A

ad hoc report

A summary of computer information conceived after the master files have been created, usually produced by an after-the-fact reporting system designed for the purpose.

ADI

(area of dominant influence) Geographic division of markets by Arbitron, based on preponderance of television viewing.

algorithm

A sequence of instructions that describes how to solve a particular problem.

alphanumeric

Any coding system that provides for letters, numbers, and special symbols.

analysis of variances

The method of determining the significance of observed differences among the means of the subject or group variables. Multiple regression approaches are used to solve complex research problems in more powerful and more practical ways.

ANOVA

(analysis of variance) In research, the results of testing the impact of a variable upon the desired response. If more than one independent variable is tested, the approach is called a "two-way ANOVA."

ASCII

(American standard code for information interchange) Standard code and sorting sequence for representing characters as binary numbers used in microcomputers. (Pronounced: as-kee.)

attribute

Non-graphic information associated with a point, line or area element in a GIS.

automated cartography

The process of drawing maps with the aid of a computer-driven device, such as a plotter and graphic screen. Does not include information processing.

auto-sexing

The assignment of gender, by a computer routine, based on a dictionary of names.

B

The process of completing a mailing beyond list procurement including printing, lettershop, and fulfillment. Also, a measure of a customer's subsequent purchases after the first.

backend results

The measurement of a mailing, including paid and unpaid order information. Used for profitability analysis.

backup

A copy of a data file or program, that can be used if the original is lost or damaged.

bandwidth

The transmission capacity of a communications channel, usually expressed in bits per second (baud rate) or bytes per second.

beta test

The second stage of testing a software program being prepared for market. Usually, testing is conducted in situations similar to the intended usage.

binary

A number system that uses only the digits 0 and 1. Also, a situation in which there are two possible choices.

bit

The smallest unit of binary data; on = 1 or off = 0.

block group

A subset of a census tract, usually no smaller than 600 people. In rural areas, the term is "enumeration district."

brand image

A group of characteristics or associated relationships which a consumer attributes to or identifies with a specific brand.

bulk mail

Second, third, and fourth-class mail serviced on a non-preferential basis by the U.S. Postal Service.

business-to-business

Any business activity directed at or intended to influence corporate decision makers, corporate purchases, or corporate awareness.

byte

A sequence of adjacent bits, usually 8, which comprise a character.

C

carrier presort

Mail identified by carrier route number for cost-efficient mail delivery.

carrier route coding

The identifying carrier route number for mail delivery.

CASS

(Coding Accuracy Support System) A system of the U.S. Postal Service which tests the accuracy of in-house mailing systems.

CD-ROM

(compact disc read-only memory) A form of write-once, disc-based, random-access data storage capable of holding over 500 megabytes of data.

CDS

(Computerized Delivery System) U.S. Postal Service file which provides delivery of city addresses sorted by carrier delivery order.

census tract

A small geographic area, approved by the Census Bureau, into which cities and adjacent areas have been divided for statistical purposes. The average tract is 4,000 residents or 1,200 HH (households). There are over 61,000 tracts in the U.S., according to the 1996 census.

central-location study

In research, a survey conducted to measure consumer preferences in which consumers are recruited to a specific "central" place.

Cheshire label

In direct mail, a mailing label that is computer printed on continuous forms specially designed for affixing to envelopes by a Cheshire machine.

cleansing

The process of correcting or removing a name or address from a mailing list because it is no longer correct or appropriate. (Also see hygiene.)

client/server architecture

A systems architecture that divides functions between two or more computers. The client is the information-requesting machine and the server is the information-supplying machine.

cluster analysis

The science of comparing a series of variable profiles of one subject or group to that of another subject or group and progressively discriminating and associating the subjects or groups into homogeneous clusters. This process statistically minimizes the overall estimate of variation within clusters.

cluster sampling

In research, collectively selecting a target and variables.

cluster selection

A choice of segments, usually based on their defined characteristics, which most closely match a target audience.

coding

Information on reply devices used to identify the source of the address. Also, the structure of characters used to classify characteristics of an address on a list.

compiled list

Information derived from censuses, telephone directories, warranty registrations, and other public records.

confidence level

The degree of statistical significance which can be attached to an observed correlation.

confidence limits

Limits based on the extent of variation of the data and the time horizon of forecasting. Based on statistical theory and probability distributions, a confidence interval (a set of confidence limits) can be established for future forecasts.

conversion

The reformatting from one data processing system to another.

correlation analysis

Used to quantify the relationship (if any) between the independent and dependent variables. (Also see regression analysis.)

correlation coefficient

A standardized measure of the relationship between two variables, X and Y. Commonly designated as r, its values range from -1 to +1, indicating a strong negative relationship to a strong positive relationship with 0 (zero) as neutral. It is the co-variance between a pair of standardized variables.

