Robbin - Curriculum Vitae
Academic BackgroundEntreprenurial Career
Affiliations in the 1950's
Activities in the 1950'sComputer Programs for Factor and Cluster Analysis
Computer Programs for Scientific Applications
Founding of Research Data Processing CorporationFurther Background
Founding of General Analytics Corporation
Founding of Claritas Corporation
Founding of Auricom Corporation
Founding of Ricercar, Incorporated
Activities in the 1960'sManhattan Journey to Work and Park and Ride StudiesActivities in the 1970's
Advertising Research and Benefit Segmentation Analysis
Southern California Gas Company Cross-level Study
Office of Economic Opportunity ContractThe Community Profile ProjectOther Developments during the 1960's
Bibliography of Federal Statistical Data
Typology of Poverty Populations
The Index of Susceptibility to Civil DisorderThe IBM 1130 Computer Statistical System and HCAP
Census Data Access and Use Laboratory ContractCharter of Claritas CorporationActivities in the 1980's
Development of Census Data at the ZIP Code Level
Publication of REZIDE and launching of geodemographic
market targeting applications
Geodemographic market targeting projects for TIME and
Development of the Claritas Neighborhood Lifestyle
Claritas Cluster Applications to Direct Marketing
Claritas Cluster Applications to Political Campaigning
Contract with NIMH - the Mental Health Demographic
Profile SystemDevelopment of PRIZMActivities in the 1990's
Contract with U.S. Army Recruiting Service
Value-added Data ProductsP$YCLE
BUSINESS PATTERNSAuricoms's Active Investors List
Targeting the Affluent Market
Contract with the National Centers for Resource
Retail Site Location System for Goodwill Industries, International
High Resolution Targeting®
Retail Sales Potential Analysis for Planning and Development with Kissel Consulting Group
Consultations with Federal Agencies
Expert Testimony, Lectures, Courses Taught and Publications
Return to Ricercar Home Page
Robbin - Curriculum Vitae
Jonathan E. Robbin received his A.B. from Harvard College (1951), his A.M. from Columbia University with the Cristo Miron-Lovneau Prize (1955) and pursued graduate studies in sociology, demography, statistics and research methodology at New York University (1956-1960).Affiliations in the 1950's
Before embarking on his entrepreneurial career, Mr. Robbin was a staff member of the Harvard University Laboratory of Social Relations (1952-1953), a staff member of the Watson Scientific Computing Laboratory of International Business Machines Corporation at Columbia University (1954-1955), a Research Associate at the Population Council of the Rockefeller Foundation (1956), and a faculty member of New York University, Department of Sociology (1957-1961).Activities in the 1950's
Return to top
Computer Programs for Factor and Cluster AnalysisEntreprenurial CareerDuring the 1950's, he developed some of the first computer software for large-scale multivariate statistical analysis, including systems to perform factor and hierarchical cluster analysis of data sets with up to two-hundred variables using the IBM 650 and 7090 computers. Mr. Robbin also developed a method for iteratively approximating a SIMPLEX clustering array in which variables or objects were linked (grouped) by maximizing the secondary diagonal of their correlation matrices (ri,i+1) through permutation of vectors. These programs were applied to a number of innovative projects, including:the ecological classification of Bahamian facies for Dr. John Imbrie, a geologist at Columbia University (later a MacArthur Fellow),
a typology of American cities based on 1950 census data and numerous other projects in sociological aggregate data analysis for Dr. Edgar Borgatta of New York University and the Russell Sage Foundation,
and computation of very large correlation matrices for Dr. Charles Westoff's work, "The Third Child," at Princeton University's Office of Population Research.
Computer Programs for Scientific ApplicationsAt the same time, Mr. Robbin also developed ground-breaking software in the scientific computing field, including:the first program for automated design of kinematic linkage mechanisms for Dr. Ferdinand Freudenstein, Professor of Mechanical Engineering at Columbia University,
a system for analyzing the vibrations of rocket hulls including the effects of rotatory inertia and shear deformation for Dr. George Hermann, Professor of Civil Engineering at Columbia University,
and a numerical solution to the six-pad bearing problem as applied to air-driven and suspended magnetic drums for Heard Baumeister, then principal investigator in a number of research projects in mechanical engineering at the Watson Scientific Computing Laboratory of International Business Machines Corporation
Return to top
Founding of Research Data Processing CorporationActivities in the 1960'sMr. Robbin began his entrepreneurial career as President and Chairman of Research Data Processing Corporation (RDPC) which he founded in 1961.Founding of General Analytics CorporationIn 1964, Mr. Robbin became President and Chairman of General Analytics Corporation (GAC), RDPC under a new name. In 1968, GAC became a publicly held corporation. Subsequently, in 1970, the company was acquired by the Republic Corporation.Founding of Claritas CorporationIn 1971, Mr. Robbin repurchased the assets, operating facilities and business of GAC from Republic and founded Claritas Corporation as a wholly owned private entity with offices in Bethesda, Maryland. and was its Chairman, Chief Executive Officer and principal owner from that year until 1984 when the company was partially sold to VNU Business Information Systems (VNUBIS), the U.S. subsidiary of Verenigde Nederlandse Uitgeversbedrijven, a major Dutch publishing concern. From 1984 to 1986, Mr. Robbin was Chairman and a limited partner in Claritas, L.P.. From 1987 to 1989, he held the position of Founder and Chairman Emeritus of the current Claritas Corp., which replaced the partnership and is wholly owned by VNUBIS.Founding of Auricom CorporationIn 1989, Mr. Robbin left Claritas to its new owners and in 1990, founded Auricom Corporation with John Anderson, a former partner at Claritas. Auricom, of which he was Chairman, has a charter of providing services to companies with significant databases who wish to gain added revenue through the derivation of new marketing and information products from their data resources and who desire the most widespread and effective distribution of these products.Founding of Ricercar, IncorporatedMr. Robbin sold his interest in Auricom to Mr. Anderson in 1994 and founded Ricercar, Inc. of which he is Chairman, President and sole owner. Ricercar provides his individual consulting services along with those of his professional associates in the performance of contracts primarily for the building of custom market sizing, targeting, segmentation and siting models for direct, retail and mass marketers. The company also designs and develops new analytical, typological, interpretive and predictive systems supporting marketing and communications, business planning, location and development, human ecology, household sector and small-area economics, demography, social geography and epidemiology. The commercial, scientific, political and educational applications of these systems are packaged by Ricercar in the form of proprietary software, databases and other information products and sold to a wide clientele both directly and through a number of "partner" vendors. In addition, Ricercar is making a major investment in the building of expert systems for marketing information and analysis applications that can be used interactively in both Internet and Intranet environments.
