University of Wisconsin-Milwaukee
Employment and Training Institute

.

Confronting Anti-Urban Marketing Stereotypes: A Milwaukee Economic Development Challenge

by Lois M. Quinn and John Pawasarat, Employment and Training Institute, School of Continuing Education, University of Wisconsin-Milwaukee, June 2001.

Working with the Helen Bader Foundation and the Milwaukee Department of City Development, the Employment and Training Institute has developed purchasing power profiles for Milwaukee neighborhoods and commercial districts. The profiles are posted on the City of Milwaukee website at www.mkedcd.org/PurchasingPower. These profiles utilize current state income tax records, local property files, business phone listings, state driver's license files, motor vehicle records, and state and federal mortgage lending databases to identify the purchasing power that comes from urban population densities and concentrations of income and to describe economic trends of interest to retailers and other commercial businesses. As part of the project, ETI examined the information reports and stereotypes about Milwaukee neighborhoods presented by national and international marketing firms. This paper examines the methodologies firms use to develop marketing information on Milwaukee neighborhoods, traces the history of marketing clustering systems, and explores the damage marketing stereotypes pose for Milwaukee and other cities. See also, the Brookings Institution discussion paper on "Exposing Urban Legends: The Real Purchasing Power of Central City Neighborhoods."

Findings

  1. Marketing firms frequently use racial and class-based stereotypes to describe urban neighborhoods. On its website CACI, an international data firm, stated that African Americas in Milwaukee "splurge on fast food and spend leisure time going to bars and dancing." Milwaukee Hispanics, CACI stated, "splurge on videos, long-distance phone calls, cable TV, and theme parks and casino visits." Claritas, Inc., another international data company, reported on its website that a fourth of residents on the near southside of Milwaukee (in zipcode 53204) "smoke 9 or more cigarette packs per week ... [and] don't know the amount of money needed to retire comfortably." Claritas described the residents of northside African American neighborhoods in Milwaukee as "very low income families [who] buy video games, dine at fast food chicken restaurants, [and] use non-prescription cough syrup." Meanwhile, according to Claritas, North Shore suburbanites are "interested in civic activities, volunteer work, contributions and travel." A review of zipcode profiles for other cities uncovered identical descriptions used to characterize hundreds of urban neighborhoods around the country.
  2. In order to produce data reports at low expense, most marketing firms rely primarily upon decennial census data and broad assumptions about "types" of households. Ignoring state and local databases, CACI denigrated four Milwaukee zipcode areas (53205, 53206, 53208, and 53212) as "distressed neighborhoods" where people "who work have only part-time jobs." Yet, working age tax filers in these zipcodes reported adjusted gross income totaling $804 million in 1999. Working age filers reported $37 million income per square mile in zipcode 53205, $67 million per square mile in zipcode 53206, $75 million income per square mile in zipcode 53208, and $55 million per square mile in zipcode 53212. Oak Creek residents, whom CACI described on its website as "a prosperous population who have opted for semirural locales and lifestyles," had just $20 million adjusted gross income per square mile in 1999.

Total 1999 Adjusted Gross Income Per Square Mile:
Working   Age Tax Filers

  1. When retail spending patterns are estimated based on detailed income tax data, the purchasing power of the central city is very high -- given the high concentrations of workers and data showing that lower-income households spend a greater portion of their income on retail expenditures than higher-income households. The data show that Milwaukee residents in the four zipcodes described as "distressed" by CACI spend an estimated $373 million annually for retail purchases for food, apparel, entertainment, household furnishings and equipment, and personal care products and services. These central city residents are spending $22 to $38 million per square mile on retail expenditures -- well above the estimated $4.5 million spent in Franklin and $17.4 million spent in Greendale, for example.

Estimated Annual Retail Spending Per Square Mile

  1. Rather than comparing purchasing power per square mile, marketing companies usually rank neighborhoods based primarily on race and class. Sparsely populated suburban areas with high average household income are rated as "winners" while densely populated urban areas with higher concentrations of income are ranked as "losers." These rankings are then used to steer businesses toward upper income neighborhoods while writing off urban centers. In a research report for the National Retail Hardware Association, Claritas encouraged home improvement store owners throughout the country to target suburban neighborhoods while ignoring urban centers. ETI estimates that eight central city Milwaukee neighborhoods in the categories written off by Claritas spend over $93 million annually for household furnishings and equipment. These "non-targeted" central city neighborhoods spend an estimated $3.7 million per square mile each year for household furnishings and equipment while seven Milwaukee County suburban areas with Claritas categories identified as important markets for store expansions spend one-third that -- $1.2 million per square mile.

