segmentation and marketing analytics, Bruce Carroll was an innovative leader in the marketing services industry throughout North America and Europe. He co-founded Claritas Corporation and served as its President until it was acquired by the VNU Group in 1990. A key developer of the original U.S. PRIZM segmentation system in the 1970s, Bruce returned to his native Canada in 1991 to serve as President of Blackburn Marketing Services Inc., which included Compusearch, then Canada’s leading marketing services provider. When Environics Analytics was created in 2003, Bruce served as a trusted advisor to the company. After his death in 2015, EA President Jan Kestle said his greatest gift to the industry was his ability to create new, insightful services without abandoning the proven methodologies of the past. “He was the master of Big Data and analytics before the terms were even coined,” she observed.
THE ANALYTICS JOURNAL ❯ Fall 2016
1998 Originally published in the June 1998 edition of FYI, a newsletter published by Compusearch.
The changing world of target marketing By Jan Kestle
M
arketers know that the consumer world is becoming increasingly fragmented, requiring more sophisticated tools to reach smaller and smaller segments. Channels are changing: the net and the store, how do they work together? What are the latest techniques in channel and category management? New locations are popping up on large and medium-sized chains, and more dollars are earmarked for better site-selection decisions. All major companies are investing in enterprise-wide databases. Marketers are looking to leverage the database not only for one-to-one purposes but also to help with a full range of marketing-related decisions. It’s in this context that Compusearch is seeking input from our customers and users regarding future directions and product development plans through vehicles such as the user conference and this newsletter. Geodemography is a relatively young industry and Compusearch goes back to the very beginnings—a little “younger” than Claritas and ESRI, a little “older” than MapInfo. And the first 20 years saw the steady development of geodemographic products: trade areas of census data, current demographics, clusters—first Lifestyles and then PSYTE—links to important third-party survey and location databases, boundary files and streets, and desktop software to make it all meaningful. Through this whole period (the ‘70s and ‘80s) computing power became more accessible and affordable. Database marketing, involving modeling and mining of customer records, became the “alternative” to geodemography. At least some people attempted to position it that way. But what really happened is that the most data-literate users combined their customer records with the geographic overlays, and geodemographic applications continued to grow in scope and use.
Fall 2016 ❮ THE ANALYTICS JOURNAL
19
The integration of geodemography with database marketing Compusearch’s response was to develop tools to facilitate the integration of these two worlds— database marketing and geodemography—and it’s still our business development focus today. The executions don’t look much different as a result of this strategy but they are based on better quality input data, often on actual customer purchase behavior. A flyer or promo piece that’s dropped to routes based on PSYTE-cluster-plus-dollars-spent-in-the-store will be a better target for most applications than one based on clusters alone. And so we have progressed rapidly down this path, developing products that make this integration possible. These include better geocoding resulting from a much better street product (CompuStreets), expansion of consensual data programs like ConsumerLink, new wealth and disposable income data, enhancements to MarketLink, and a full blown data hygiene and enhancement service. And we’ve thrown a lot of hardware at it (fast Alpha servers). Have all of our geodemographics customers adopted the integration with database marketing? By no means: of our 200 largest customers, about 10% have gone this route so far. But those who have find it’s resulting in big payoffs. And after a little over a year of offering this “full service” approach, Data Services and MarketLink are our fastest growing areas of business. The challenge as we pursue this strategy is twofold: to endure that we have the critical mass of brain power and tools behind the analytics components and to stay abreast or ahead of our customer requirements in terms of technology.
Looking ahead Our analysts and methodologists, led and mentored by Dr. Tony Lea, are a top-notch group of MAs and PhDs who keep our products and services moving forward. We are staffed now so that we can offer their expertise in custom analysis, complementing our standard research services and desktop systems. And we are still growing—bringing experts on board every week to respond to our customers’ complex challenges. We’ve launched our first web-based delivery tool, DemographicsCanada.com. But that is only the beginning. The next 12 to 18 months will see the migration of our tools to the web. The transition will be orderly and modular. The 5.5 Release of MarketMath and the 4.7 Release of MarketLink will form the foundation of the next generation of products. Our goal: a product that helps you access our unit record data—TotaList and The Lifestyle Selector—combined with our aggregate data, including PSYTE, seamlessly working with your data (which can be in our computers or yours), and all of this “spatially” enabled. Applets like profile, query and rank will be modular and available via the net—either because they can be accommodated in a thin client application or because they are embedded in a middle tier that “looks after” some of the more complicated math. Different types of users such as regional managers or executives will access just those reports and pieces of information that they need. That’s our vision. The technology touchstones are “three tiers,” “data anywhere,” “spatially enabled” and “distributed access.” Throughout the course of our conference, we received a lot of encouragement regarding this direction. And we’ve continued to hear from our customers about their needs and suggestions in terms of this next phase of product design. We are now proceeding to formal design meetings and we want to hear from you as the “specs” evolve and so we can continue to meet your needs, current and future. THE ANALYTICS JOURNAL ❯ Fall 2016
2008 Originally published in the April 2008 edition of Direct Marketing magazine.
Alphabet card deck strikes tourism chord, ECHO gold By Susan Maclean
W
ith its mandate to promote the growth and development of the tourism industry in the province, Tourism BC faced the challenge of a decline in visits to the province by Americans and “long haul” Canadians—fellow residents who live a multi-day driving distance away. The decline was attributed to confusion regarding U.S.-Canada passport requirements and the exchange rate which had changed dramatically—and would even more, later—from the U.S. dollar worth 1.60 Canadian dollars in January 2002. Not much could be done about those two causes but a third reason for the decline—a “been there, done that” indifference— was definitely something that marketing could address. Countering that “already done that” factor among Americans meant positioning B.C. “as a diverse neighbour with a unique trinity of cities, mountains and water making B.C. the perfect place to refresh, recharge and renew,” says Joel Tkach, marketing manager, consumer programs, North America, Tourism BC. “Among Canadians, it required communicating “British Columbia’s diversity to show it’s a lot more than a pretty face.” In that context, Tourism BC agency Blitz Direct created the direct mail campaign “Get to Know BC from A to Z”—with that letter pronounced as “zee” for the sake of rhyming and appealing to Americans. Tracey Chapple, Blitz Direct general manager, reports that the objectives were to entice prospects to order the BC Escapes Guide and to learn more about the spectacular beauty and vast array of destinations and activities that B.C. has to offer. She adds that another objective was to achieve a 15% response rate and generate $1 million in tourism revenue. A multi-wave direct mailing matched elements of B.C. with each letter of the alphabet. But this wasn’t just A for “Aquarium,” B for “Bowen Island,” C for “Capilano” and so on. A is for “Adrenaline-pumping adventure,” B for “Beaches,” C for “California roll”… all the way to Z is for Fall 2016 ❮ THE ANALYTICS JOURNAL
21
“Zzzzs,” the latter referring to comfortable lodgings. “We had lots of discussion about the right mix of A-Z items,” Tkach reveals. “It was important to include many B.C. icons as consumer feedback consistently tells us that’s the primary appeal of our destination. However, we wanted to balance that with the unknown and even the obscure. “The best debate was around the letter ‘W’. The original idea was “Wreck Beach,” which is a legal clothing-optional wilderness-surrounded beach on the shores of the University of British Columbia. It’s safe to say that not everyone was on side with this so we decided to go with “West Coast cuisine” instead. The nail in the coffin of the Wreck Beach concept was the suggestion that my butt be used so we could avoid any model release issues,” he comically recalls. A novel focus group test was held at the Toronto Yonge & Bloor subway station where “consumers were happy to tell us their ‘B.C. A to Z’ ideas as well,” he adds. Using an in-house database of past bookers and past inquirers based in the U.S. or Canada, males and females aged 35 to 55 with mid-high dual incomes who had indicated they enjoy travel and adventure were divided into two groups totalling about 67,000 prospects.