Return to top
Manhattan Journey to Work and Park and Ride StudiesActivities in the 1970'sIn 1962, Mr. Robbin directed RDPC's first large-scale contract in statistical data processing for the Penn-Jersey Transportation Agency, the analysis of a sample of 90,000 commuters into the Manhattan central business district producing the first automated journey to work study of New York City (using the IBM 1410 and 7074 computers). This project was also the first to summarize trip data using postal geography. New York City 1962 Postal Zones provided the "to" and "from" small areas for a commuter origin/destination matrix. Metropolitan Postal Zones were the predecessors of the United States Postal Service's ZIP Codes ("ZIP" is a mnemonic for "Zone Improvement Plan").Advertising Research and Benefit Segmentation Analysis
This work was followed by a project for the Tri-state Transportation Agency (a regional transportation planning operation supported by the state governments of New York, New Jersey and Connecticut). The Tri-state effort entailed a forecast of commuter demand for a contemplated new railroad park and ride stop on the Pennsylvania Railroad's main line, the Jersey Avenue Station in New Brunswick, New Jersey. Mr. Robbin designed the questionnaire instruments and sampling procedures, directed data gathering on local and interstate highways, buses and railroads, oversaw the data processing and performed the statistical analyses which developed the study's findings.In the early 1960's, while heading RDPC, Mr. Robbin also directed many cutting edge projects in advertising and marketing research using new multivariate statistical programs for clients including Grey Advertising Agency, Benton and Bowles, Kenyon and Eckhardt, IBM, Pepsico, Lever Brothers, Ford Motor Company, Merck, Warner Lambert and others. Statistical models and computer programs for large-scale Q-sort (transpose) factor analysis and multi-dimensional linear optimization were originated by Mr. Robbin to support Dr. Russ Haley's implementation of "Benefit Segmentation Analysis" at Grey. Dr. Haley was the innovator of this powerful new method of uncovering segments of consumers defined by the benefits they valued most in specific goods or services and the degree to which the satisfaction of such benefits by competing products determined ultimate purchase behavior and market shares. These procedures are still widely and effectively used in contemporary marketing and advertising research.Southern California Gas Company Cross-level StudyIn 1965, while Chairman of GAC (General Analytics Corporation), the successor of RDPC., he also directed the design, analysis and implementation of a marketing information system for the Southern California Gas Company, a subsidary of Pacific Lighting Service and Supply. In this project, 1960 census tract demographic data were joined to hand-geocoded customer accounting records, customer surveys and metered household gas-burning appliances to model and explain local residential consumption of utility gas and forecast consumer demand for this fuel by small geographic areas. This was one of the first large-scale marketing research studies to utilize "cross-level" data analysis - i.e., to merge individual consumer survey and administrative data with demographic small area context data in a predictive model.Office of Economic Opportunity Contract
Return to topThe Community Profile ProjectOther Developments during the 1960'sIn 1966, Mr. Robbin established GAC's Washington, D.C. office where he directed the development and implementation of the Community Profile Project for the federal government at the Office of Economic Opportunity (OEO). This large effort for the lead agency in the "War on Poverty" produced an 187,000 page report containing natural language narrative descriptions of the general and poverty-related social, economic, geographic and demographic characteristics of each of the 3,134 United States counties using updated 1960 census small area summary data merged with 29 other agencies' data. These reports were designed to report clearly and simply the results of expert statistical analysis, comparison and interpretation of the data for immediate consumption and use by lay persons who were charged with the amelioration of the level of living in their small local constituencies -- in keeping with the Community Action Program (CAP) initiative of the OEO to involve local entities in the reduction of poverty.Bibliography of Federal Statistical DataA byproduct of this effort included one of the first comprehensive annotated bibliographies of federal statistical data and information products for small areas with documentation and metadata at the item level of detail, organizing and hierarchically cross-referencing over fifteen-thousand variables.Typology of Poverty PopulationsWhile contractor to the OEO, Mr. Robbin also created a typology of poverty populations for the agency, including a multivariate "Poverty Index" for small areas simultaneously measuring several dimensions of poverty including: its severity (degree of deprivation), concentration (proportion of population poor) and magnitude (number of poor persons and families).The Index of Susceptibility to Civil DisorderIn addition, he developed for OEO an "Index of Susceptibility to Civil Disorder" for American cities, a discriminant classification model which clearly specified the social and economic antecedents of the 1967 "ghetto riots". The model identified with 87% accuracy the cities in America that experienced the incidence of riots in their Black neighborhoods in 1967. This rate of correct classification of all cities into "riot/non-riot" types rose to 93% after several new disturbances occurred in reaction to Dr. Martin Luther King's assassination in 1968.