CPI

(cost per inquiry) A method of evaluating a direct marketing campaign based upon the total cost of a mailing divided by the number of inquiries received.

CPM

(cost per thousand) A relative measure of media cost. Total cost of an ad divided by the potential audience reached, expressed in thousands of persons or HH (households).

CRIS/CROSS

U.S. Postal Service file of standardized carrier routes identified by ZIP+4 address lists.

CRM

Acronym for customer relationship management. CRM entails all aspects of interaction a company has with its customer, whether it is sales or service related. Computerization has changed the way companies are approaching their CRM strategies because it has also changed consumer-buying behavior. With each new advance in technology, especially the proliferation of self-service channels like the Web and WAP phones, more of the relationship is being managed electronically. Organizations are therefore looking for ways to personalize online experiences (a process also referred to as mass customization) through tools such as help desk software, e-mail organizers and Web development apps.

critical path

A step-by-step description of the shortest time in which an entire project can be completed.

cross section

A group of names and addresses selected to be representative of an entire list. (Also see nth selection.)

D

data

Factual symbols, words, or numbers used in computing and printing reports.

data compression

The storing of data in a manner that requires less space on the storage medium.

data dictionary

A document, usually drawn from a computer file, which describes each field and its location within a database.

database

A collection of data, stored on a computer medium, that can be used for more than one purpose.

database management

The task of storing, modifying, and retrieving database information to produce reports, answer queries, and record transactions.

database marketing

Targeted marketing that makes use of customer data captured through customer transactions and communications. This data may be used independently or in conjunction with outside data to guide prospecting, customer retention, promotions and other marketing efforts.

data warehouse

A collection of data designed to support management decision-making. Data warehouses contain a wide variety of data that present a coherent picture of business conditions at a single point in time. Development of a data warehouse includes development of systems to extract data from operating systems plus installation of a warehouse database system that provides managers flexible access to the data. The term data warehousing generally refers to combine many different databases across an entire enterprise.

decile

The portion of a frequency distribution containing one tenth of the total sample.

de-dupe

To check that a mailing list contains only one household (see HH), name, or address for a specific offer and, then, to eliminate these duplications. (Also see merge/purge.)

demographic

The statistics describing factual aspects of a target audience, such as age, sex, race, religion, income, and geographic location.

demography

The vital statistics of a population within an area. U.S. Census data is the major source of demography.

dependent variable

A variable that is determined by some other factor or factors. The variable being predicted in regression analysis.

direct mail

Promotional material delivered to consumers. Unsolicited and bulk mail are often referred to as "junk mail" by the general public.

direct mail advertising

A promotional effort using a mail delivery service for distribution of an advertising message.

direct marketing

An interactive system of marketing which uses one or more advertising media to affect a measurable response and/or transaction at a given location. Direct marketing is an aspect of total marketing that is characterized by measurability and accountability, with heavy reliance on lists and data. (Also see direct response.)

direct response

Marketing directly to consumers, by mail, catalog, or other home delivery, attempting to solicit orders by mail or a toll-free number.

direct response advertising

Advertising through any medium designed to generate a measurable response.

discriminate analysis

The science of determining significant differences, and the nature of these differences, between two or more subjects or groups. These subjects or groups are defined in terms of many variables.

download

To transmit a file or program from a central computer to a smaller or remote computer.

dynamic pricing

The popularity of eBay and priceline.com, marks the beginning of an important business trend. From now on, online prices for a growing number of consumer and business goods will change frequently according to supply and demand. For instance, some companies are using variable-pricing strategies to maximize profits, much as airlines do with ticket prices. Numerous firms are producing the necessary software.

E

EBCDIC

(extended binary coded decimal information code) The numeric representation of characters on large computers.

80/20 rule

The "heavy user/heavy seller" theory which maintains that 20% of consumers buy 80% of a product, 20% of retailers sell 80%, etc. Also known as "Pareto's Law," after the 19th century economist.

enhancement

Using compiled and proprietary data to upgrade information contained in a list or database. Often referred to as "appending" data.

extract

A logical or physical file composed of pertinent name and address records usually selected by passing or failing discriminate criteria.