Dimensional mailing The tier one group of slightly more than 36,000 received the dimensional mailing with the getaway contest to encourage response. It included the following: • An attractive 4.5” x 6” thick envelope sealed with a resealing sticker promoting Super, Natural British Columbia; on the address side was the notice of a chance to win one of three B.C. getaways and the other side showed a man looking over a lake and mountain scene and the words “Get to know…” • A letter on card stock from the vice president of marketing at Tourism BC; it unfolds to reveal a self-mailing, postage-paid request for free BC Escapes Guides and a chance to enter the getaway contest. • A deck of 27 3.5” x 5” cards, one title card and one for each letter of the alphabet, are held together by a clear elastic band. The cards are full-coloured and contain friendly, chatty copy that also invites the recipient to contact the Web site HelloBC.com or 1-800 number for the guide. The title card tops explain “our first limited collection of BC: A to Z.” • A brochure slightly larger than the cards when folded promises “Your letter-perfect vacation begins here.” It opens half-way to the same mountainous terrain reflected in the lake as on the cover and the letters B and C, then opens to its full 14.5 inch width to show the A card on the far left and the Z card at the right with details on the guides and booking vacations. Recipients could respond through one of three channels: a personalized postage-paid business reply card, toll-free number or online landing page.
Postcard mailing The tier two prospects were sent a 6” x 9” postcard that showed the colourful cards as a page of peel-away stickers on one side and, on the address side, a list of what each letter stands for and an invitation to click or call for information or to book a “letter-perfect vacation.” THE ANALYTICS JOURNAL ❯ Fall 2016
22
The postcard was also sent as a follow-up to non-responders among the premium group four weeks after the initial mailing. “Many visitors to the province have a general perception of what they believe British Columbia has to offer,” Chapple explains. “While this has worked to our benefit in the past, we wanted to show potential visitors who may not have considered B.C. as a primary vacation destination that there is much more to learn about B.C. than they might think. In doing so, we wanted to use a device that was not only educational but also engaging. “The creative hit the mark for both U.S. and Canada, which isn’t always the case,” says Tkach. “For Canada, we wanted to demonstrate that British Columbia was much more than a pretty face and worth visiting despite the novelty of a strong Canadian dollar that opens up many U.S. destinations. For the U.S., we wanted to talk to people who have been here before and show that we’re worth a repeat visit despite their dollar not going as far as it once did and the ever-changing passport requirements (or lack thereof ).”
Response and ROI The campaign pulled in a 21% response rate from the core direct mail audience, greatly exceeding the objective. As well, 40% of direct mail respondents travelled to/within B.C. generating $763,000 in hotel booking revenue with additional spending estimated at about $2 million—for a 12:1 return on investment. “We’re huge fans of direct response marketing, and direct mail is a big part of that equation,” Tkach concludes. “The BC Escapes budget is a $6 million dollar investment for us, of which, about 16% is spent on direct marketing, inclusive of electronic and traditional direct mail.” Susan Maclean was a freelance writer and editor of Contact Management magazine.
Update Since this article was first published, Destination BC replaced Tourism BC as the crown corporation responsible for promoting tourism in British Columbia. In an effort to create a strong, united marketing network, Destination BC recently began working collaboratively with the province’s regional and city-based destination marketing organizations (DMOs) to better understand prospective visitors, such as skiers in Canada and the U.S. In one initiative, analysts linked Destination Canada’s “Explorer Quotient” (EQ) segmentation system to EA’s PRIZM system to help the province understand the needs and motivations of target travelers. The target segments allowed local DMOs to develop more targeted marketing campaigns and share a common language with each other, with Destination BC, and, through the EQ-to-PRIZM link, with Destination Canada. The goal: “Elevate the brand and profile iconic experiences to trigger a strong emotional desire to travel to B.C.”
Fall 2016 ❮ THE ANALYTICS JOURNAL
2009 Originally published in the January 2009 edition of Direct Marketing magazine.
Environics Analytics launches ENVISION, a web-based micromarketing tool that provides ease of use An interview with developer Gary Wood
W
hen Gary Wood set out to create a new micromarketing platform, he aimed high. “I wanted a solution that would offer advanced applications but not require an advanced degree to operate,” says Wood, vice president of Custom Applications at Environics Analytics. With this month’s launch of ENVISION—Environics Analytics’ web-based, userfriendly analytics tool for Canada—Wood appears to have succeeded. At the heart of ENVISION is a comprehensive set of demographic, PRIZMC2, marketing and media databases that power a suite of “one-click business applications” for creating custom maps, customer profiles, executive summary reports and market, behaviour and product rankings. To Wood, an accomplished software developer and the chief architect in the 1990s of Compusearch’s MarketMath and Zephyr desktop systems, developing an online tool that can analyze customers, prospects and trade areas across Canada was a result of the technological changes going on in the industry. Recently, Direct Marketing sat down with Wood to learn more about this solution.
Q: Why does the industry need another micromarketing tool? A: Most of the existing systems are complex add-ons to GIS technology. That’s fine if you have a GIS technician on staff who is accessible to marketers. While GIS-based analysis is essential to many companies, for others marketers just want results. They want to see maps and reports to figure out what it means to their business. With ENVISION, you just start your browser and, with a few clicks of your mouse, you can get a marketing report about your customers and trade areas.
Q: So is it fair to say that, with ENVISION, Environics Analytics is moving away from GIS-based solutions? A: Not at all. We also offer powerful GIS-based desktop systems and partner with providers of THE ANALYTICS JOURNAL ❯ Fall 2016
24
server-based GIS. But we wanted to have different ways to deliver our data and analytics that meet a wide variety of needs. ENVISION is adding a new approach to our suite of analytics delivery systems.
Q: What are the major features ENVISION offers? A: It’s all about ease of use. Everybody’s got a Web browser so there’s nothing to install. The average entry-level marketing analyst could learn how to use the system and be productive in about 15 minutes. ENVISION integrates your customer and location data with its own geodemographic data. If you give ENVISION a file of customers and postal codes, it will geocode them with PRIZMC2 and produce a customer segmentation profile in minutes that will tell you all about your customer’s preferences and habits through PRIZMC2-linked surveys. And you’ll have a profile that you can retrieve immediately, tomorrow or in a month.
Q: What are some examples of how ENVISION can be used? A: In the retail industry, ENVISION can help companies understand their customers and purchasing habits. In banking, it can identify site locations for opening and closing branches and installing ATMs. And in the public sector, it can pinpoint areas of need for delivering community services such as health care, senior assistance or child care. One of ENVISION’s features is that it customizes reports and maps to suit an industry’s terminology. So a report to a fundraising user will describe donors, not customers, and the map for a government user will discuss catchment areas instead of trade areas.
Q: How would you describe the user interface? A: Intuitive and minimalist. With a few clicks of the mouse, you can get a cluster profile of your customers. And then with a few more, you can see where they live, what they’re buying, which Social Values they score high on and what media they prefer according to BBM RTS Canada and PMB surveys. It’s fast and comprehensive—the whole package.
Q: What’s the logic behind ENVISION’s architecture and user interface? A: It’s designed with wizards that guide the user in creating profiles and analyses. The user interface prompts users when to import data and then offers some typical market analysis workflows ranging from standard to increasingly complex analyses. But the wizards always let you go back and modify your initial parameters. If you’re doing a trade area analysis and you realize that you’re looking at the wrong store or you won’t get enough customers in a five-mile radius, you can go back and change that definition instantly. You don’t have to go back to the very start. And then the next time you need the map, the system won’t ask for the definition or customers all over again. It already has them.
Q: What were some of the technical challenges you had to overcome? A: I’m proudest of the report cue management. If you ask for a bunch of reports, you can come back a few minutes or weeks later and pick them up. If you delete them or lose your own copies, they’ll still be there on our server. With typical desktop systems, you run a report and print it out and it disappears. But ENVISION archives the output so it’s always available. It’s like Fall 2016 ❮ THE ANALYTICS JOURNAL
25
the old mainframe days. You can run eight reports, three maps and an executive summary, and then come back an hour later and they’ll all be waiting for you. ENVISION can geocode 5,000 customers in about five minutes. Now on the Internet, that may seem like a long time, but in the meantime your analysts can continue to be productive while delegating report production to ENVISION. And reports are stored on our server so you can always access them.
Q: How can ENVISION help companies in today’s tough economic times? A: In bad economic times, you need loyal customers who will continue to buy from you. ENVISION helps companies reach those loyal customers using precision marketing that explains who they are, what they think and what they want. And that’s important even during economic downturns because brand awareness is built every day.
Update Now the Senior Vice President for Software Strategy at Environics Analytics, Gary Wood is responsible for the technical direction of EA’s software development and its role in the marketing platform ecosystem. An accomplished software designer and developer, he has over 25 years of experience in geodemography, market analytics and statistical analysis. ENVISION was not the first micromarketing platform he created. A former software developer at Compusearch, he was the chief architect of micromarketing software, including MarketMath and the map-centric application Zephyr. At Compusearch Gary met Jan Kestle, who was in charge of the company’s not-for-profit practice. When Jan started her own company, Environics Analytics, Gary signed on as one of its first 10 employees.