Return to topThe IBM 1130 Computer Statistical System and HCAPOther projects in statistical systems design completed under Mr. Robbin's direction at GAC and RDPC in the 1960's included the design and implementation of the 1130 Statistical System for International Business Machines Corporation and the development of HCAP, the "Hierarchical Cluster Analysis Program" for the IBM 7094.INFORMSAt GAC, Mr. Robbin also conceptualized and oversaw the development of INFORMS, a general-purpose Information Management System (IMS) running on the IBM 360 series of computers. INFORMS embodied Mr. Robbin's conviction that statistical analyses should be totally packaged as expert systems, eliminating any need for the end-user to understand or apply statistical methodologies and techniques in order to utilize their results. To this end, INFORMS incorporated a unique report program generation language into a powerful proprietary IMS enabling simplified computer production of narrative texts interpreting the results of statistical analysis and reporting them to end-users in natural language.Census Data Access and Use Laboratory ContractUnder GAC's contract with the Data Access and Use Laboratory of the United States Census Bureau, Mr. Robbin directed the development of reports utilizing INFORMS technology for evaluation in the 1967 New Haven Census Use Study. The Census Use Study was a pioneering effort for the testing and refinement of new technologies in the use of small-area census data. These included the earliest implementations of geographic base systems and record matching, geocoding and computer mapping software as well as trials of special tabulations of microdata, special sample surveys of family health and area travel patterns and systems for satisfaction of local data user interests and needs. The application of INFORMS to 1967 New Haven special census data automatically produced narrative descriptions of all New Haven census tracts which statistically interpreted and comparatively evaluated tract characteristics in natural language. The results of this exercise were reported in Census Use Study No. 5, Data Interests of Local Agencies, published by the Census Bureau in 1970. Therein, the tract narrative profiles were described as the most valued and widely used reports of census data of all the alternative forms presented to non-technical city government personnel in executive administration, health, public safety, social service, city planning, election, transportation, maintenance, and waste management operations.
Return to top
Charter of Claritas CorporationActivities in the 1980'sMr. Robbin organized Claritas in 1971 with the primary purpose of focusing his company's business on the opportunities for growth he foresaw in adding value to census data resources at small-area levels for guiding the planning, execution and evaluation of marketing actions. At that time, census data were tabulated directly only for arbitrary administrative and political units of small-area geography such as minor civil divisions, tracts, block groups and enumeration districts which could be linked to an individual address only through expensive and approximate data processing procedures. Thus, the business community made practically no use of local area census statistics in spite of their vast potential as a market targeting information resource.Development of Census Data at the ZIP Code LevelIn 1972, however, the Census Bureau also attempted to make indirect estimates of its tabulations of 1970 data for five-digit postal ZIP Codes in major metropolitan areas, the "Fifth Count." Through the ZIP Code, data regarding a neighborhood context could be appended to an individual address record at no cost whatsoever. Unfortunately, the Fifth Count contained usable data for only about 4,500 out of the 36,000 ZIP's in the U.S.. One of the first tasks Mr. Robbin completed at Claritas was the massive job of reallocating tabulated 1970 census data from its given administrative and political geographies to all 36,000 five-digit ZIP Code areas. The resulting database was immediately put to use bydirect marketers, who wanted to target their mailings to specific ZIP's with high densities of prospects qualified by census-based socioeconomic indicators predictive of response, credit or repeat purchase.Publication of REZIDE and launching of geodemographic market targetingOne-hundred-and-twenty-one selected characteristics of U.S. five-digit ZIP areas drawn from this database were also published by Claritas in 1974 in a 3,000 page databook, REZIDE, The National Encyclopedia of REesidential ZIP Code DEmography, designed and edited by Mr. Robbin. Several new multivariate indices describing small-area populations were created for REZIDE by Mr. Robbin. These included ZQ, a T-score for each ZIP measuring socioeconomic status and the Family Life Cycle Code describing the dominant pattern of household composition in each ZIP. For the following ten years, this publication and its updates were the only comprehensive printed source showing the economic and demographic characteristics of the populations served by each of the nation's 36,000 five-digit ZIP Codes.Launching of geodemographic market targeting applicationsThe publication of REZIDE and development of the comprehensive ZIP Code database of census summary information made generally available the first comprehensive and practical implement for performing what he termed "geodemographic" market targeting -- the use of demographic data summarized for small geographic areas as independent variables explaining and predicting sales results in the same small areas.Geodemographic market targeting projects for TIME and other magazinesTIME, Inc. and other print media clients made early and extensive use of the new ZIP-based data. Systems built by Mr. Robbin supported their sales of advertising space with circulation profiles geodemographically linked to product purchase data and with geodemographically created special editions (such as TIME-Z). Initial geodemographic mail response models crafted by Mr. Robbin dramatically improved the efficiency of direct promotion of new subscribers and renewals as well as payment performance.Development of the Claritas Neighborhood Lifestyle Cluster System
Return to topA major technical innovation produced in 1974 by Mr. Robbin for Claritas was a system of classification of all U.S. ZIP's, census tracts, enumeration districts and block groups into forty categories. Each of these forty types of neighborhood contains areas with similar socioeconomic characteristics yet is distinctly different overall from the other types.Claritas Cluster Applications to Direct Marketing
The quantitative categorization of local population aggregates was first applied to selected metropolitan areas by the "Chicago School" of human ecologists and the early social area analysts such as Shevky, Bell, Moore, Williams and Tryon. Mr. Robbin extended it to a national level through a multivariate statistical model of his design.