Est. Spending for Household Furnishings & Equipment

  1. In order to produce reports that can be easily and inexpensively replicated for every possible neighborhood in the United States, local statistics are often ignored even when they directly measure the trend under analysis and when the data are readily available. CAP Index, Inc., a leading seller of crime data, estimates "crime risk" based on a statistical model that links national crime statistics to decennial census data on the types of people living in neighborhoods where crime often occurs. The company reported that the "risk" of murder, rape or robbery at Miller Park is 3 to 4 times the national average -- based on its statistical analysis of census data for people living in a 12-square mile area around the ballpark. Under this formula, every baseball stadium in the country is considered an unsafe area, except the ballpark in Arlington, Texas (home of the Texas Rangers). CAP Index, Inc., rates the University of Wisconsin-Milwaukee campus as having "risk" of violent crime at five to nine times the national average, not because of reported crime around UWM, but because there are census tracts with large numbers of poor people living within five miles of the school.
  2. To minimize costs and maximize production of data, national marketing firms even "adjust" local data to fit their national statistical models and computer programming needs. For example, Claritas, Inc., secured data from the Milwaukee Department of City Development showing a 3 percent decline in housing units in the census tracts around zipcode 53204, yet reported a 9 percent decline. The firm reported that the area had 36 percent fewer owner-occupied housing units than the City of Milwaukee property file showed, 29 percent fewer vehicles than the number registered with the state Department of Transportation, and 16 percent fewer working families than the number of working age single and married income tax filers with dependents. Using its statistical model (and ignoring local school, health department and tax revenue records suggesting growth), Claritas projected a 10 percent decline in population over the decade while the population actually showed a 5 percent increase.

  3. Most marketing companies do not reveal the sources of their data or when the data were last updated. Because commercial marketing information is readily available (although often at exorbitant prices), some government departments and community agencies have also relied on such data, hampering their own ability to sell their neighborhoods.

Computer-Generated Stereotypes Hinder Economic Development

Early ETI reviews of commercial marketing firm reports focused on two neighborhoods - -Mitchell Street on the southside and Fond du Lac and North avenues on the northside. At the request of the Helen Bader Foundation, the Employment and Training Institute prepared a purchasing power profile for the neighborhood around Fond du Lac and North for a national site selection team from Kmart. The site is at the center of one of Milwaukee's most densely populated neighborhoods, with a growing labor force, rising incomes and available workers. A suitable 14.7-acre tract of vacant land was available on two heavily-traveled arterials and served by three bus routes. The site had much to recommend it -- except for its perceived reputation as an inner city area. An urban neighborhood like this, which could potentially benefit from the 300 jobs and modestly-priced retail goods a Super Kmart might bring, is plagued by deficit-based marketing stereotypes and urban legends that lead companies to avoid the central city.

Consider the description of this neighborhood offered by CACI Marketing Systems, an international firm headquartered in London, England. CACI described the neighborhood as "distressed" and "characterized by youth, single parenthood, poverty, high unemployment, and public assistance." Its residents, CACI announced on its website, "splurge on fast food and spend leisure time going to bars and dancing." "Unemployment is high," CACI reported, "those who do work have only part-time jobs."(1) A similarly negative portrait was provided by Claritas, Inc., a subsidiary of VNU Marketing Information Services, a Netherlands- based media conglomerate. Claritas labeled the neighborhood "Difficult Times" where residents "buy video games, dine at fast food chicken restaurants, [and] use non-prescription cough syrup." Another Claritas marketing scheme labeled the area as a combination of "Inner Cities" (a type it ranked 47th in desirability), "Single City Blues" (ranked 45th) and "Mid-City Mix" (ranked 30th).(2)

If the Kmart team were to consider locating in the nearby suburb of Oak Creek, they would learn from CACI Marketing Systems that the residents there are "a prosperous population who have opted for semirural locales and lifestyles." As for leisure time activities, CACI reported: "As homebodies, they are interested in reading, needlework, and cooking." Claritas offered a similarly positive description of Oak Creek: supposedly 33.81% of residents are a "Family Ties" population who "eat at fast food pizza restaurants, shop at Builder's Square, and own a bowling ball," while 33.77% fall in the "Great Beginnings" category, meaning residents who "dine at fast food Asian and Mexican restaurants, shop at 7-Eleven, and lease their car," and 17.82% are in the "White Picket Fence" grouping, where residents "play softball [and] eat at Dairy Queen."

It would appear from the Internet that a host of international marketing firms have traveled to Milwaukee's central city and surveyed neighborhood residents, but no such visits have occurred. A review of neighborhoods around the country from Robert Taylor Homes in Chicago to Scottsdale, Arizona --uncovered identical descriptions (and dozens like them) used to characterize hundreds of urban neighborhoods around the country. When the methodology used to create these "market segments" was examined in detail, it was discovered that most of the marketing statistics and stereotypes are based on ten-year-old census data and broad assumptions about purchasing habits of minorities and urban residents. Average household income was used as the primary factor in rating neighborhoods while income density was ignored.

For example, 1999 state income tax returns showed that the residents in the Fond du Lac and North area (zipcode 53206) had 10,166 working age single and married tax filers, or 3,737 filers per square mile. This is one of the highest concentrations of working age income tax filers in the Milwaukee area. By contrast, Oak Creek had 405 tax filers per square mile. Working age families and single persons in zipcode 53206 reported $183 million in total adjusted gross income in 1999, or $67.3 million per square mile, a concentration of income more than three times that earned by "prosperous" Oak Creek residents -- $20 million. When the 1999 adjusted gross income of Wisconsin tax filers was examined for the four zipcode areas that CACI Marketing Systems describes as "distressed" where employment is supposedly only part-time, the income totaled $804 million. Income earned in 1999 per square mile in these neighborhoods ranged from $36.9 to $75.1 million well above that of Oak Creek and Franklin and comparable to other suburban neighborhoods.