THE ANALYTICS JOURNAL ❯ Fall 2016
2010 Originally published in the March 2010 edition of Direct Marketing magazine.
When unaddressed mail produces upmarket results Holt Renfrew finds surprising success with unaddressed catalogues By Michele Sexsmith
I
n the world of luxury retailing, the personal touch has always been touted as the best way to handle affluent customers. Just ask Holt Renfrew, Canada’s leading luxury retailer. For years, Holts has sent personalized direct mail regularly to some of its best customers, always tailored to their previous purchases and preferences for the finer things. But when Holts wanted to attract new customers for its fine jewellery, it decided to forego the kid-glove treatment and instead turned to an unconventional approach: unaddressed mail to targeted postal walks. And the results proved surprising. Holts has never been shy about trying new approaches to win customers. For 173 years, the retailer has used new and unusual marketing strategies to promote its exclusive fashion apparel, footwear, handbags, cosmetics and jewellery. Every spring and fall, it mails a catalogue with its to-die-for wares to customers who treat the beautiful book like a collector’s item. In November of 2007, Holts launched an outreach effort with its so-called White Catalogue of jewellery items. The company hoped to raise awareness of its jewellery collection, acquire new customers and compare the effectiveness of a catalogue to direct mail for connecting with prospects. To reach the target audience, the company printed 100,000 copies and inserted them in a Saturday edition of The Globe and Mail distributed in Toronto, Montreal and Vancouver. According to Pierre Montagnier, Holts’ Senior Manager of Customer Analytics, the campaign was well received, creating a lot of buzz for the chain’s jewellery department. But the approach had limitations. It didn’t screen out existing customers and the newspaper distribution zones were too large to allow Holts to select preferred areas for targeted marketing. “Geographically inflexible” is how Montagnier describes it. More problematic, the results were hard to measure. “The campaign was successful on the surface,” says Montagnier. “People came into the store talking about the catalogue and Fall 2016 ❮ THE ANALYTICS JOURNAL
27
we had strong aggregate growth in our fine jewellery business. But when we attempted a deeper analysis of this particular marketing tactic, it was hard to measure the results. Our response measurement was fuzzy.” Unable to see a significant difference in sales patterns or customer acquisition between the targeted cities and those that didn’t receive the catalogue, Holts couldn’t determine the program’s effectiveness. For Montagnier, an analytical marketer who holds a Ph.D. in mechanical engineering, this lack of quantification was particularly frustrating. “If marketing isn’t measurable, why bother doing it?” he asks rhetorically from his Toronto office. “We wanted to crunch the numbers to make sure that we’d made the right decision using newspapers. But the numbers just weren’t available.” To design a more measurable campaign, Holts revamped its approach. In 2008, it contacted Environics Analytics (EA), the Toronto-based marketing analytics company, to help identify its best customers and determine where to find more of them. Using PRIZMC2, EA’s segmentation system that classifies Canadians and markets into 66 lifestyle types, analysts first categorized Holts’ existing customers into one of the clusters. Not surprisingly, many hailed from the nation’s most affluent segments, such as Cosmopolitan Elite (very wealthy middle-aged and older families), Suburban Gentry (wealthy, middle-aged suburban families) and Les Chics (sophisticated, urban Québec couples and singles). Then researchers mapped the presence of key clusters to neighbourhoods within the trade areas of Holts’ stores. By knowing the top clusters and customer sales per neighbourhood, marketers could pinpoint the under-penetrated areas—that is, neighbourhoods where like-minded prospects live who were not already regular Holts’ customers. Holts selected the 294 most promising postal walks (consisting of about 300 households each) near its flagship stores in Toronto, Montreal and Vancouver. In the fall of 2008, it mailed 75,000 lavish, art-directed, high-gloss catalogues to unaddressed households in those targeted areas. But it also identified “control” neighbourhoods with a high concentration of the key clusters and held back 16,500 catalogues to test the difference in sales between the two groups. Then it sent 40,000 of the 20-page catalogues to existing customers who’d never bought jewellery in order to compare a direct mail approach to customers with the targeted unaddressed strategy to prospects. Finally, to obtain an accurate budget comparison, Holts made sure that the cost of mailing the catalogue unaddressed to the target postal walks was comparable to the cost of the previous year’s newspaper insertion. “We wanted a scientifically designed test,” says Montagnier, “so when it was time to compute ROI, we’d have hard numbers and not just intangible feedback. By distributing catalogues at the postal walk level, we felt we could better trust the results.” With its careful design, the less personalized campaign yielded a number of surprises. First, the response rate of prospects who received the unaddressed catalogues matched the response of existing customers who were sent direct mail—both in number of purchasers and amount spent by each responder. In addition, residents in the postal walks who received the new catalogues showed a 60 per cent lift in sales over those in areas who didn’t receive a piece. “That was the icing on the cake,” says Montagnier. “The campaign proved that receiving a catalogue made a difference in sales. Even when comparing customers who share the same lifestyle, those who receive a catalogue tend to shop more than those who don’t.” THE ANALYTICS JOURNAL ❯ Fall 2016
28
Beyond the lessons learned, the campaign made money for Holts. Despite the $3.30 cost for each catalogue, the unaddressed campaign not only broke even thanks to the immediate sales to the new customers, it eventually proved profitable. A follow-up analysis showed that a significant percentage of newly acquired customers returned to purchase jewellery within six months—a good indicator of customers with a promising lifetime value. “The campaign wasn’t just a one-time wonder,” says Montagnier. “It really created a larger base of jewellery customers.” The White Catalogue campaign also indicated which PRIZMC2 clusters were the most responsive to the catalogue mailing, and Montagnier’s team hopes to refine their cluster selection for future marketing campaigns. Although he recognizes that Holts’ indirect approach to marketing fine jewellery through unaddressed mail may be unusual, he believes it can be extremely effective if married to well-designed analytics behind the scenes. “There is no magic to this kind of target marketing,” says Montagnier. “You just have to go about your analysis in a very disciplined manner. If you look for your most profitable customers and address them specifically, even in an unaddressed way, you cannot go wrong.”
Update Michele Sexsmith is now the Senior Vice President and Practice Leader for the media, retail, real estate and entertainment industries at Environics Analytics. In that role, she assists clients in leveraging omnichannel opportunities, from identifying the best sites for new stores to developing the optimal media mix for reaching targeted consumers for future growth. Originally planning on a career in experimental psychology—“the analytical framework is very much like direct response modelling,” she says—Michele earned an honour’s bachelor’s degree in psychology from Hamilton College. She held positions in research, marketing and sales at Compusearch, before focusing on the automotive industry in the 1990s; she joined Blackburn Polk Vehicle Information Services and later became Chief Operating Officer of Polk Canada. But when she learned in 2003 her Compusearch colleague Jan Kestle had founded Environics Analytics, she left academia behind and hasn’t looked back.
Fall 2016 ❮ THE ANALYTICS JOURNAL
2011 Originally published in the September 2011 edition of Direct Marketing magazine.
Geodemography Q&A with Jan Kestle
Q: Geodemography has been around for nearly 40 years, how has it managed to survive? A: Geodemography has survived— indeed thrived—for a number of reasons. First, Canada offers a wealth of survey data that geodemography can leverage. Government and private sector organizations conduct reliable surveys on spending, media preference, technology adoption, leisure activities, tourism and many other aspects of day-to-day living. For most surveys, the sample size is sufficient to release data for Canada and the provinces and the larger markets. But these survey-based variables only become usable for trade area analysis and local marketing when they are combined with geodemographic segments using typological inference. Because Canada has more than 1,500 good-quality census variables at the neighbourhood level, analysts can develop a robust segmentation system built on a broad range of reliable and comprehensive data. This allows an analyst to combine the survey-based measure by segment for a behaviour like “go the movies,” for example, with information on which segments live around a location to determine the viability of a new cinema. Without geodemography the amount of local marketing data would be greatly reduced. Another reason geodemography endures is less technical: segmentation clusters are easy to understand, and they capture the essence of a population with fanciful icons and clever nicknames. Nielsen’s Shotguns & Pickups, MapInfo’s Kindergarten Boom and Environics Analytics’ Lunch at Tim’s all conjure up images of groups of consumers much more readily than cumbersome descriptors like “upscale empty nesters in condos.” But perhaps the most important reason is that the results of these systems are executable and measurable. While much market research is descriptive and can help with product conceptualization, brand awareness and advertising, only geodemography can link the customer, product and brand profiles to site selection, local marketing (including direct mail or flyer drops), merchandising, category management and media planning. And because marketers can tie segments back to the ground (stores, postal codes, markets, etc.), to channels THE ANALYTICS JOURNAL ❯ Fall 2016
30
and to their own transactional data, campaigns can be measured and continually modified.