Mr. Robbin began the task of defining a national typology by partitioning over six-hundred demographic and economic indices derived from 1970 census small-area summary data into six domains of related content:socioeconomic status,A bi-level factor analysis reduced the six-hundred indices to thirty-four composite measures explaining 84% of the variance between American neighborhoods over the whole scope of small-area census data. The largest factors extracted replicated the Shevky, Bell dimensions, thus providing a validation over time. All of the nation's ZIP Codes were then compared by a modal-seeded, iterative K-means cluster analysis in the thirty-four dimensional hyperspace derived through factoring. Hierarchical (Ward's) cluster analysis was applied to the vectors of orthogonal measures summarized for each of from twelve to two-hundred cluster solutions in order to reveal the latent structure of the typology. After evaluation of a number of alternative models, forty area types were isolated as the most meaningful and statistically efficient configuration for analysing transactional or sample-survey marketing data containing customer address information. Each of the United States' five-digit ZIP Code, minor civil division, census county subdivision, census tract and block group/enumeration district populations was assigned to one of the forty types. This work was the first application of factorial ecology which described and classified neighborhood-sized social areas on a nationwide scale.
family configuration and
The resulting system of forty small-area types, the Claritas Neighborhood Lifestyle Clusters, first developed with 1970 census data and updated later with the 1980 census and exogenous information was a quantum leap forward in marketing science. It enabled marketers for the first time to discover rapidly latent groups or segments of customers using only their address. This system permitted segments with highest sales potential to be located geographically with great precision for the purpose of efficiently targeting new prospects. In addition, the characteristics of each segment's neighborhood context provided a rich description of its members for use in developing effective and relevant selling messages.
This form of precise, empirically-based, quantitative market segmentation finally obviated the need for qualitative segments based on highly subjective small-sample behavioral studies and intuitive guesswork. The incidence of such "psychographic" segments could not be estimated for small area populations since their underlying data were not measured by census or any other geographically detailed systems. Thus, their value could not be directly evaluated in practice. Occasionally, Claritas had an opportunity to analyze "psychographic" segments that were assigned to survey respondents who were geocoded -- i.e., linked to census small area context data through the relation of their addresses to census geographic codes. Studies of these linked data unequivocally revealed "psychographic" categories to be weak or null predictors of aggregate sales when demographic and economic context data were held constant.
However, the Claritas Neighborhood Lifestyle Cluster system provided a reliable method of projecting meaningful behavioral data to small areas by exploiting the strong relationship between individuals' behavior and their neighborhood contexts. For example, when questionnaire data measuring product consumption were profiled by Claritas Cluster, the sales potential of local areas could be estimated in a manner unbiased by prior market domination or unequal concentrations of distribution, competition and/or advertising effort. This technique of typological prediction contributed a new methodology to survey-based marketing research which significantly incremented its value by extending its use to tactical large-scale marketing operations such as selective promotion by direct mail, "spot" broadcast media, special metropolitan editions of national publications and local delivery zones of newspapers.
Areas of greatest opportunity now could be located by simply ranking all neighborhoods on the difference between actual and estimated sales or on an index of sales potential comparing each local area to a national or regional average. Calculation of a continuous "lift" statistic displayed the descending cumulative benefit of any level of target selection. Also, by Cluster profiling syndicated research or broadcast rating respondents, a linkage was established enabling discovery of the most appropriate and efficient way to reach qualified segments of prospects through broadcast, print, outdoor, direct or other measured media.
Use of geodemographic selection in practice unfailingly demonstrated real and substantial increments in return on marketing investments in application after application. Conceptual problems which had stunted the investigation of cross-level (individual/aggregate) methodologies in academic social research (such as Robinson's "ecological fallacy") were, therefore, relegated to irrelevancy in the real world of marketing. The Claritas Neighborhood Lifestyle Cluster system irrefutably demonstrated the benefit of taking collective actions guided by "ecological" predictions.
Descriptions of the Clusters, including their principle demographic attributes, buying behaviors and cultural highlights may be obtained from a book written by Michael Weiss, The Clustering of America (Harper & Row, N.Y., 1988). Mr. Weiss validated the statistical construct of the Clusters by actually visiting neighborhoods assigned to each Cluster throughout the United States.