Research by the Bureau of Labor Statistics Consumer Expenditure Survey (CEX) finds that lower-income families spend far more of their income on retail purchases than higher income earners. Using detailed income tax data from the Wisconsin Department of Revenue by neighborhood and block area, ETI estimated the purchasing power of zipcodes and blocks in Milwaukee County based on patterns of spending found by the CEX for large cities in the Midwest by income range and family composition. The model for this work was developed by Frank Stetzer and John Pawasarat of the University of Wisconsin-Milwaukee after a review of five quarters of CEX data, state income tax data and HCFA (Health Care Financing Act) files on the elderly population. The data showed that residents in zipcodes 53205, 53206, 53208 and 53212 annually spend over $373 million for retail purchases. These central city residents are spending about $29 million per square mile on retail expenditures.

When purchasing power per square mile is examined for the Milwaukee area, it becomes immediately apparent that central city neighborhoods show high concentrations of annual income and offer important economic opportunities for retailers. These opportunities have yet to be realized by many national marketing firms. By ignoring the income density of the nation's cities and focusing their neighborhood rankings on average household income, marketing firms mislead retailers and other businesses into thinking that central city neighborhoods are a "distressed" market where business is unprofitable. A case in point is the work performed by Claritas, Inc. for the National Retail Hardware Association described below.

Steering Business From the Central City

In the last three decades a host of national and international marketing companies have created neighborhood descriptions based on cluster analysis of census tracts, zipcodes, block groups and ZIP+4 areas. Typically, the population of the entire United States is segmented into 40-60 groupings (see detailed discussion below) which are then grouped and/or ranked according to their market potential. For example, Claritas, Inc. used its 50 "MicroVision" clusters (which included the "Difficult Times," "Great Beginnings," "Family Ties," and "White Picket Fences" descriptions cited above) to advise the National Retail Hardware Association on where its members should locate and expand their businesses. Based on the Claritas research, home improvement center owners throughout the country are encouraged to target suburban neighborhoods while ignoring urban centers. Fourteen market segments were identified by Claritas as "suburban" and all fourteen were recommended as good business prospects. Meanwhile, only 2 of the 18 market segments identified as "urban" or "urban/suburban" were recommended as neighborhoods to target for home improvement retail business.

The Claritas rationale for its recommendations was explained in a marketing article on the NRHA website: "The reason for identifying these marketing targets is based on two tried and true observations. First, the best customers are existing customers. The people who are already shopping these stores or other people similar to existing customers are most likely the best customers. Secondly, birds of a feather flock together. In choosing a place to live, people tend to seek out neighborhoods compatible with their lifestyles. ... By identifying the types of neighborhoods in which they find existing independent stores' shoppers, researchers can accurately predict the types of neighborhoods where the industry can find future customers."(3)

Again, the data on Milwaukee neighborhoods call these marketing assumptions into question. The estimated purchasing power for the CEX category of household furnishings and equipment was calculated for each Milwaukee neighborhood based on 1999 income tax data and other current files on the population. In the central city Milwaukee neighborhoods given category labels dismissed by Claritas as low-priority markets for home improvement stores, spending for household furnishings and equipment ranged from $1.6 million (in zipcode 53233) to $5.6 million per square mile (in zipcode 53210). In suburban neighborhoods with types recommended by Claritas as prime customer areas for home improvement sales, spending for these categories ranged from $0.6 million per square mile in Franklin to $3 million per square mile in South Milwaukee.

"Discrimination in the Digital Age"

Milwaukee is not alone in feeling the effects of marketing stereotypes that inaccurately describe urban communities. In June 2000, the Acorn Housing Corporation announced a lawsuit against the Wells Fargo Bank for linking Internet users exploring housing moves to websites with racial stereotypes designed by CACI. The housing group objected to the description of inner city and African American neighborhoods as "distressed" and to such quotes as, "They splurge on fast-food and chicken restaurant take-out. They listen to urban contemporary music and read Playboy magazine." According to Mike Shay, executive director of the advocacy agency, the Wells Fargo website "brought the old discriminatory practices of block-busting and racial steering into the digital age." Shay explained, "If you were moving from one city to another they'd ask what zip code you currently lived in, and then they'd refer you only to similar zip codes with similar racial characteristics in the new city."(4) In an immediate response, Wells Fargo dropped the website link to the CACI data. The lawsuit is ongoing.

In Richmond, Virginia, a jury returned a $100.5 million verdict against Nationwide Mutual Insurance Company, a decision based in part on the insurance company's use of marketing strategies to rank each zipcode in the metropolitan area according to its Claritas "MicroVision" category. Agents were encouraged to target zipcodes with more desirable "MicroVision" rankings. All of the City of Richmond zipcodes which were predominantly minority were excluded from the list of targeted areas. In a divided opinion, the decision was overturned on appeal on the grounds that the fair housing organization bringing the suit, Housing Opportunities Made Equal of Richmond, lacked standing under Virginia law.(5)

The two marketing firms whose clustering plans were the basis for the housing lawsuits (Claritas and CACI) now team up with non-profit organizations to sell special "inner city" databases and analyses to promote urban communities; at the same time these firms' websites and client reports continue to denigrate central city neighborhoods. The housing lawsuits may represent the tip of the iceberg as cities identify the negative impacts of marketing stereotypes commonly provided to Fortune 500 companies, local entrepreneurs and Internet surfers by international marketing companies. For cities like Milwaukee, the negative marketing stereotypes associated with city neighborhoods amount to computer-generated "urban legends" that have to be overcome in order to attract new businesses. The scientific nature of the statistics presented lead credibility to the reports, even when these statistics are contradicted by local, more accurate data, as seen in the example below.