Q: How have geodemographic systems evolved over the years? A: Originally, most systems were built using demographic data to form the customer segments but at Environics Analytics we take some pride in having pushed the envelope by incorporating psychographic Social Values to provide deeper insights into the behaviour and mindset of customers. Our first PRIZM segmentation system, created in 2004, classified Canada’s neighbourhoods into 66 unique lifestyle types using Social Values from our sister company Environics Research. And in 2006, we updated that system with new census data to create PRIZMC2. It captured Canada’s growing diversity, with 15 francophone segments and 13 segments that are home to visible minorities, new immigrants and those who speak a language other than French or English. Because it’s linked to many important marketing surveys and databases, it helps marketers create the most complete picture available of their customer segments. We continue to move geodemography forward with other innovations. This year, we launched two new postal code-level segmentation systems, PRIZM∆ (“PRIZM Delta”) and DELTA. PRIZM∆ is developed at the postal code level using the 66 segments of PRIZMC2. DELTA classifies all postal codes into 121 types—nearly double the number of segments of PRIZMC2—based on key dimensions such as age, income, education, marital status and dwelling type. These systems provide clients with additional ways to differentiate customers for profiling and modeling applications.
Q: What are the main applications for geodemography today? A: Geodemographic systems are used for everything from targeting direct mail and crafting advertising messages, to defining product mix, selecting store sites and planning media. Because the segments can be linked to customer data and a number of media and marketing surveys, decision-makers get data-driven analysis on what consumers are buying, doing and thinking. Marketers can identify their most profitable segments and locate areas where they can find similar consumers.
Q: Is geodemography used in one particular sector more than others? A: That’s the great thing about geodemography’s long history—after proving its value in consumer marketing, it’s been adopted by a variety of industries and not-for-profits. Banks and insurance companies use geodemographic segmentation to develop products, services and messages that increase client retention through cross-selling and up-selling. Retailers can identify underserved markets and areas where operations should be combined or curtailed. Fundraisers can focus on potential donors that are likely to have the highest response to their direct marketing campaigns. Even governments use geodemography to ensure that the right services are available in the appropriate areas. Our system has helped the City of Mississauga better market its libraries and parks programs to its residents.
Q: Privacy is a primary concern for direct marketers and consumers. Is geodemography good or bad for those particularly sensitive to privacy concerns? A: It’s definitely a good tool because geodemographic systems don’t capture individualized Fall 2016 ❮ THE ANALYTICS JOURNAL
31
data. They classify people by common interests, targeting homes one block at a time and avoiding potential privacy concerns. Geodemographic overlays are a privacy-compliant way to enrich transactional databases. Many analysts use the clusters as well as individual variables that are proprietary to their businesses in custom models. Viewed another way, geodemography also helps businesses and not-for-profits who do not have unit record customer data. While large retailers, financial institutions and charities typically keep track of their customers and donors, many businesses still do not collect much personal demographic information on their customers. Geodemographic segmentation systems allow those companies to develop a profile of their customers using something as basic and non-intrusive as postal codes, which have long been used to determine how far a customer is willing to travel to a particular outlet.
Q: How is geodemography moving into the digital age? A: In a word: enthusiastically. Direct marketers today can target their customers with geodemographic segments as part of their interactive campaigns. By knowing the online, mobile and social media habits of their customers, marketers can create highly targeted and interactive applications such as coupons, contests, promotions and loyalty programs delivered over mobile phones, through social networks and via standard websites. Last year, we partnered with 80/20 Solutions to help clients determine which PRIZMC2 segments of their customers are more likely to respond to digital programs. They can then customize offers and creative based on the imaging we’ve done for their target groups. And through 80/20’s marketing platform, clients can conduct contests and email marketing campaigns in real time, gauge campaign ROI and goal achievement, and then look at media effectiveness— comparing one geodemographic target group to the other. And because it’s all measurable, clients can do a segmentation analysis to get a good sense of who’s engaging and how the campaign should be modified for better results. At our User Conference last fall, we demonstrated targeting mobile messaging using PRIZMC2 with another partner, Broadplay. Attendees sent a short code on their smart phones followed by a postal code and each instantly received an offer and coupon tailored to their needs. In the digital world, good targeting is even more important than in the off-line world as consumers are happy to receive messages on their devices that understand who they are. And we’re making sure that PRIZMC2 stays on top of the digital habits of consumers. This year, we partnered with Delvinia to connect its AskingCanadians mobile and social media surveys to PRIZMC2. This enhancement enables marketers to target their products, messages and channels based on their customers’ mobile and social media habits and lifestyles. For example, with this information clients can identify influential shoppers who frequently rate products online and have more than 500 friends on their social networks. They can determine which immigrant groups use social media to stay connected with families back home. And they can even pinpoint which geodemographic segments of Canadians like to tweet, retweet and read Twitter postings. For the first time, marketers can target their products and messages based on the online, mobile and social media habits of all segments of Canadian society. Jan Kestle is the President and CEO of Environics Analytics. THE ANALYTICS JOURNAL ❯ Fall 2016
2012 Originally published in the October 2012 edition of Direct Marketing magazine.
Direct mail devotees in the digital age Just because they’ve gone digital doesn’t mean consumers don’t like direct mail By Jan Kestle and Michael J. Weiss
T
he media landscape may be more fragmented than ever with emerging digital channels like mobile apps, tweets, Facebook likes, email ads and online banners. Yet through it all, the power of direct mail remains undiminished. While some believe that flyers, catalogues and addressed mail are “so last century,” a new study by the global interactive marketing company ExactTarget confirms that nearly two-thirds of consumers have made a purchase as a result of a direct mail piece. According to a recent channel preference survey of Canadian and U.S. consumers, nearly half of respondents say they prefer direct mail to email. In today’s era of always-on channels like email, SMS and social networks, consumers say they appreciate direct mail’s substance, flexibility and once-a-day tempo. Just who are the folks who invariably open a direct mail envelope—whether it’s a sales pitch, donation request or informative brochure? According to Environics Analytics (EA) and the Print Measurement Bureau (PMB), nearly 13% of Canadians “always” or “often” open and read direct mail letters. These ardent envelope openers are a different breed than the average recipient of unsolicited mail. But that’s not to say they’re all alike. Canada’s direct mail devotees fall into roughly three target groups. In the city, you’ve got your young singles and families who like getting announcements about the latest brands and boutiques. Out in the suburbs, a number of upscale families look for envelopes that hold promotions from stores where they already do business. And you’ve got your older, settled residents of towns who’ve been responding to fundraising appeals for years. As a group, direct mail fans are concentrated geographically in cities and small towns along Canada’s southern tier, with particularly high percentages, for example, in Montreal, Quebec, Petawawa, Ont., and Dawson Creek, B.C. Demographically, direct mail fans sit squarely in the centre of the Canadian mainstream: They mostly mirror the general population in terms of age, education, marital status and family Fall 2016 ❮ THE ANALYTICS JOURNAL
33
type. True, they include a higher percentage of immigrants, university graduates and those who work at white-collar jobs. But their incomes are close to the Canadian average—$92,322 annually—and they’re about evenly divided between homeowners and renters. Despite this middle-of-the-road demographic portrait, their lifestyles are all over the consumer map. PRIZMC2, EA’s segmentation system that classifies Canadians into 66 segments, shows that the top-ranked segments for direct mail enthusiasts include Daytrippers & Nightowls (young, mobile urban singles and couples), La Cité Nomade (downscale, young and mature Quebec singles), Asian Affluence (wealthy, suburban Chinese families) and New Homesteaders (rural, midscale couples and families). But they share some patterns of behaviour that may help marketers better target their products and messages to these consumers. Research shows that these are active households into playing a number of sports—soccer, basketball and tennis—as well as attending pro basketball, football and baseball games. But direct mail fans also have a cultured streak, as seen in their higher-than-average attendance at operas, art galleries and museums. Their media preferences follow the general population, with average rates for watching TV, listening to the radio and reading newspapers and magazines. They show a slight preference for news radio, TV sports and publications that cover business, finance, travel and entertainment. Among their favourite titles are Starweek and Canadian Business and preferred TV networks include MTV Canada and the Food Network. And reflecting the many citydwellers among their numbers, they also respond to outdoor advertising in buses, streetcars and subways. Given their interest in analog communications, perhaps the most striking media trait shared by direct mail fans is their fondness for the Internet. They’re heavy online consumers with strong rates for using email, instant messaging, Where direct mail is always read blogs, podcasts and social networks. They use their Province/Territory Index mobile phones for web browsing, purchasing products, entering contests and downloading Quebec 110 coupons. They like to visit websites that feature Prince Edward Island 105 beauty and fashion, home improvement, Saskatchewan 101 investments and newspapers and magazines. And Alberta 99 they’ll go online to buy books, home electronics, toys and games, computer hardware and software British Columbia 98 and travel-related purchases. Nova Scotia 98 For businesses and not-for-profits wanting to New Brunswick 97 make direct-mail-friendly consumers part of their Ontario 96 marketing mix, research by EA and PMB shows that they’re particularly responsive to retailers who sell designer clothes, home electronics and vacation travel packages. And they like messages that recognize they’re stressed for time and promote convenience as part of their shopping experience—online or in bricks-and-mortar stores. They look for products that embody THE ANALYTICS JOURNAL ❯ Fall 2016