Return to topMr. Robbin's cluster model was validated most effectively in the area of direct response promotions where targeting effectiveness could be quantitatively measured. Through selection of prospects in mailing lists who resided in Claritas Clusters shown to be the most responsive by prior analysis, subsequent mailings achieved significantly higher response rates than earlier untargeted mailings. Marginal lists were turned into acceptable promotional resources thereby, increasing both the mailable universe and the profitability of mailing. Claritas Clusters also succeeded in maximizing the profitability of many other diverse marketing actions. Applications of the Claritas Clusters consistently achieved measurable benefits in market targeting efficiency when applied to retail site location, optimization of sales territories, localized analysis of market shares, launching of new businesses, branch location, route delineations, merchandising strategy, distribution and inventory allocations, fund raising and political campaigning.Claritas Cluster Applications to Political CampaigningIn 1978, working with Matt Reese and Hamilton and Associates, Mr. Robbin applied the Claritas Clusters for the first time in support of a large-scale political action, the effort by a consortium of labor unions in Missouri to defeat a "right to work" initiative on the ballot that year. Indices were created by Cluster-profiling Hamilton's polls which pinpointed in geodemographic terms the "leaning persuadables" (voters who could be easily converted to the consortium's position). Cluster-profiling the poll also identified the most effective message for positively influencing each of nine mutually exclusive groups of voters with differing interests and orientations towards the issue.Contract with NIMH - the Mental Health Demographic Profile System
Over a half-million telephone calls, millions of mail pieces and direct house to house canvassing actions were targeted to deliver the nine messages to leaning favorable voters selected by analysis of geodemographic indices at the block group level. Use of television was minimized so as not to incite the opposition. This campaign turned a potential 40%/60% defeat (as predicted by a July poll) into a 60%/40% victory at the polls in November. Dubbed "the New Magic", the system has been frequently used by political analysts since that time to target campaign, fund-raising and public relations activities. It has been applied with particular effectiveness in guiding communications designed to heighten public acceptance of a given point of view on specific issues.Among the larger projects managed by Mr. Robbin at Claritas during the 1970's was a seven year contract with the National Institutes of Mental Health of the federal government for the development and maintenance of the "Mental Health Demographic Profile System." This system organized, analyzed and retrieved small-area 1970 census data in the form of social indicators for targeting services to the populations of Community Mental Health Centers' catchment areas nationwide.
Return to top
Development of PRIZMActivities in the 1990'sUnder Mr. Robbin's direction, Claritas updated the Claritas Neighborhood Lifestyle Cluster System based on the 1980 Census of population and Housing and developed the PRIZM system (Potential Ratings in ZIP Markets) , a widely used, comprehensive marketing data base which linked the forty 1980 U.S. Claritas Neighborhood Lifestyle Cluster types to a broad range of databases containing consumers' sales records or survey respondents identified by ZIP Code. These included:COMPASS GIScustomer information files with frequency, recency and amount of purchases,
print media circulation files,
magazine readership and consumer buying habits for many merchandise and service items surveyed by syndicated research studies such as Mediamark and Simmons,
TV, radio and cable broadcast ratings gathered by Scarborough, Birch, A.C. Neilsen and Arbitron,
newspaper circulations and delivery areas,
R.L. Polk's new automobile registrations and
consumer behavioral and expenditure surveys and panels maintained by Market Research Corporation of America, National Family Opinion and National Purchase Diary.Claritas integrated these PRIZM-coded databases with digital cartographic information, small-area census data and clients' transaction files in a microcomputer software package, COMPASS, one of the first GIS packages specifically developed to support geodemographic market targeting applications.Contract with U.S. Army Recruiting CommandMr. Robbin was also Principal Investigator for Claritas' 1980 contract with the US Army Recruiting Command at Fort Sheridan, Illinois. This extensive project supported General Maxwell R. Thurman's objectives as commander of the Recruiting service to change the perception of the Army as a job of last resort and target high-quality recruits capable of managing challenging technological tasks in a modern fighting force. For this purpose, Mr. Robbin designed custom geodemographic models explaining and predicting success in recruitment of qualified soldiers able to handle a variety of specific military occupational skills (MOS). These models were then applied to the relocation, deployment and quota settings for all recruiting stations throughout the United States, contributing to the successful achievement of General Thurman's goals.Value-added Data Products
Return to topIn addition to overseeing the development of statistical systems, segmentation models and geodemographic applications, Mr. Robbin also directed Claritas' activities in acquiring and adding value to new data bases and updating, projecting, statistically reducing and enhancing the company's principle data resources.
Some of the more widely adopted products and services invented and developed by Mr. Robbin and offered by Claritas include:
P$YCLE, a dedicated geodemographic segmentation system for financial service marketers based on a multivariate analysis of the 1986 Federal Reserve Board's Survey of Consumer Finance and projected to the neighborhood level using the 1980 census Public Use Microdata Sample and small-area summary databases. This system determined the most powerful socioeconomic predictors of consumer usage of financial products and services and employed them to delineate and size twenty-three segments within each of the U.S.'s 250,000 block groups and enumeration districts as well as other small-area geographies. P$YCLE was extensivly used by Chase Manhattan, Citicorp, Chemical, First Interstate, American Express and a large number of other significant entities in the banking and financial services sphere. A number of these institutions have used P$YCLE to guide the reorganization of their businesses from a "product" into a "personal" orientation.
WEALTHBASE, estimates of the distribution of households and persons within small areas in intervals of very high assets and income. These estimates were developed by models which capitalized interest and dividend income data given by the census and greatly extended the Census Bureau's low open-ended intervals (such as $50,000 and over) through extrapolation of least-squares fitted Pareto curves. WEALTHBASE showed distributions of income and financial assets cross-tabulated by age and work status for every neighborhood in the United States. Brokers, bankers and merchandisers marketing to the very affluent found this resource to be highly effective in concentrating their efforts. A by-product of WEALTHBASE was an annual estimate of the number of U.S. millionaires.
NIGHT/DAY, a database covering all metropolitan areas with significant "commutersheds", or persons traveling outside their residential communities on their journey to work. This resource calculated for small areas such as census tracts and ZIP's, the distribution of PRIZM Clusters and selected socioeconomic characteristics not only for area residents (nighttime population) but also for the commuters from other neighborhoods, non-working residents and residents who work where they live (daytime population). Thus, for the first time in this application area, the persons patronizing local establishments during the day could be targeted using a substantially more accurate picture of market potential for neighborhoods impacted by heavy incoming commuting.