A "Weighted Statistical Average": Crime Risk, Crime Reality

The most popular crime indices use a demographic-based model which estimates crime risk based on the kinds of people who reside in neighborhoods where crime is often committed rather than using current data on where crimes actually occur. Data companies use national statistics from the Federal Bureau of Investigation, the decennial census population reports and other sources, in order to make broad generalizations about the type of people who are more likely to live in areas with certain types of crimes. Once this crime "risk index" is developed for the country, the crimes reported nationally (or in some cases by metropolitan area) are then "reallocated" to each neighborhood based on the characteristics of its residents. New crime indices are issued each year even though the demographic characteristics of the neighborhood population may be ten or more years old.

Using this methodology, the APB News website reports that the risk of homicide, rape or robbery around Milwaukee County Stadium (now Miller Park) is 3 to 4 times the national average. In this case, the crime risk is based on the CAP Index statistical model applied to the 1990 demographic characteristics of residents living within a 12-square mile circle around the stadium. Given the methodology used, it is not surprising that twenty-seven major league baseball stadiums in the U.S. are designated as in high crime risk areas. According to the CAP Index, only the Texas Rangers play in a "safe" area Arlington, Texas. Indeed, almost any place urban will show up as dangerous, irrespective of the actual crimes that occur against college students or baseball fans.(6)

The CAP Index, the leading provider of these risk numbers, goes so far as to claim, "Police Reports and the FBI Uniform Crime Report (UCR) are useful compliments to CAP Index scores, but are not thorough enough by themselves."(7) Local residents are admonished, "If your risk surprises you, keep in mind this is a weighted statistical average for the entire ZIP code."(8) Samples of crime risk maps displayed on the Internet typically portray hundreds of square miles of city neighborhoods as dangerous places. Parents logging onto the Internet are told, for example, that the University of Wisconsin-Milwaukee is in a "high risk" crime area. The CAP Index rates the UWM campus as having "risk" of violent crime at five to nine times the national average, not because of reported crime around the campus but because there are census tracts with large numbers of poor people living within five miles of the school.(9)

The level of neighborhood specificity helps give these crime indices their believability --to parents debating whether to allow their high school graduate to attend an urban university or to a business executive considering locating a retail operation in a highly populated neighborhood. Yet they are far from accurate. To test the accuracy of crime information offered by marketers for Milwaukee neighborhoods, the Employment and Training Institute reviewed Milwaukee police department reports and selected 22 census tracts with the same number of violent crimes (murders, rapes and robberies) in 1998. The CAP Index ranked six of these tracts as "low risk" for violent crimes (i.e., less than one-half of the national average), seven tracts as "moderate risk" (at the national average), eight tracts as "moderately high risk" (two to four times the national average) and one tract as "high risk" (five to nine times the national average). Thus, the maps of these neighborhoods on the APB News website shaded some census tracts in red for "dangerous," some in yellow for "caution," and some in green for "safe." The maps suggested the risk of violent crime was up to 18 times higher in some tracts than in others. In fact, the true crime level, as measured by official police statistics on violent crimes committed, was the same in all 22 census tracts.

Selling crime vulnerability statistics (and giving away even worse data) has become a thriving Internet activity particularly on websites advertising security systems, car locks and homeowners insurance. The impact of these "statistics" in scaring away potential businesses, homeowners and other development is impossible to estimate but it could hardly be a positive one. Other data about urban areas are equally suspect, as seen in the case study below.

"Complex Methodology": Portraying Stable Neighborhoods As in Decline

While preparing a purchasing power profile report for zipcode 53204 on the near southside, ETI compared the data it had amassed for the Milwaukee Neighborhood Indicators Project with reports offered by commercial firms. The demand for annually updated, geographically detailed demographic estimates by neighborhood has resulted in the proliferation of marketing databases. Nearly all marketing firms construct their current estimates based on 1990 demographic data from the decennial census updated from an impressive array of public and private data sources. These sources are usually fed into a formula to estimate the current and projected demographic and purchasing patterns but in some cases these databases have serious limitations that impact negatively on measures of neighborhood well-being.

An exhaustive examination of data sources revealed a number of commercially developed sources that were in serious conflict with local institutional sources of data for zipcode 53204. One very popular commercial source of recent, detailed demographic data was Claritas, which for $99 offered a 12-page demographic report with 1999 data and projections for 2004 as well. The Claritas data, however, portrayed 53204 as a neighborhood with declining population, home ownership, housing units, vehicles and working families. For example, while Claritas secured tract level data from the Milwaukee Department of City Development, which showed a 3 percent decline in housing units from 1990 to 1999, the firm "adjusted" the data to report a 9 percent decline in housing units -- triple the city counted loss. The firm also reported that the area had 36 percent fewer owner-occupied housing units than the City of Milwaukee property file showed.

The Wisconsin Department of Transportation file showed 14,135 registered vehicles but Claritas reported 10,141, or 29 percent fewer vehicles. The state's motor vehicle file showed that 10,303 households had valid license plates for their vehicles, but Claritas estimated that 6,998 households had vehicles, or 32 percent fewer. A total of 6,756 working age single and married tax filers with dependents submitted income tax returns to the state Department of Revenue in 1998, but Claritas reported only 5,703 working families, or 16 percent fewer.