Yukon Territory
96
Newfoundland and Labrador
94
Northwest Territories
58
Nunavut
18
Index of 100 = national average Sources: Environics Analytics, PMB, 2012
34
Direct mail readers like:
Direct mail tossers prefer:
Basketball
Hunting
originality, physical beauty and technology. And despite their solid, Art galleries Bingo halls midscale incomes, they Yoga/pilates Snowmobiling like to save money and Attending food shows Country music concerts appreciate direct mail pieces with money-saving Collecting stamps Bird watching offers. Direct marketers MTV Canada Mystery also should be aware “The Bachelor/Bachelorette” TV figure skating that these shoppers Canadian Business Outdoor Canada are also influential Following brands on Twitter Use Google social media consumers. As selfdescribed “consumption Beauty/fashion sites Video games evangelists,” they like Sources: Environics Analytics, PMB, 2012 spreading the word about brands and products they prefer. “I am always one of the first of my friends to try a new product,” they tell researchers. Given this multi-layered portrait of direct mail fans, even the most dedicated digital marketers would do well to recognize the importance of integrating mail as part of a multichannel strategy. Despite the hyperactive interactivity of Twitter, Pinterest and Groupon, many Canadian consumers still like to sit down and mull over a tangible, real-world letter in a way that they never can with websites. And they’re more than likely to respond if that envelope is filled with a money-saving promotion for a unique outfit, vacation package or tablet computer. While direct mail may not be able to match the lower cost of online advertising, a well-targeted, perfectly positioned pitch can have a better chance of connecting with the right consumer right where they live—away from the clamour of the Internet.
Update President and CEO of Environics Analytics, Jan Kestle has been a leader in the marketing information industry for more than 40 years. Prior to founding EA in 2003, Jan was president of Compusearch and spent 19 years at the Ontario Statistical Centre. An expert in using statistics and mathematics to help solve business challenges, she has helped hundreds of customers turn data and analytics into insight, strategy and engagement. Michael J. Weiss, Marketing Consultant at EA, is an internationally known expert on geodemography and has had a hand in developing nearly 20 lifestyle-based segmentation systems in the U.S. and Canada, including EA’s PRIZM5, PRIZM5 QC, PRIZM C2 and PRIZM CE. An award-winning writer, he is the author of three books on consumer segmentation: The Clustering of America, Latitudes & Attitudes and The Clustered World.
Fall 2016 ❮ THE ANALYTICS JOURNAL
2014 Originally published in the February/March 2014 edition of Direct Marketing magazine.
On becoming an analytics organization By Catherine Pearson
T
wo years ago, as many companies began overhauling their market strategies following the recession, Alterna Savings faced some familiar challenges: tough competition, complicated decision making, and imperatives to grow its customer base. Alterna’s response, however, was anything but conventional. Rather than deal with the problems on a piecemeal, departmental basis, Alterna revamped its operations to become an analytics-based organization sharing a 360-degree view of its customers across the enterprise. Of course, no company can overhaul its operations overnight—especially one that’s been in business since 1908. As Ontario’s second-largest credit union, Alterna Savings currently serves more than 100,000 members as a financial cooperative. With 24 locations, the member-owned institution offers a full range of financial products and services, all with a commitment to supporting the well-being of members, employees and communities. But to Pamela Stewart, Alterna’s Vice President and Chief Marketing Officer, the September 2011 decision to change the way it did business was “the beginning of a journey.” At the time, Alterna segmented its members by financial holdings and made marketing decisions “based on intuition from our old legacy records,” says Stewart. “We couldn’t tell you who our highest potential members were. And our marketing team was concerned that we weren’t talking to the right people.” In today’s era of Big Data, the credit union didn’t lack for customer information in its member database. Its big challenge, however, was identifying and extracting the key data that would provide the insights necessary to help them better connect with members and prospects. “We were segmenting on dollar holdings only,” recalls Stewart. “It was one dimensional and we learned that we were missing opportunities to connect with our membership.” Specifically, Alterna wanted to know how to augment the credit union’s own rich data
THE ANALYTICS JOURNAL ❯ Fall 2016
36
with external databases and tools to build customized member segments that would help them acquire new members and better engage existing ones. Just as important, they were interested in spreading their customer-centric analytics across the organization. “We wanted to show how everything works together,” says Stewart. “We needed to bring together IT, product, marketing, retail and our wealth partners to help us identify new business opportunities across our lines of business. So it was critical to see what our customers look like.” Starting with the question “Who are our best customers?” Alterna’s marketing team gathered members of other departments—including senior executives, IT and retail experts, and product managers—to participate in the customer-centric initiatives. The company also engaged Environics Analytics (EA) to guide them on the journey to becoming an analytics-based organization. With the first order of business to understand its in-house data, EA analysts undertook a large-scale segmentation of its members using EA datasets like PRIZMC2 (lifestyles), WealthScapes (wealth) and DemoStats (demographics). Analysts then turned to Alterna data to identify the top-performing member segments based on member behaviour such as their level of product engagement and length of membership. The dominant PRIZM segments that emerged from the member analysis were combined into distinct target groups, which were detailed in summaries that provided comprehensive views of members’ lifestyles and habits. And the results helped Alterna officials better understand their customers, calculate wallet share and determine opportunities for growth. With the imaging profiles, all departments now had a common understanding of Alterna’s members and how to communicate with them. By classifying members into target groups, the analysis showed that Alterna’s best customers in Toronto—such as members of Winner’s Circle (well-off, middle-aged exurban families) and Pets & PCs (large, upscale suburban families)—are different from those in Ottawa—which included members of Money & Brains (upscale, educated couples and families) and Suburban Rows (younger, thriving immigrant families). A gap analysis allowed analysts to calculate market and wallet share, and it indicated where branches had an opportunity to increase sales of financial products like GICs and mortgages—and where they didn’t. It also helped Alterna develop a “win-back strategy” to reengage lapsed members. “Branch managers appreciated the imaging profiles,” says Stewart. “They could finally visualize who they were talking to. And marketers could use the profiles for unaddressed mail campaigns. They helped us develop the wording and images for the messaging and creative.” The analysis also helped Alterna reduce attrition. EA Analyst Mark Harrison developed a retention model that identified early warning signs of members contemplating moving their business elsewhere—key triggers to reengage those members. As Stewart puts it, “if we can understand the triggers, we can get proactive to try and keep them as a customer.” With its members now at the centre of its business strategy, the credit union has applied data-driven insights across the organization—from executives to front-line staff—to strengthen its relationship with current members and attract new ones. Social Values data have given Alterna branch managers insight into how to communicate with members during phone calls using language that resonates with them. The PRIZM-linked media and Fall 2016 ❮ THE ANALYTICS JOURNAL
37
lifestyle data helped marketers develop unaddressed admail to high-potential markets and target groups. “For the first time, we had a common understanding of our members and a starting point for developing common metrics to measure the outcome of our strategies,” observes Stewart. “There’s been a change in thinking through our tactics.” Although Alterna’s new customer-centric approach is still in its early stages, branch managers offer anecdotal evidence that their new strategy is working. “We know we’re talking to the right people now,” says Stewart. “We’re no longer talking to the same people on the same lists. We’re seeing an increase in sales. There is positive momentum in branches.” By linking member behaviour to PRIZM, Alterna’s team is now able to better leverage the role of analytics in marketing and other activities—and link everything together. Indeed, the effort to take Alterna to an analytics-based organization has been so successful, says Stewart, that credit union executives have no interest in seeing the initiative come to an end. “The work left Alterna with a thirst for more quantitative insights to define opportunities and increase engagement,” says Stewart. “We saw the work as a journey to becoming an analytics organization, and the results are worth the time and effort.” For Alterna, when it comes to connecting with customers—both current and future—the journey never really ends.