F

factor analysis

The science of reducing a set of subject or group variables into a set of non-correlated variables which retain the most desired information. Exposes the essential determining constructs behind a set of observable behaviors.

fifth-generation computers

Machines of the future which will be capable of knowledge processing instead of merely data processing.

fixed field

A computer format wherein each record is allocated to a fixed number of bytes, regardless of its actual length. (In contrast, see variable field.)

flat file

A computer readable file which contains no information in any field about the order in which to read records. A two-dimensional array of data items. (In contrast, see relational database.)

floppy disk

A small, portable, data storage medium constructed of plastic with a magnetic coating.

focus group

A form of research utilizing a moderated session of 10 to 12 consumers. Usually conducted in a research facility with a one-way mirror, so that consumer behavior can be observed. Results are qualitative.

forecasting

The prediction of values of a variable based on known past values of that variable or other related variables. Also may be based on expert judgments which, in turn, are based on historical data and experience.

frequency

The number of times a person transacts business (buys, orders, charges, etc.) within a specific time period. (Also see RFM.) In media, the number of times a target audience is exposed to an ad or promotional message.

F-test

In statistics, the ratio of two mean squares (variances) can often be used to test the significance of some item of interest.

FTP

(file transfer protocol) The standardized computer guidelines for retrieving files from the Internet.

function

A statement of the relationship between variables. Virtually all of the quantitative forecasting methods involve a functional relationship between the item to be forecast and either previous values of that item, previous error values, or other independent variables.

G

geo-coding

The process of assigning geographic designations to name and address records.

gigabyte

One billion bytes.

GIS

(geographic information system) A system of capturing, checking, integrating, analyzing, and displaying data about the earth that is spatially referenced. Normally includes a spatially referenced database and appropriate applications software.

H

HH

(households) A living unit. Many modified definitions include: single-family HH, multi-family HH, female-headed HH, etc.

HOH

(head of household) The person primarily responsible in a given HH.

hot-line lists

The most recently added names available on a specific list, usually not older than three months.

house list

A proprietary list of names owned by a company as a result of either a compilation of inquiries and transactions or an acquisition from external sources on a buy-out basis.

HTML

(hypertext mark-up language) The standard format for documents on the World Wide Web.

hygiene

The process of correcting or removing names and addresses from a mailing list.

hypermedia

A system where documents, pictures, sound, movie and animation files scattered across many sites on the World Wide Web are directly linked.

hypertext

A system where documents scattered across many sites on the World Wide Web are directly linked.

I-J

independent variable

A variable whose values are determined outside of the system being modeled. Used in a causal relationship to predict values of a dependent variable.

index

The ratio of a specific group against the national or market-area.

K

key

The field or fields used to identify or locate a record.

L

latitude

Angular distance, expressed in degrees and minutes, along a meridian north or south of the equator.

lifestyle

A way of life or purchase behavior that reflects the attitudes and values of an individual or group. Lifestyle variables are usually measured through samplings or from sales statistics.

list broker

A specialist who makes arrangements for one company to rent the list(s) of another company.

longitude

Angular distance, expressed in degrees and minutes, east or west from a standard meridian to the meridian of Greenwich.

M

magnetic tape

A storage device for electronically recording and reproducing defined bits of data via computer.

market segmentation

Dividing the total heterogeneous market into smaller, homogeneous segments based on geographic, demographic, psychographic, and/or lifestyle variables.

mean

The sum of the values of the items divided by the number of items. Also known as arithmetic average. Frequently used as a measure of location for a frequency or probability distribution.

median

The value of the middle item when all the items are arranged in either ascending or descending order of magnitude. Frequently used as a measure of location for a frequency or probability distribution.

megabyte

One million characters of computer information. Usually applied to storage devices such as tape, disks, or random access memory.

memory

The space within a computer where information is stored while being actively worked on.

merge

To combine two information files into one in a logical fashion.

merge/purge

To combine two files into one in such a way that duplicates are recognized and eliminated. (Also see de-dupe.)

metadata

Data about data. Includes information about the quality, condition, accuracy, collection method, and timeliness of the data.

MIPS

(millions of instructions per second) A measure of computer hardware speed.

model

A symbolic representation of reality. In quantitative forecasting methods, a specific model is used to represent the basic pattern contained in the data. This may be a regression model, which is causal in nature, or a time-series model.

modem

(modulator/demodulator) A device that translates digital pulses from a computer into analog signals for telephone transmission.

module

Part of a larger system. In programming, a part written and tested separately, then integrated into a large system.