Claritas reported a 10.1 percent decline in population, from 40,530 in the 1990 Census to 36,430 in 1999, and projected a further decline to 34,643 by 2004. CACI Marketing Systems estimated an even larger decrease of 10.9 percent over the nine-year period, with the 1999 population estimated at 36,102. Yet, almost all sources of data assembled by ETI showed a stable neighborhood with indicators pointing in a positive direction over time. Births had remained stable, with 1,026 in 1993 and 1,031 in 1998. The school census showed a 4 percent increase in the school age population. State income tax filings showed a 6.3 percent increase in working age tax filers from 1993 to 1998 and a 22 percent increase in adjusted gross income. When questioned about the serious discrepancies in their population data for 53204, researchers at Claritas responded that they had rechecked their sources and methodology. They explained that they utilize "a complex methodology that uses post-1990 inputs (Donnelley) to create accurate block group estimates." They stated further that "Every source of data that we have shows this ZIP code as declining."(10) The 2000 Census data confirmed the observations of residents and local data findings: the 2000 Census population count totaled 42,747, a 5.5 percent increase over 1990.

Puzzled why a marketing firm would alter current administrative databases (like the city property file and the state department of transportation motor vehicle files), ETI began examining the statistical models used to drive the generation of city statistics by the marketing firms. Ensuing discussions with Claritas revealed that many of the sources used by Claritas/Donnelley were federal, state and local databases commonly accessed by marketing firms: city housing units at the tract level, state department of transportation vehicle files, and files from the United State Postal Service, U.S. Census, Internal Revenue Service, and the Health Care Financing Administration. ETI purchased or obtained these data files and then compared them to the Claritas estimates. The institutional data sources provided by federal, state and city agencies were compared to the results of the "complex methodology" used in the 1999 demographic and housing statistics for 53204. In reviewing methodologies used by Claritas and other marketing firms to adjust local and state administrative databases with zipcode data, Employment and Training Institute researchers identified several critical areas where the population and economic assets of urban neighborhoods are often undercounted.

A primary factor in explaining the discrepancies in data produced by marketing firms is their heavy reliance placed on historically under counted U.S. census data. Many firms compound urban undercount problems by utilizing other databases which are often less accurate in urban areas. For example, monthly statistics on active U.S. Postal Service delivery route addresses and postal boxes are often used to "update" census counts, but postal workers warn that these counts are not intended for this purpose and tend to be less accurate in city neighborhoods with high residential mobility and several families sharing the same postal address. The Health Care Financing Administration (HCFA) count of the Medicare population is another valuable national database, but it must be used with caution in neighborhoods with large numbers of immigrants and other elderly persons not eligible for Medicare. Also, telephone counts are frequently used as population trend factors, yet a lower percentage of low-income households have listed phones and many may use cell phones if they have poor credit history. Likewise, driver's license files, while valuable data sources, should be used with caution in states like Wisconsin where license suspensions are used as a method of collecting unpaid fines and civil forfeitures unrelated to driving. Other limitations of generating neighborhood reports by computer (one researcher called it "as easy as printing money") become evident once the history of cluster marketing is understood.

History of Cluster Marketing: "You Are Where You Live"

The methodology used to warn retailers, college applicants and even baseball fans away from central city neighborhoods had its ironic genesis in research funded by the U.S. Office on Economic Opportunity, as part of the War on Poverty. In 1966 the OEO funded Jonathan Robbin, a developer of multivariate computer software for the IBM 360 and former sociology professor at New York University, to prepare descriptions of the demographic and economic characteristics of U.S. counties utilizing 1960 census files and other federal statistical data. Robbin developed indices to compare areas of need across the country including a "Poverty Index" and an "Index of Susceptibility to Civil Disorder" that was used by the U.S. Department of Housing and Urban Development to select cities for "Model Cities" anti-riot funds.(11)

Robbin saw the potential for analyses of decennial census data at the zipcode level to help identify desirable markets for mass mailings and advertising firms. Robbin's oft-stated premise for the neighborhood clusters he created was simple "Birds of a feather flock together." Michael Weiss, who has written three books promoting the Claritas clustering model, quoted Robbin's philosophy: "Basically, people seek out neighborhoods that are most congenial to them. At each stage in their life cycle, people tend to join their peers in appropriate communities, whether it's in high-rise city apartments or single homes out in suburbia."(12) The 36,000 U.S. Postal Service zipcodes, initiated in 1963, proved an ideal geographic area for Robbin's clusters. By linking his cluster groupings with consumer files and client mailing addresses, Robbin could identify new zipcode markets for products and target consumers for direct mail and retailing efforts.

Robbin founded a marketing firm, Claritas, to sell his databases to business and government clients. Initially, Claritas created 40 PRIZM (Potential Ratings in ZIP Markets) clusters using 1970 U.S. Census data. This approach provided a model for numerous competitive systems that have evolved since 1974. Once the clusters were created, each zipcode, census tract and census block group was assigned to one of these categories. The clusters were ranked from most to least prosperous, and given catchy titles to convey their market potential or lack of it. The result was to effectively divide up the nation's total market into clusters of winners and losers.