Update Catherine Pearson is now Senior Vice President and Practice Leader for the finance, insurance, travel and telecommunications industries with Environics Analytics. With nearly 25 years of sales and account management experience, she provides companies with extensive consulting and applications support to implement their marketing programs and custom modelling solutions. She is one of the driving forces behind the development and popularity of WealthScapes, EA’s financial database that offers 178 key financial and investment statistics and serves as a balance sheet of Canadians’ assets and liabilities. Originally designed in 2007 for bankers and financial planners, it has become popular among retailers, university fundraisers, real estate developers and university professionals—for direct mail, mass media and out-of-home buys.
THE ANALYTICS JOURNAL ❯ Fall 2016
2016 Published in the May 2016 edition of Direct Marketing magazine.
Big Data analytics: Isn’t it still just analytics? Tactics must evolve but the mission hasn’t changed By Richard Boire
A
s “Big Data” continues to dominate discussions in the analytics space, along comes the notion of “Big Data Analytics” to add confusion in the marketplace. If Big Data analytics warrants its own discipline, then its methodologies and approaches should be significantly different from what has been used in traditional analytics. On closer examination, I would argue that the analytics at the heart of Big Data analytics remain fundamentally the same, but require even greater focus on the core business problem to be solved. Big Data has always been with us, it just wasn’t discussed as widely as it is today. The traditional users of Big Data were direct marketing firms and credit card companies but the growth of digital technology and new devices have altered the paradigm so that many organizations now have easy access to large volumes of information. Technologies like Hadoop have facilitated the processing and consumption of ever-increasing volumes of data. Indeed, data scientists—or “data miners” in the last-century vernacular—have always contended with volume. And they have always earned their salaries through their ability to transform raw source data into meaningful insights. In any exercise, creating a meaningful analytical file is still the most important first step but now data scientists must also be able to both identify the business problem and create a data environment that provides the information foundation to develop a business solution.
The new reality Before the digital explosion of the Internet and social media, a typical project would involve the data miner asking for as much data as possible. The rationale was to allow the data miner to filter out all the noise in the data which represented structured data. But in our Big Data world, massive volumes of semi-structured and unstructured data no longer lend themselves to this approach. The initial “ask” of the data needs to be filtered. Fall 2016 ❮ THE ANALYTICS JOURNAL
39
Historically, raw data consisted of transaction records, customer files, campaign data and, perhaps, geodemographic data. All this data was structured but the information was meaningless in its raw state. Data miners had to “work the data,” applying an extensive variable derivation to process it all into meaningful variables or fields. It wasn’t unusual for this type of data transformation process to generate several hundred variables. By contrast, in much of today’s exploding digital environment, the data arrive either in semi-structured or unstructured format. The newer challenge for data scientists is to first convert this raw data into meaningful variables. Extraction tools now allow the data scientist to identify key fields and information without knowing the data structure or location of the information. The use of NOSQL databases and programming languages such as Python, R and Java provide one approach to transforming semi-structured and unstructured data into some meaningful format. But this extraction is meaningless unless a further transformation occurs. Data scientists need to remember the business problem they are trying to solve. For example, if I am trying to understand how engagement with Coca-Cola in social media has changed both prior to and after a marketing promotion, I might do the following: 1. Extract all tweets with keywords related to Coca-Cola that occurred two months prior to the promotion date and two months after the promotion date. 2. Convert that data to JSON objects and extract the date field using Java type programming or some API. 3. Create an analytical file of a structured table with only one date field. 4. Create a graphical trend report using a tool such as Tableau that depicts tweet counts— prior to and after the promotion. Further, if I want to learn whether a tweet refers to Coca-Cola in a positive or negative manner, I could turn to sentiment analysis tools and create a graphical trend report—again using a tool like Tableau—to graph the different sentiments over time. But this general reporting of tweet behavior over a period of time is insufficient to effectively determine how a promotion has altered social media engagement. The extraction process needs to be much more focused in order to address the specific business question. Especially when it comes to social media, the old “give me everything” approach simply consumes too many resources in the attempt to make sense of the data. Identifying and understanding a business problem traditionally is one of the four key steps in the data mining process, but it is even more critical when dealing with social media data today.
Content adds context Besides identifying simple engagement and sentiment, the analysis should also probe more deeply into the content. Are certain themes or topics emerging in the social media conversations? The use of text mining and text analytics tools allow this type of more exhaustive probing. But again, what is the business problem we are trying to solve? If the challenge is creating more customer engagement, text mining may reveal that certain themes or topics are more relevant in driving this engagement to higher levels as a result of the marketing campaign. Clearly, the business problem must dictate how data scientists use social media. Suppose THE ANALYTICS JOURNAL ❯ Fall 2016
40
we want to build a customer retention model that uses social media such as tweets to determine customer satisfaction. The first issue concerns the ability to match customer records from the company’s database against the individuals who are engaging in social media. The second issue is one of reliability: some current research questions whether the comments of people on social media truly represent the opinions of the “silent majority.” Furthermore, there may be privacy issues raised in using this type of information. If our intention is to build better retention models, we might seriously question the usefulness of appending social media to customer records given these issues. Big Data and especially social media data will continue to grow. As analytics practitioners, we can no longer respond by “extracting everything.” Today more than ever, we truly need to understand the business problem so that we can effectively extract the right information when building the solution. While we continue to see great developments in software and technology, the real challenge for analytics and data science is human related: having the right analysts who are trained and educated on the principles of data mining as well as business analysis. This ability to understand the domain knowledge of a given business, grasp its major issues and dissect its challenges will become even more paramount in any data scientist’s skill set. In that sense, Big Data analytics may differ from traditional analytics, but I regard it more as an enlargement of the discipline that makes data scientists even more valuable to the 21st century organization. Richard Boire, senior vice president, Boire Filler, at Environics Analytics, is the author of Data Mining for Managers: How to Use Data (Big and Small) to Solve Business Challenges (2014: Palgrave MacMillan).
Update In March, 2016 Environics Analytics acquired the Boire Filler Group, a Toronto-based marketing company that specializes in predictive analytics, data management, business intelligence and customer-centric services. Boire Filler’s offerings are now integrated into EA’s products and services suite, providing clients of both companies an expanded range of data and analytics services, including data cleansing, integration, modelling and management. Since its founding in 1999 by industry veterans Richard Boire and Larry Filler, the Boire Filler Group has been assisting clients—such as Hudson Bay Company, Nestle, CAA and Whirlpool—by providing predictive modelling and data management support for marketing and communications strategies. Boire, with a 30-year career in data mining and analytics, is a nationally recognized expert in the database and data analytics industry. Filler has over 25 years of experience in relationship and database marketing, most recently focusing on transforming data into insights that can be leveraged to drive more effective CRM results. Together, they have assisted clients representing a range of industry sectors, including technology, financial services, retail, automotive, packaged goods, loyalty, travel and not-for-profit.
Fall 2016 ❮ THE ANALYTICS JOURNAL
2016 This article is a new contribution written for this Journal and previously unpublished.
The value matrix
By Stephen Shaw
There are many ways to segment customers based on value. But to know who is truly valuable, data analysis must be combined with market segmentation and loyalty research. This triangulated view will not only inspire creative thinking—it can even be used to predict future behaviour.