AIM, the "AIDS Impact Model", a database at the small-area level resulting from joint analysis of census data and local morbidity and mortality data on AIDS (acquired immunodeficiency syndrome). At-risk populations were estimated for 1,726 areas (PUMA's) tiling the United States using the 1980 census Public Use Microdata Sample (PUMS) and were then projected to small areas by ecological inference, employing census summary tabulations (STF3B) as marginals or bounds. These data were merged with morbidity and mortality statistics from the Centers for Disease Control and Prevention (HHS) and rates of incidence of AIDS estimated thereby. An epidemiological model specified by Dr. John Pickering of the University of Georgia, Athens was evaluated which locally forecasted for health service providers and life and health insurers HIV seropositives and AIDS cases and deaths annually to the end of the century under varying assumptions regarding the length of periods of incubation and progress of the illness from onset to death.
BUSINESS PATTERNS, applications of a special summary produced for Claritas by the Census Bureau at the five-digit ZIP Code level of the data used to develop "County Business Patterns". These data are summaries of administrative statistics gathered from a universe of businesses withholding social security contributions from employees. They include counts and payrolls of establishments by size (number of employees) and by four-digit Standard Industrial Classification codes for each ZIP. The primary models developed using such data were geodemographic targeting of commercial ("business to business") marketing actions, estimation of the volume of retail sales using the association of sales by merchandise line with store type and size, and further evaluation of the characteristics of working (daytime) populations in small areas.
Claritas/VNU currently provides a wide range of products and services for direct, mass and retail marketers closely based on Mr. Robbin's original models, which support local area targeting of sales and promotional actions for a wide variety of commercial clients in almost every industry throughout the world. A number of other "data vendor" firms have also made and marketed simulacra of Mr. Robbin's and Claritas' systems with varying degrees of competence.
Return to top
Since leaving Claritas in 1989, Mr. Robbin has independently pursued a program of continued development of innovative and advanced marketing information systems, services and value-added data resources, first at Auricom Corporation and presently at his wholly owned company, Ricercar, Inc..Further Background
Auricoms's Active Investors ListAt Auricom, Mr. Robbin built the Active Investors List, a direct marketing resource targeting the affluent market. He also directed Auricom in its performance of ad hoc contracts for the building of custom market targeting and segmentation models for a variety of national marketers and financial service institutions.Targeting the Affluent MarketAt Ricercar, Mr. Robbin continues his specialty of targeting the affluent market for a variety of high-end financial service clients. He has recently developed a surprisingly effective method of identifying names and addresses of individuals with significant assets residing in compiled lists. He used this method in creating samples of respondents who were interviewed to provide material for Dr. Thomas Stanley's best-selling books: Ordinary Millionaires, The Millionaire Next Door and The Millionaire Mind (forthcoming). The lists developed by Mr. Robbin for Dr. Stanley provided a remarkable 80% concentration of qualified households (millionaires in liquid assets) for inclusion in Stanley's studies of wealthy Americans. Such households have an incidence of less than 2% in the general population.Ricercar's Contracts with the NCRI, ChesapeakeIn a recent demonstration project Ricercar performed for the National Centers for Resource Innovations (NCRI Chesapeake), Mr. Robbin developed multivariate linkages of social and economic data to the agricultural, environmental and physical characteristics of small rural areas. Factor and cluster analysis were employed to reduce this integrated information resource into a typology of small areas vividly reflecting a variety of homogeneous categories of farm operations and operators. Ecological inference was employed to estimate Concentrated Animal Operations for the Pennsylvania Department of Environmental Resources, identifying small areas where excessive generation of animal produced nutrients could be expected.Retail Site Location System for Goodwill Industries, International
Mr. Robbin's rural clusters in a demonstration area encompassing the Susquehanna watershed were joined to National Resource Inventory (NRI) data measuring both environmental problems affecting agriculture and steps farmers were taking to surmount them. These data were mapped and analyzed in conjunction with the clusters to provide guidance to the United States Department of Agriculture (USDA) and State of Pennsylvania environmental protection agencies in local targeted marketing of conservation practices to farmers in a variety of ecosystems.
In another project undertaken by NCRI, Chesapeake, Mr. Robbin used a maximum likelihood ecological inference technique (iterative proportional fitting) to distribute the data available in Table H16 of the STF3 1990 Census summary files (mode of housing unit sewage disposal) by occupancy status (vacant/occupied), urbanization (urban/rural nonfarm/rural farm residency) and household size (number of persons per household). These estimates gave a population basis for NCRI's estimation of pounds of nitrogen loadings emanating from septic and other non-sewered types of waste disposal systems for each of the 12,388 census blockgroup small areas throughout the Chesapeake Bay Watershed. The underlying interactions of sewerage disposal with urbanization and household size as well as household income, age of housing unit, plumbing, water source and number of units in structure were determined by cross tabulation of the 5% Public Use Microdata Sample of the 1990 Census within each of the 121 Public Use Microareas in or intersecting the Watershed. Mr. Robbin's imputations permitted an accurate assessment of the specific contribution of human nutrient to nitogen loadings to ground water, surface water and the Chesapeake Bay. After industrial waste disposal, the largest source of ground water contamination is nitrates from septic field effluent.