After the 1980 census, Claritas' cluster groupings were modified slightly, with some renamed and others added or subtracted. The market winners in the 1980s groupings began with "Blue Blood Estates" ("America's wealthiest neighborhoods suburban homes and one in ten millionaires"), "Money & Brains" ("posh big-city enclaves of townhouses, condos and apartments"), "Furs & Station Wagons" ("new money in metropolitan bedroom suburbs"), and "Urban Gold Coast" ("upscale urban high-rise districts"). The last of the forty clusters were the losers: "Share Croppers" ("primarily southern hamlets devoted to farming and light industry"), "Downtown Dixie Style" ("aging, predominantly black neighborhoods, typically in southern cities"), "Hispanic Mix" ("America's Hispanic barrios"), "Tobacco Road" ("predominantly black farm communities throughout the South"), "Hard Scrabble" ("the nation's poorest rural settlements"), and the neighborhoods ranked at the very bottom, "Public Assistance" ("America's inner-city ghettos").(13)

While Robbin spoke of reviewing over six hundred variables and millions of market records, three variables appear to be critical to his model: PLACE (big city, suburban community, small town, or rural), RACE (white, black, Hispanic, older white immigrant or young recent immigrant), and INCOME.(14) Claritas now offers 62 PRIZM rankings based on groupings which include some age and family type variables derived from the 1990 census in addition to place, race and income to help identify popular markets for retirees, empty nesters and young professionals. The cluster stereotypes are also now applied to the twenty-two million ZIP+4 delivery areas (i.e., smaller geographic units that typically include several blocks, a group of streets, an apartment building or small geographic areas within each zipcode). The basic division of the American marketplace into the most attractive and least attractive residents is retained: "Blue Blood Estates" remains at the top of the cluster rankings; the three bottom groups are "Hispanic Mix" ("urban Hispanic singles and families") "Inner Cities" ("inner-city solo-parent families") and "Southside City" ("African American service workers").(15)

After Robbin had divided the nation's zipcodes into PRIZM clusters, he began collecting customer databases and client address lists to determine which clusters had households more apt to buy specific types of products and services. Thus, the residents of "Inner Cities," according to a Claritas website, are most likely to "buy baby food, buy soul/r&b/black music, pay bills by phone, watch pay-per-view sports, read the National Enquirer." By contrast, "Old Yankee Rows" residents ("Empty-Nest, Middle-Class Families") are most likely to "belong to a union, believe ad campaigns, buy pop music, watch bowling, and read Spin."

Even in their crude state, these marketing generalizations proved to have utility for Claritas's customers. As Weiss pointed out, a direct mailing campaign would be considered phenomenally successful if it raised the customer response rate from two percent to three percent, even though 97 percent of those targeted did not respond.(16) One Claritas website concedes that PRIZM "is not intended to predict with absolute certainty how every neighborhood's household lives or what they buy or don't buy, rather it is used to predict the likelihood that a certain type of behavior or purchasing preference drink diet soda or go skiing, for instance would be found in higher concentrations within that neighborhood versus another."(17)

Other marketing firms copied the Claritas approach. The CACI Marketing Systems produced 42 unique neighborhood clusters from "Top One Percent" to "Distressed Neighborhoods." Internet surfers are told that "Top One Percent" residents typically "drive luxury cars and visit museums" and are interested in "civic activities, volunteer work, contributions, and travel" while "Urban Hispanics" a group that is far down on the list of clusters "read religious books" in their leisure time.(18)

The Claritas "MicroVision" scheme presents a softer image with more subtle negative stereotypes. The "Close Knit Families" webpage (MicroVision Segment 41) shows a smiling Hispanic family; the population is described as "low income young adults, age 18-34, with a large number of children, seven or more people." They "dine at midscale Mexican restaurants, attend movies 2-3 times a month, purchase doll toys, and had transmission service done in the last year." An African American family with three children is pictured for "City Ties" (MicroVision Segment 24). They are described as "medium-income families, age 22-29, with a relatively large number of children." They "eat at fast food chicken restaurants, smoke menthol cigarettes, and shop at Lady Foot Locker." A visual review of the photographs for each of the "MicroVision" rankings shows minorities pictured in eight of the ten worst clusters and in only one of the ten best clusters.(19)

The market segmentation approach has become so easily and widely duplicated that Global MOSAIC by Experian (the United Kingdom's largest crediting and marketing information firm) now purports to cluster much of the entire world -- over 779 million consumers in 18 countries. Global MOSAIC divides the world population into 14 lifestyle types that it says "can be found in every modernized country based on a simple proposition that the world's cities share common patterns of residential segregation." All 18 nations, according Experian, "have their ghettos of 'Inner City Melting Pot,' suburbs of 'Midscale Metro Office Workers' and communities of 'Educated Cosmopolitans.'"(20) Ethnic stereotypes are common within these classifications. "Chicano Blues" residents are "very reflective of the Hispanic culture (listen to Spanish radio, buy Latin music, travel to Mexico and Central and South America)....[and] rank first for visiting gambling casinos, playing the lottery, and eating fast food 14 or more times in the last month." "Motowners" are "predominantly young, Black (77%) renters with low incomes. They rank second for single parent families and total unemployment. Notable purchases include foods that are directed at children such as powdered soft drinks, bubble gum, Hostess Twinkies, and Kellogg's Froot Loops."(21)