I
t was a product so effective that its users swore by it. But the manufacturer was worried. The patent license was about to expire. A flood of generic products was expected to swarm the market. The customers were loyal—exactly how loyal was about to be tested. How many of their long-time users would switch to the cheaper generic brands? What would be the likely revenue exposure? What could be done to convince the most valuable customers to stay? In the absence of a crystal ball, the answers would have to come from some mix of research and data analysis. The majority of end users were small independent businesses, most of them operating out of a single location. With such a large and geographically dispersed market to serve, the company’s sales force was forced to concentrate on the more lucrative multi-location businesses. That left marketing with the job of protecting the rest of the portfolio. But of the many possible countervailing strategies—price discounts, loyalty offers, a generic product of their own—which option made the most sense? One advantage the company enjoyed was a large transaction database linked to individual businesses enrolled in a purchase rebate program. Using this database, the company could stratify firms by past spending, helping them identify the most valuable customers. But the company would have to stretch that analysis to answer a host of other questions: Did customers vary significantly in needs? Did they differ in their decision-making process? What factors made them more or less brand loyal? And importantly, how much cheaper would a competitive product have to be before a customer considered switching? THE ANALYTICS JOURNAL ❯ Fall 2016
42
If the company answered those questions, they could start to craft a defense strategy. They could think about how to extend the value of the product beyond its functional benefits. They could introduce new value added services that would make the brand even more integral to the success of their customers. They could reposition the brand to stand for something more than simply product superiority. Or they could aggressively expand the product portfolio to cushion the inevitable fall in market share. All of those options were on the table, especially given how much revenue was at stake. But what was the best way to get the answers they needed to defend their flagship product?
True insight If the primary goal was to minimize flight risk, the company would have to focus its retention strategy on the best customers outside of the key accounts. These were firms whose business could not be allowed to slip away. Since true insight lies at the intersection of behaviour, motivation and intent, the analysis would have to examine the past purchase patterns of these customers, and then probe for the attitudinal factors that best explained their behaviour. Knowing why customers acted in a certain way—such as sharp jumps or declines in spending volume, a noticeable deviation in order frequency, or a longer-than-usual interval between purchases—might inspire a creative response to the looming crisis. A systematic approach was needed, combining behavioural profiling, attitudinal segmentation and loyalty driver research. The analytical database would have to take the shape of a three-dimensional matrix: one side denoting customer value; another side representing attitudinal drivers; and the third side reflecting the state of the brand relationship. That way the matrix could be pivoted to identify nested segments within each value tier. More than that, the analysis would be able to estimate the potential portfolio loss due to customer attrition, helping to build the business case for retention. Before any of this work could proceed, however, it was essential to do the proper groundwork. The company needed to understand how their customers—80% of whom were small- to medium-sized businesses—preferred to operate: how they made decisions; what their ideals and aspirations were; their management acumen; their appetite for growth; how they treated employees; their business confidence and outlook; their views on supplier relationships, and more. All of that knowledge could be used to shape the segmentation study to follow. To gather this feedback a series of one-on-one discovery interviews were conducted with select customers who were representative of the market. What the company learned was that these business owners and managers varied dramatically in entrepreneurial drive and business sophistication—not surprising, given their humble origins. However, they did share a lot in common: concern about the stagnant economy; about government regulation; about the risk of litigation; about the health, safety and training of their employees; and, of course, about keeping their customers happy. A collective picture emerged of business owners passionate about their work; heavily reliant on distributors for advice; and loyal to products, not brands. Most worrisome: none of them were averse to trying generic products if it made financial sense. As far as they were concerned, the product manufacturers were all the same. None of them had any meaningful presence in their lives—in fact, as corporate brands, they were very nearly invisible: the Fall 2016 ❮ THE ANALYTICS JOURNAL
43
brand relationship did not extend beyond the box.
Strategic riddles No matter how conversant everyone is with the Pareto principle, it always comes as a shock to owners when they see how dependent their business is on so few customers. In this case, the top 20% of the customer portfolio accounted for 83% of total revenue (no surprise there). But a top decile account was worth almost seven times more than the average customer—a worrisomely high multiple. And on closer inspection almost half of the revenue from the top two deciles was being generated by just two per cent of customers. Many of those elite accounts, of course, were already being served directly by the sales force. Yet mixed in with those key accounts were many less important customers, diffusing the sales effort. Across all of the value segments there were sub-groups of customers whose year-overyear spending growth far exceeded the rest of the portfolio. In fact, about one-fifth of the customer base had doubled their spending in the past five years. But almost two-thirds of customers had reduced their purchases, largely due to adverse market conditions. An intriguing story began to emerge: within each value tier there were significant numbers of customers whose historical spending trajectory made them highly desirable accounts to retain. Therefore, the value segmentation had to incorporate two other critical variables: relationship status (entrenched, ambivalent or inert) and spending velocity (fast-growth, moderate, flat or negative). Together these variables could be used to identify high value customers whose behaviour was suggestive of business vitality and continued growth. What the behavioural segmentation could not reveal was how these customers differed in their attitudes and needs from everyone else. At a time when the market was shrinking due to declining demand, how did the more successful businesses manage to thrive? And what was stopping them from growing even faster? The answer to those questions could only be revealed through a segmentation study of the entire market based primarily on two factors: growth orientation and business sophistication. Once the market segments were defined they could be mapped back to the customer base to determine the precise distribution across the value tiers. Within this value matrix, custom strategies could be devised for each segment based on their respective business drivers and management maturity. This approach would help solve a number of strategic riddles: How should they extend the value proposition to be seen as more than just another product manufacturer? What would they have to do to earn true partnership status? And what really mattered to these customers beyond the functional product benefits?
Beyond the box The segmentation study found that the eligible universe of companies was bifurcated: half of the target population was made up of firms in survival mode, struggling with profitability and growth, while the other half consisted of businesses which were flourishing despite the decline in overall market demand. Within the no-to-low growth market, there were two kinds of business owners: those who recognized their weaknesses and aspired to reach the next level of maturity—and a self-reliant but cynical group of established businesses wedded to the status quo. On the high-growth side, operators could be characterized by their level of THE ANALYTICS JOURNAL � Fall 2016
44
adroitness in business planning, or by their zeal for top-line sales growth. All four segments were distinct in their operating style, in their openness to new ideas and in their business savvy. At that point it was obvious which market segments were the most attractive— but the harder question was, which of them was least brand loyal? When the company mapped these attitudinal segments to their own customer base, they found that the A behavioral percentage distribution mirrored the segmentation of the customer base market. But when revenue contribution integrates lifecycle was factored in, one segment stood stage, current value and historical buying out from all the others: the business patterns and trends. planners who kept careful control of their bottom line. In fact, that one segment accounted for half of all product sales. Clearly, these customers were critical to retain. There was just one problem: they were the most value conscious—which made them the most likely to switch if the price was right. In fact, it was estimated that more than 80% of that segment would be at risk of defecting if the price variance was greater than 20%. The implications were apocalyptic: potential revenue losses could run as high as 60%. What could be done to dissuade these customers from bolting at the first sign of a better deal? Before proceeding with any kind of strategy, the company needed to reimagine the brand experience through the eyes of customers. To help with that exercise, personas were developed for each segment, bringing to life their hopes, dreams and ambitions, while describing their defining attitudes. The pragmatic and sensible business planners, for example, were personified as an archetype by the name of Mike. His main goals: to balance growth and cost management. He is depicted as a “calculated risk taker” always on the lookout for better and more cost-effective products. Thorough and rational in his decision style, Mike was the sort of manager who revels in intelligence gathering and trend watching— someone who would appreciate a manufacturer playing a more active role in the industry and generously providing post-sale advice and support. With a clearer idea of what this valuable customer segment wanted, the brand manufacturer could confidently march forward with a defined game plan: justify the premium pricing of its flagship product by offering exclusive informational services which could not be easily mimicked by low-cost generic competitors. The company would not have to yield ground on price—nor would it be forced to offer loyalty inducements: it would just have to hold out a helping hand to customers seeking value beyond the box. Stephen Shaw is the Chief Strategy Officer of Kenna, a marketing solutions provider specializing in customer experience management. He can be reached via e-mail at sshaw@kenna.ca Fall 2016 ❮ THE ANALYTICS JOURNAL
2016 Originally published in 2016 in Geodemographics for Marketers: Using Location Analysis for Research and Marketing by Barry Leventhal (Kogan Page Limited).