MSA

(Metropolitan Statistical Area) Areas that contain a city of 50,000+ population or an area of 50,000 people within a larger city.

multi-collinearity

In regression analysis, the problem when any linear combination of one subset of regressors (independent variables) is nearly perfectly (perfect = -1 or +1) related to a linear combination of any other subset of regressors. May not be easy to detect. (Also see perfect collinearity.)

multi-contingency frequency rates

The science of taking large volumes of subject observations, breaking them into Boolean subsets, and reporting the frequency and percentage occurrences and other statistical values for the interval distribution of the desired variables.

multiple regression

A statistical technique used in direct marketing to compare the relationship between responses to mailings with Census demographics and/or to characteristics of a selected mailing list. This technique is an extension of simple regression. It allows for more than one independent variable to be included in predicting the value of a dependent variable. For forecasting purposes, a multiple regression equation is often referred to as a causal or explanatory model.

multiple regression analysis

Analyzing and correlating historical observations of variables pertaining to subjects or groups. Then determining the mathematical relationships which predict the dependent variable values as a function of the independent variables.

multi-variate analysis

A statistical analysis involving more than one independent variable. The four forms are regression, discriminate, factor, and cluster.

N

NCOA

(National Change of Address) The U.S. Postal Service's system which consolidates and standardizes all address changes in the U.S. Licensees are updated every two weeks.

neural network

An information processing model that attempts to emulate the way the densely interconnected, parallel structure of the human brain processes information. Good for pattern recognition and classification, able to make generalized decisions regarding imprecise input data. Resistant to input distortions and capable of learning. Best for solving problems that are too complex for traditional methodologies.

normalization

The disintegration of complex data structures based on a set of dependency rules, intended to produce simpler, more stable data structures.

nth selection

A fractional selection that is repeated in sampling a list or database. For example, every 10th would be a selection of records 1, 11, 21, etc.

O

observation

The value of a specific event as expressed on some measurement scale by a single data value. In most forecasting applications, a set of observations is used to provide the data to which the selected model is fit.

OCR

(optical character recognition) A machine identification of printed characters through use of light-sensitive objects.

OLAP

Acronym for Online Analytical processing, a category of software tools that provides analysis of data stored in a database. OLAP tools enable users to analyze different dimensions of multidimensional data. For example, it provides time series and trend analysis views. The chief component of OLAP is the OLAP server, which sits between a client and a database management system (DBMS). The OLAP server understands how data is organized in the database and has special functions for analyzing the data. There are OLAP servers available for nearly all the major database systems.

P

partial correlation

Provides a measure of the association between a dependent variable and one or more independent variables when the effect of the relationship with other independent variables is held constant.

penetration

A measure of marketing effectiveness in terms of either customer or response transactions. The ratio, expressed as a percentage, of items of interest to the base count, such as the number of mailed pieces or the number of available households. In statistical terms, penetration is the dependent variable whose variation over all ZIP Code clusters is to be explained by one or more selected independent variables.

perceptual mapping

The research technique that utilizes graph coordinates to analyze the relationships among a group of products, positioning, etc., to identify market gaps and opportunities.

perfect collinearity

In regression analysis, the problem when any linear combination of one subset of regressors (independent variables) is perfectly (-1 or +1) related to a linear combination of any other subset of regressors. (Also see multi-collinearity.) The computational problem arising when two or more regressors (independent variables) are perfectly correlated with one another.

predictive modeling

A statistical process which estimates the value of a dependent variable, given data values of predictor variables. Used to pre-determine response rates of mail offers based on historical response data.

probability

The degree of plausibility, based on available data, of a given event to occur. Expressed as a number from 0 through 1 (impossible = 0; certain = 1).

product differentiation

The real or advertised difference that positively distinguishes one's products from competitors.

product positioning

The process or image projected employing product differentiation to make one's product appear better.

profiling

To build a picture of a target customer by utilizing information from various sources.

psychographics

The data describing the lifestyles, behavioral traits, or mind sets of a target audience - as opposed to factual (or demographic) data.

purge

The process of eliminating duplicates and/or unwanted names and addresses from a mailing list.

Q

qualitative

In research, relates to or involves quality or kind. Considered not projectable.

quantitative

In research, relates to or involves the measurement of quantity or amount. Often used to describe projectable data.

query

A question structured in a database language to sort, group, or select records from a table or multiple tables.

quintile

The portion of a frequency distribution containing one fifth of the total sample.

R

R

The coefficient of correlation which reflects the error in regression analysis. An R value of 1.0 is a perfect correlation. R squared gives the fraction of variance in the dependent variable described by the correlation.