National advertisers were among the first users of the cluster markets running ads in "zoned" editions of news and entertainment magazines to target zipcode areas most likely to buy their products. Governments followed suit. In 1980 the Defense Department contracted with Claritas for help in locating areas where it might find better success in recruiting for the all-volunteer army. Claritas matched zipcode data on three years of Army recruits with their home address's cluster classification and determined that lower-income Southern and rural areas (Shotguns & Pickups, Mines & Mills, Blue-Collar Nursery, Coalburg & Corntown) were top recruitment areas rather than the urban neighborhoods previously targeted. The Army increased the number of recruiting centers in these cluster areas, set new enlistment quotas based on the findings, and increased advertising in selected magazines (Field & Stream, Sports Afield, Outdoor Life, Mechanics Illustrated) commonly sold to residents in the clusters identified. Claritas also analyzed the zipcodes of Americans killed in highway accidents for the U.S. Department of Transportation to determine where the department should concentrate its "buckle up" campaigns (Norma Rae-ville and Grain Belt).(22)

"You are where you live" the Claritas motto takes on self-fulfilling prophecy as institutions target customers based on cluster descriptions. According to Weiss's examples, a West Coast bank used clusters to decide whether to offer area residents a charge card with $500 or $1,000 credit or no card at all. An East Coast developer decided where to locate shopping malls and which stores to seek as anchors. A college in Nebraska sent different recruitment materials inquiring high school students depending upon their zipcode address. Students from higher- ranking marketing clusters were mailed a brochure on "A Stepping Stone to Your Future" while students from lower-ranking clusters were sent a brochure on "An Affordable Education." A college in Maryland assigned roommates based on their zipcode cluster "to reduce conflicts."(23)

Competing Images of City Life

"Tell me someone's zipcode, and I can predict what they eat, drink, drive -- even think," promised Jonathan Robbin, creator of the first Claritas clustering system.(24) As observed above, marketing companies use computer programs to crunch decennial census data looking for pockets of homogeneity, the leitmotif of the advertising world, and then to label neighborhood residents according to pre-packaged cluster stereotypes. While this approach may work fairly well in some relatively homogeneous rural and suburban areas, city populations tend to be far more diverse, with city blocks including people of many backgrounds and lifestyles. Consider this description of the area along Lincoln Avenue on Milwaukee's near southside, described by Milwaukee Magazine senior editor Kurt Chandler: "From third-generation Europeans to first-generation Mexicans to Puerto Ricans, African Americans, American Indians, Hmong, professionals, blue-collar workers, gays and straights there has to be more diversity along this street than any other in the city. On the street, a woman in full leathers rides a Harley down the center line, while a wiry Asian boy kicks a rubber ball along the sidewalk. Lincoln Village is self-sustaining in many ways, a foreign land a city within a city with a different set of cultures and traditions, different faces, too rarely known by outsiders."(25)

Log onto Internet sites of international marketing companies and a very different image of this area is given. CACI Marketing Systems website defines the population as "Low Income: Young & Old." This consumer type, says CACI, "incorporates the extremes of the age distribution, the very young and the elderly, which is reflected in their lifestyles: purchases of infant toys, toy cars and trucks, jewelry, and household items such as flatware and washing machines."(24) The Claritas "MicroVision" model reports that the zipcode area has three major populations. Exactly "28.80%" of the households are "Close-Knit Families," defined as "low income young adults, age 18-34 with a large number of children, seven or more people." Another "25.36%" are labeled "Trying Metro Times," described as "very low income young single adults with young children, age 25-34, and seniors, over 75 years, one to six people." The third group, making up precisely "17.89%" of the population, is catalogued as "Metro Singles" who are "medium-low income young singles and some couples, age 25-39, with few children, one to two people." On its website entitled "A Zip Code Can Make Your Company Lots of Money!" Claritas claims that each of these market segments "consists of households that have similar demographic characteristics, lifestyles, interests, media habits, and purchasing behaviors."(27) Others find city neighborhoods a bit more complex.

Conclusion

National and international marketing firms currently exert great influence on business decisions in the United States. The largest firms amass incredible amounts of information on individuals and localities from public databases (like driver's license and motor vehicle files, postal service route information, federal Medicare files and federal census files) and from private sources (like loan application data, credit card receipts, medical histories and warranty card returns). These databases are fed into national modeling programs used to generate reports for clients in any locality or subpopulation market. Given the mass production nature of the industry, local residents are not consulted on whether the resulting statistics are accurate for a given neighborhood and data inconsistencies are often adjusted to fit the model used. As Jonathan Robbin, the founder of Claritas, envisioned, the data sets are packaged as "expert systems" where the user is not given information on the research methodology, the data sources used or the age of the data sources.(28) Error rates for the demographic estimates are not shown, if ever calculated.

The experience in Milwaukee neighborhoods shows that these data reports should be used with caution both for the accuracy of the information reported and for the stereotypes perpetuated. Increasingly, local agencies will need to "fact check" the profiles being created to describe their neighborhoods and challenge the assumptions on which they are based.

The typical market research firm uses a deficit-based model of city neighborhoods overlooking the income density attributes that could make them attractive to retailers. Retailers who base their decisions on these deficit models will continue to lose out on valuable economic opportunities. At the same time, central city residents will be forced to drive longer distances for products they need and miss opportunities for employment in their neighborhoods. Hopefully, as Milwaukee takes leadership in developing its own marketing data to map neighborhood assets, the negative influences of international marketing firms can be diminished and the strengths of central city areas recognized.