Mr. Robbin has recently advanced the science of retail site location and store performance evaluation by adapting Ricercar's proprietary site location models to a comprehensive automated system which Ricercar programmed for Goodwill Industries, International's (GII) central headquarters. GII services 192 agencies and their 1,400 stores throughout the United States. This system applies the results of empirical research to make a model of nonlinear gravity effects which use the unique configurations of population density and dispersion around a prospective store location to specify a trading area radius encompassing 95% of shoppers. These gravity gradients are then used to customize sales-adjustment factors based on store selling space in square feet and distance (travel time) to store of each consumer household in the trading area.Management Profiling®
The absolute sales potential of each small neighborhood in the trading area is determined by multivariate analysis of household demographic and economic characteristics. Simultaneously, the system models the effects of sales transfer to nearby existing stores and to competitors of varying size offering the same merchandise and price lines. The effects of local economic conditions are also considered in relation to store sales, including comparative current area employment, income, inflation, cost of living and retail sales. The system has proved highly accurate in forecasting sales and in estimating both shopper volumes and dollar sales per shopper. It is being used not only for evaluating specific prospective store sites but also for determining the intrinsic potential of all alternative sites in a large territory. This latter system use supports Mr. Robbin's strategic concept of "proactive" siting. The proactive approach replaces the traditional reactive method of waiting for sites to be proposed by realtors and then running a specific evaluative study on them. Proactive siting enables retailers to identify the most promising small areas in a territory and then to seek appropriate properties located in these areas.Mr. Robbin also has continued to innovate and launch new concepts which add value to large-scale public data resources through new uses of ecological inference in the development of market targeting applications. In particular, he has found ways to extend target marketing instrumentality to small businesses through his proprietary system of "Management Profiling®" which analyzes input from store managers to define market targets.High Resolution Targeting®Through an extensive analysis of microdata in the Census Bureau's PUMS files, he has provided Ricercar with block-group level estimates of numerous powerful new classifications of individuals and households into multivariate categories not found in standard census tabulations and summary files. Until the advent of this system, for example, it was impossible to use census summary data for small areas to target double-earner, middle and high income couples without children living at home and occupying a variety of age brackets. Such important segments, at the heart of many specific promotions (such as cruise packages), and many more custom-defined and complex multidimensional groupings of consumers are now easily tracked, mapped, and appended to prospect files at small-area levels. Ricercar now offers a new service based on these concepts called "HRT®" (High Resolution Targeting®). Clients with prior survey research or transactional databases can now use these resources as input into their custom definitions of high potential consumer market targets and, through High Resolution Targeting®, locate concentrations of these special prospects geographically and in geocoded mailing lists.Geoeconomics®
Return to topIn searching to control more of the variance in the prediction of sales penetration (the proportion of buyers to households in retail territories or of responses to promotions in direct selling), Mr. Robbin has isolated many powerful context variables exogenous to the traditionally used demographic items. In particular, Mr. Robbin has discovered that local variations at small area levels in cost of living, general prosperity and consumer confidence create significant differences in selling success versus a national average after primary demographic factors have been held constant. Economists and federal-level makers of fiscal policy have traditionally focused on macroeconomic indicators and their cyclical movements in highly aggregated (national) form. Distracted by this, marketers inevitably have tended to underestimate the intrinsic value of localized economic data. This situation has led Mr. Robbin to develop a new set of Geoeconomic® variables proprietary to Ricercar for use in supplementing the geodemographic measures and systems now available to target marketers.Retail Sales Potential Analysis for Planning and Development with KCG
Geodemographic market targeting models are powerful predictors of that part of consumption expenditure which is a consequence of life-style, social status and demographic factors. But even persons of identical life-styles, statuses and demographic characteristics who live in different places can have dramatically different expectations of economic conditions and radically different levels of discretionary expenditure based on the local value of their money and the health of their local economies.
Ricercar's Geoeconomics® is a wholly new market targeting system which measures on a current basis those economic aspects of small areas which affect consumer spending. Fundamentally, Geoeconomic® measures predict the varying propensity of consumers to purchase specific goods and services in given small areas, segmenting them by shared household and personal economic characteristics indexed to reflect local economic conditions.
Mr. Robbin's new system of Geoeconomics® measures the income-driven prerogatives of consumer spending and adjusts them to account for local differences in the purchasing power of a dollar. Incomes, expenditures, savings, investments, prices, credit balances and limits, taxes or discounts are meaningfully compared place to place in Mr. Robbin's system by the use of geographically constant "Geodollars®". Thus, the discretionary spending capacity of households and its influence on sales statistics can be computed in standard terms for local areas using Geodollars for the purpose of making unbiased sorts or rankings of the value of marketing opportunities in these local areas. Temporally constant dollars (i.e., current dollars adjusted by consumer price indices) are valuable for meaningful comparison of economic trends over many years. Marketers, however, typically are concerned only with the few months in which their promotions take place. Cross-sectional geographic differences in the value of money over their selling territory at the time of their activity are of much greater relevance to their immediate success than lengthy secular trends.
Another key element of Geoeconomics® is its current measurement of local levels of prosperity and their very recent upward or downward movements. Local boom or bust conditions mediate consumer confidence, hence spending and other household economic behavior, no matter what the socioeconomic spectrum. Geoeconomics® assesses these conditions in real time at small-area levels.
In essence, the Geoeconomic® database which Mr. Robbin developed for Ricercar groups consumers by their personal (household) financial characteristics standardized to remove the effects of local differences in the cost of living and then measures each segment's propensity to consume specific goods and services considering the health of their local economy.