RAID

(redundant array of inexpensive/independent disks) A storage device that uses several disks working together to provide large storage capacity and redundant backup.

random access

A computer mode in which records are organized into a storage file in a non-sequential manner so that any record can be directly accessed.

random sampling

A statistical sampling method involving selection, in such a way that every unit within that population has the same probability of being selected as any other unit.

record

Data elements that are grouped together and treated as one unit, typically stored in a table. Each element is identified by a unique field name.

reformatting

The changing of a format or storage mode from one medium to another, more usable format.

regression

Any modeling of a dependent variable Y as a function of a set of independent variables X1 through Xk. The term dates back to Sir Francis Galton and his work with the heights of siblings in different generations. The heights of children of exceptionally tall (or short) parents "regress" or go back to the mean of the population.

regression analysis

A technique for finding and describing a functional (mathematical) relationship between a dependent variable (such as sales results) and one or more independent variables (such as demographic characteristics of buyers).

relational database

A storage format in which data items can be stored in separate files but linked together to form different relations, thus giving great flexibility.

response device

The turn-around document included with a mailing which identifies the prospect by source code, offer, and name/address.

RFM

(Recency, Frequency, Monetary) A mail-order designation of the quality of a customer. Used to evaluate sales potential of a database list.

rules-based system

A basic "if then" programming routine that can be automated in online stores, such as "If someone buys Hamburger Helper, then they may also buy Coors Light."

S

sales force automation

CRM systems that let salespersons access and manage information about their corporate clients.

sample

A finite or limited number of observations or data values selected from a universe or population of such data values.

SCF

(Sectional Center Facility) A U.S. Postal Service distribution unit comprising all ZIP Codes beginning with the same first three digits.

scoring

The process of using the correlation derived from a model to project and forecast the potential propensity of lifestyles and penetrations.

seed protection

Decoys added to a list. Assures that files rented for one-time usage are not reused.

segmentation

The practice of dividing a customer base into categories or markets - such as income level, geographic location or gender - to better serve them.

set

A named collection of records. Each set must have one record type specified as the owner, which must occur once and only once within the set, and any number of record types specified as members.

SIC

(Standard Industrial Classification) The U.S. Department of Commerce's categorization of businesses.

simple regression

A special case of multiple regression involving a single independent variable. Assumes a linear relationship between the independent variable and the dependent variable. That relationship is estimated using the method of least squares and a set of observed values.

SKU

(stock-keeping unit) The lowest level of inventory identifier. Uniquely identifies every item a company inventories, down to size, shape, color, and weight.

standard deviation

A summary statistic (parameter) for a sample population which is the square root of the variance. It is the square root of the average squared deviation from the mean. Useful for indicating the distribution of the variable's population.

stepwise multiple regression analysis

The science of using the techniques of factor analysis and multiple regression analysis jointly to progressively reduce a set of subject variables to a set of essential, non-correlated, and statistically significant variables.

stepwise regression

The statistical procedure which minimizes the dispersion of a set of data from a point of central tendency. The best mix of predictors for multiple variables is computed by adding a variable to the equation and purging it if dispersion is not reduced. (Also see predictive modeling.)

T

table

A collection of data, in the form of records, with each item uniquely identified by a label or by it's position in relation to other items.

tape density

The number of bits of information written per inch of tape (BPI).

tape dump

The printout of data stored on magnetic tape. Printed for the purpose of editing and checking for correctness, readability, consistency, etc.

terabyte

One trillion characters of computer information.

thematic map

A map displaying selected kinds of information relating to specific themes, such as population density, land use, etc.

time sharing

The multiple utilization of computer time.

topographic map

A map showing natural and man-made features as well as relief, often in the form of contours.

tuple

An element of a table. (Also see record.)

U-V

URL

(uniform resource locator) A type of address that points to a specific document or site on the World Wide Web.

validation

The process of testing the degree to which a model is useful for scoring and making forecasts. The sample data is often split into segments to test different aspects of the model. There are many variations on the process of validation.

variable field

A layout format of list information which assigns a specific sequence to the data, but not specific positions or lengths. (In contrast, see fixed field.)

variance

A summary statistic (parameter) for a sample population which is the average of squared deviations from the mean.

W

WWW

(World Wide Web) A hypertext and hypermedia system that enables users to find information on the Internet.

WYSIWYG

(what you see is what you get) In electronic imaging, a feature that enables all elements of a project to be previewed on the monitor prior to print-out.

X-Y-Z

ZIP Code

A group of five digits used by the U.S. Postal Service to designate specific post offices, station branches, buildings, or large companies.

ZIP+4

The U.S. Postal Service's nine-digit code for a particular block, building, apartment, or business location providing narrowed demographic information. An average ZIP+4 area contains 10-15 households.

Z value or index

The statistical measure of a variable's deviation from its average relative to the variable's total population. The Z value of zero is the variable's population mean. A Z value of +1 signifies that it lies in the upper 80% of the total population and is significantly above the variable's average. A Z value of -2 indicates that it is in the bottom 3% of the population and is significantly below average.