Why geodemographics is relevant in an age of Big Data By Richard Webber
T
he idea of geodemographic classification was introduced to the marketing industry in the late 1970s. By comparison with other marketing innovations of that time its use has been remarkably resilient, notwithstanding the changes in the structure of the marketing industry, new media and the plethora of new sources of data that have become available, particularly in the Internet era. The purpose of this section is to explain the reason for this resilience. During the late 1970s, when geodemographic classification gained its first adherents, it overcame two serious problems that had been troubling the industry. The first benefit was that it could provide information on what sorts of people were buyers of very particular products and services—and it could do this quickly, consistently, cheaply and without the need to add a question to a consumer survey and wait for a response. The second was that it enabled direct marketers to target new customer recruitment campaigns using direct mail or door-to-door distribution with the level of precision they have become accustomed to when using print media, radio and television. Since the mid-1970s the mix of applications that rely on systems such as Mosaic and Acorn has evolved in response to changes in the focus of marketing practices, new communications channels and new information management technologies. Prior to the 1970s the focus of many large companies had shifted from mass marketing to what was described at the time as ‘target marketing’. This followed naturally from the declining share of commercial television in the marketing mix, and improvements in the ability of computers to store, access and manipulate customer information. As computers became able to store and process ever larger amounts of data, large organizations began to realize that the information they had used to manage customer accounts could beneficially be linked together in relational form so as to produce a more comprehensive and detailed picture of the purchasing history of each customer. Marketers
THE ANALYTICS JOURNAL ❯ Fall 2016
46
in large organizations began to realize that this information could provide an invaluable resource for identifying existing customers to whom other products in the company’s product portfolio could be cross-sold. There are two reasons why, contrary to the prognostications of some industry observers, customer data did not cause a decline in the use of geodemographic data. The first of these is that, whilst customer relationship management systems provided an excellent resource for cross-selling and up-selling in sectors where customers transacted frequently but at low value (such as in banking or telecom), there remained many other sectors such as automotive and large household items, where the tendency of customers to make infrequent but high-value purchases limited the volume of transactional information held about them. Second, in markets such as credit cards, where a consumer is likely to hold an account with multiple providers at any one time, previous transactional information is necessarily restricted to that part of the customer’s wallet that he or she awards to any one supplier. In the absence of geodemographic information it is nigh impossible for the marketer to distinguish on the one hand the customer who is faithful to a single supplier but makes little use of the product category and, on the other, the customer who is an extensive user of the product category but who undertakes the majority of his or her transactions with a competitor. In this situation, whilst a customer relationship management (CRM) system is capable of describing the profile of an existing customer’s spend, it is not able to place this in the context of the opportunity that the customer represents for additional business. A geodemographic classification is in a much better position to help gauge the size of the opportunity. However, direct marketing has never been the sole use to which marketers have put geodemographic classification. The late 1970s, when Acorn was first introduced, was a period when many retail multiples were expanding their presence on Britain’s high streets. In those days, when a multiple retailer operated a limited set of formats, there would be a requirement to establish how well the geodemographics of a potential new store catchment area matched the profile of what constituted the profile of the profitable customer. As logistics improved, multiple retailers became able to support a much wider variety of retail formats, indeed even to apply the concept of mass customization to the process of deciding what to stock on a store-by-store basis. Where geodemographics had been used primarily to decide where to open a new store, increasingly it became used to decide what product lines should be stocked in the store and how much shelf space should be allocated to different product categories. From the beginning, geodemographic classifications were used to exclude certain types of neighbourhood from particular product marketing campaigns. What the more perceptive marketers began to realize was something less immediately obvious but potentially more powerful. That is, that the types of people with whom a consumer comes into contact on a daily basis have a significant impact on the decisions that the consumer makes, whether consciously through emulation—keeping up with the Joneses —or unconsciously, due to people’s innate tendencies to adopt the prevailing opinions of the social groups to which they belong. Geographers refer to these influences as neighbourhood effects. Essentially their argument is that although two postmen may have identical demographic profiles, they Fall 2016 ❮ THE ANALYTICS JOURNAL
47
are likely to espouse radically different social attitudes and to have very distinct consumer preferences because of the different profiles of the people they encounter at work and in their own neighbourhood. The influence of where a person lives on their attitudes and behaviour was identified very early on in the commercial use of Acorn when Ken Baker was able to demonstrate how, despite having very similar readerships in terms of social grade, The Telegraph and The Guardian appealed to people who lived in significantly different types of residential neighbourhood. The Telegraph was more likely to be read in neighbourhoods of uniformly white, middle-class and property-owning residents. The Guardian’s readers were more likely to be found in middle-class enclaves in ethnically diverse inner-metropolitan locations, many of whose residents worked in communications, entertainment or the public sector and where the proportions of women with a university degree, and especially in the arts, was particularly high. Nor should we overlook how much consumption patterns can be affected by access to and availability of products on a local basis. Once retailers learn to locate and merchandise their stores using geodemographic classifications, residents can find it increasingly difficult to purchase products other than those for which there is a strong demand from their neighbours. In more recent years the advent of digital communications channels has hugely increased the amount of information that can be derived about an individual and that can be used to sharpen the targeting of communications. There are four principal respects in which information tends to be more specific: • Information that is known about an individual tends to be far more timely. Whereas in the 1980s the decision as to whom to target with a promotion might be based on spend in the last few months, today the decision may be based on activity in the past few minutes. • Likewise, whereas in the 1980s a company might have targeted a ‘good’ customer on the basis of total quarterly spend, or even spend on apparel, today it is normal to target customers who have bought specific items. • Prior to the digital era and the availability of Big Data, data-driven marketing could only be based on previous purchase history and response/non-response to specific mailings. Today, segmentation can be based on what products consumers have looked at on the web and how long they have stayed on any particular web page. • Finally, segmentation is becoming more specific in relation to customers’ locations. It is now possible to target consumers who are known to have passed a particular location or even to target them when they are approaching particular locations. Fabulous though these opportunities are, the more specific our targeting capabilities the easier it is to lose sight of the contextual data that can help to explain both the social influences to which consumers are subjected and their motivations and aspirations. If the highly specific and the contextual can be kept in balance, then it is much easier to identify the correct medium of communication as well as the message. In my opinion, a compelling example of the importance of contextual information is the 2014 Clacton by-election campaign, which resulted in the return of the first UK Independence Party (UKI P) member to Parliament. The dominant geodemographic segment in the Clacton constituency is labelled “bungalow retirement.” This demographic is dominated by people THE ANALYTICS JOURNAL ❯ Fall 2016
48
formerly employed in mid-level white-collar jobs, many of whom used to live in 1930s suburbs of large cities. On retirement, many of them have capitalized on the increased value of their houses and retired to modest seaside properties, typically in a street of uniform properties built by speculative developers for people on the cusp of retirement. The attraction of these streets to ex-urban migrants is that they will be surrounded by people of similar attitudes to themselves—and perhaps similar to the people who were their neighbours when they first bought their urban properties 30 years previously, people who are likely to deplore the increase in social and ethnic diversity of what had once been staunchly conservative suburbs. These are likely to be people who share a sense of responsibility for their own welfare, who are reliable neighbours and regular in their habits. They tend to value personal contact when they visit neighbourhood shops, they do not hanker after foreign holidays, find it difficult to adapt to new technology, and above-average proportions tend not to have frequent contact with their grown-up children. Whilst Big Data can clearly differentiate one household from another within this stereotype, being able to describe groups of individuals—many of whom share a broadly similar life history—would have enabled UKIP to identify Clacton as an easy seat to win. This would have told UKIP which streets in the constituency they should focus their communications on, what messages would resonate the most, and which postcodes in the constituency fitted this geodemographic segment. It is this contextual capability that, in my opinion, explains why geodemographics continues to be relevant to consumer marketers in spite of the vast array of digital data to which they now have access.
Update Richard Webber is one of the founders of geodemography. As an executive at CCN Marketing/Experian and CACI in the UK, he led the development of both MOSAIC and ACORN neighborhood classification systems, the two most popular geodemographic segmentation systems in the UK. At one time, he oversaw a sprawling complex of MOSAIC cluster systems that analyzed consumers in 19 countries (including Environics Analytics’ PRIZM5). By linking all the countries into a single segmentation system called Global MOSAIC, he could arguably claim the title of the global village’s top marketer: Webber’s analysis synthesized 631 different MOSAIC types and 800 million people in the various countries into 14 dominant lifestyles. He is currently a Visiting Professor at the University of Newcastle and founder of OriginsInfo, which pioneered an ethnic name-coding system that predicts the cultural, ethnic and linguistic origins of people based on first and last name alone. Through a partnership with OriginsInfo, EA offers OriginsCanada, which helps marketers enhance customer profiles and tailor their messages, media and products to Canada’s diverse populations. Richard’s long and successful career epitomizes industry best practices at work on a global scale.
Fall 2016 ❮ THE ANALYTICS JOURNAL