In direct marketing applications, Geoeconomic® measures are appended to lists alone or as qualifiers of individual characteristics or geodemographic types, improving their power to predict sales, hence to select the most promising names.Since 1995, Mr. Robbin has developed proprietary estimates at Ricercar of current retail sales by kind of business (SIC and NAICS) at the ZIP code level in the United States and at the FSA level in Canada. Utilizing these data and the results of running Ricercar's SELM® system (Site Evaluation and Location Model), Ricercar has contributed to a number of large department store and shopping center project planning and development studies performed by the Kissel Consulting Group (KCG) for major developers and retailers.Hyperanalysis®At present, Mr. Robbin is also developing natural sequelae to the automated analysis, interpretation and narrative report generation systems he innovated in the 1960's which have been largely underutilized or left unexploited by the purveyors, users and generators of public data. Having renewed his development of the automated expert interpretation of statistical tables and multivariate analyses and its translation into natural language, he is embedding a new technology proprietary to Ricercar called "Hyperanalysis®" in the online internet and intranet environment where, hopefully, it will turn the losing battle against "statistical illiteracy" into a winning one.
Return to top
Consultations with Federal Agencies
Contemporaneously with Mr. Robbin's business career, he has independently served the government on a number of occasions. In 1971, he held an appointment as Expert Consultant in the Office of the Director, Bureau of the Census, where he advised on development of new information dissemination technologies and applications of multivariate statistical models to small-area census data. Mr. Robbin was appointed by the Secretary of Commerce to a two-year term on the Census Bureau Advisory Committee on Small Areas beginning in 1974. He also served as the designated liaison between the Census Bureau's Advisory Committee and the American Statistical Association's Committee on Small-area Statistics. In 1980, he served the Office of the Director, General Services Administration, as an expert consultant evaluating the GSA's procurement and maintenance of data processing systems in the light of new policies circularized by the Office of Management and Budget.Professional Affiliations
During the 1960's, Mr. Robbin was an active participant in the Federal Statistics Users' Conference and has participated in the Association of Public Data Users (APDU) since the organization's founding in 1975. He served on APDU's Board of Directors from 1997 through 2002 and chaired the organization's Technology Committee (for more information on APDU, go to http://www.apdu.org). He is also a member of the Classification Society of North America (CSNA) (described at http://www.pitt.edu/~csna/csna.html) , the American Statistical Association (ASA) ( http://www.amstat.org/ ) and its Washington chapter, the Washington Statistical Society (http://www.science.gmu.edu/~wss/), the American Marketing Association (AMA) ( http://www.marketingpower.com/ ) and the International Association of Social Science Information Service and Technology (IASSIST) ( http://datalib.library.ualberta.ca/iassist/).Expert Testimony, Lectures and Courses Taught
Mr. Robbin gave expert testimony and guidance to the House of Representatives' Government Information and Individual Rights Subcommittee of the Committee on Government Operations during the hearing held September 17, 1980, Honorable Richardson Preyer presiding, regarding postal geography and its potential for enhancing the uses, scope and value of census data to both the private and public sectors during the debate on merits of establishing the ZIP+4 system.Publications
On a number of occasions from 2001 to the present, he has served the Department of Justice, Drug Enforcement Administration and U.S. Attorneys in over a dozen states as an expert witness in applied statistics and retail marketing. This work entailed the evaluation of traditional and nontraditional patterns of distribution at retail of non-prescription drugs and the likelihood of list 1 chemicals being sold over-the-counter in extraordinary quantities by specific retail industry classifications such as convenience stores. Through an analysis of the United States Economic Census of 1997 Merchandise Line by NAICS code data along with statistics from commercial marketing surveys and retail point-of-sale tracking systems, Mr. Robbin established expected values for the average monthly sale of products containing pseudoephedrine (hcl) by various types of retail outlets. Given evidence of extraordinarily large quantities of such goods sold at wholesale to retailers normally expected to deal in only minute amounts of them, he was able to establish extrinsic proof that such acticvities constituted a "gray" market through which precursor chemicals might be diverted to clandestine and illicit manufacturers of a controlled substance, methamphetamine. His analyses and testimony helped the government prevail in a number of civil and criminal cases brought against wholesale distributors and retail stores allegedly engaging in these practices.
In addition to his years on the faculty of New York University, Mr. Robbin has taught numerous courses and lectured on statistical and scientific data processing, research methodology for the social sciences and marketing research, direct response modeling, political campaign targeting, demography, human ecology, applied statistical analysis of aggregate data, small area classification and cluster analysis, privacy issues in marketing, public data access and use, value-adding to public data, retail site location and store evaluation, sizing and targeting the affluent market and the history of applied geodemography. He taught courses for several years in applied statistics, research methodology and modeling for the American Management Association and for the Nat Ross seminar series sponsored by the Direct Marketing Advertising Association. In addition, he has served as a guest lecturer on a number occasions at Georgetown University and George Washington University explicating modern geodemographic concepts, models and applications.
A bibliography is provided below listing Mr. Robbin's publications, presented papers and non-proprietary research reports as well as references to books and articles by others about his work.
Note: REZIDE, PRIZM, COMPASS, P$YCLE, WEALTHBASE, DAY/NIGHT, AIM and BUSINESS PATTERNS are registered trade names of the Claritas Corporation.
Geoeconomics, Management Profiling, High Resolution Targeting, SELM and Hyperanalysis are registered trade names of Ricercar, Inc.
Copyright, 2004, Ricercar, Inc. and Jonathan Robbin, all rights reserved.