Endnotes

(1) "CACI Free Samples: Results of ZIP Search" at www.demographics.caci.com/scripts/zipsearch.cfm accessed February 24, 2000; "Free Zip Code Report" at www.infods.com/freedata accessed April 26, 2001. [Note: the website locations and descriptions are constantly changing as companies merge and new data (and stereotypes) are promoted.]

(2) "MicroVision Segments" at laguna.natdecys.com accessed March 20, 2000; "About Claritas" at yawyl.claritas.com/ziplookup.asp accessed March 28, 2000; "MicroVision Segments" at www.delluke.claritas.com accessed May 2, 2001.

(3) Walter E. Johnson, "Who's Shopping Your Store? Identifying the Behaviors and Lifestyles of Hardware Store Customers," Do-It-Yourself Retailing, August 2000 at www.diyretailing.com/pdfs/aug.pdf accessed May 4, 2001.

(4) Marketplace newscast for Friday, June 23, 2000 posted on www.marketplace.org/archives/rundown/2000/06.23/html accessed June 26, 2000.

(5) "Nationwide Mutual Insurance Co., et al. v. Housing Opportunities Made Equal, Inc." Record No. 990733, January 14, 2000, at www.lawyersweekly.com/vasup/1990733.htm accessed April 30, 2001.

(6) "Baseball" at www.apbnews.com/resourcecenter accessed May 25, 2001.

(7) "Some Differences between CAP Index Reports and UCR Data" at www.capindex.com/cap_vrs_ucr.htm accessed May 26, 2001.

(8) "APB Data Center" at wysiwyg://6http://www.apbnews.com accessed March 23, 2000.

(9) "Check the Crime Risk in Your College Community" at www.apbnews.com accessed March 23, 2000.

(10) Correspondence with John Pawasarat, December 20, 1999.

(11) "Jonathan Robbin Curriculum Vitae" at ricercar.com/JRCV-D.html, accessed February 28, 2000.

(12) Michael J. Weiss, The Clustering of America (New York: Harper & Row, 1999), p. 14.

(13) Weiss, Clustering of America, pp. 4-5. Robin Page developed the cluster names for the three largest clustering firms (Claritas, Experian and Compusearch). Weiss writes of Page, "He alone is responsible for appellations such as Kids & Cul-de- Sacs (PRIZM), Country Clubbers (US MOSAIC), and Gerberville (PSYTE USA) a feat comparable to one person's naming every car model produced by Detroit's Big Three automakers." Michael J. Weiss, The Clustered World: How We Live, What We Buy, and What It All Means About Who We Are (Boston: Little, Brown and Company, 2000), p. 8.

(14) "Jonathan Robbin Curriculum Vitae." See the quiz on how to locate your cluster in Clustering of America, pp. 393-404.

(15) Weiss, Clustered World.

(16) Weiss, Clustering of America, p. 26.

(17) "About Prizm" at yawyl.claritas.com/faq.asp, accessed March 22, 2000.

(18) "Acorn Lifestyle Segments" at www.premierinsights.com/acorn.html [a discontinued site] accessed February 28, 2000. See also, www.demographics.caci.com/acorn_segments.pdf accessed accessed April 27, 2001.

(19) Descriptions obtained for various zipcodes from laguna.natdecsys.com [a discontinued site] accessed March 15, 2000. As of May 3, 2001 the photgraphs for each of the "MicroVision" segments were stored on the Internet at www.delluke.claritas.com:80/AWYL/segmentlookup_mv.wjsp?segment=1. Not included in the photo analysis are two clusters (Segment 49 "Anomolies" and Segment 50 "Unclassified") that are not used for marketing and do not have photographs displayed.

(20) Only one type, "Shack & Shanty," is not used for the United States. "Global by Experian Demographic Data" at www.cmcus.com/catalog/globalmos.htm accessed February 28, 2000.

(21) "Mosaic Group Descriptions" at www.cmcus.com/catalog/usmgrps.htm accessed February 28, 2000.

(22) Weiss, Clustering of America, pp. 15-16, 74, 89.

(23) Weiss, Clustered World, pp. 26, 34-35.

(24) Weiss, Clustering of America, p.1.

(25) Kurt Chandler, "The Rhythm of Change," Milwaukee Magazine, June 2000.

(26) "CACI Free Samples: Results of ZIP Search" for zipcode 53204 at www.demographics.caci.com/scripts/zipsearch.cfm accessed February 24, 2000.

(27) "Lifestyle Game" for zipcode 53204 at laguna.natdecsys.com accessed October 3, 2000.

(28) "Jonathan Robbin Curriculum Vitae" at ricercar.com/JRCV-D.html accessed February 28, 2000.


Funding for the Milwaukee Purchasing Power Project was provided by the Helen Bader Foundation, the Milwaukee Neighborhood Improvement Development Corporation, and the University of Wisconsin-Milwaukee.


To Top | Mlwaukee Purchasing Power Report | Employment and Training Institute Home Page

Page updated 2009
Employment and Training Institute
University of Wisconsin-Milwaukee

Direct comments to eti@uwm.edu