David Raab Article Archive

There is no shortage of PC products for mailing list management. For about $100, you can choose among dozens of systems that will let users create and import lists, do simple duplicate identification, create basic first- and third-class postal sorts, and generate mailing labels. If you look hard enough–say, at MySoftWare Company’s $129.95 MyMailManager (415-325-4222)–you can even get CASS-certified address standardization. Spend $1,000 and you can add unlimited size files, matchcode-based deduplication, automated postal forms and second class postal sorting, from products such as Group 1’s ArcList/Accumail (800-368-5806) and Business Computer Center’s Mail Manager 2000 (800-624-5234). For high-volume processing by service bureaus or list managers–who need mainframe-class deduplication, sophisticated batch processing and multi-job management–the main choices have been DOS and Unix solutions from Postalsoft (608-788-8700). These have recently been joined by Anchor Computer’s IPRO (800-452-2357) and Direct Marketing Software’s LISTman (800-210-9157) under DOS, and Group 1’s products for Unix. Products in this class range from $15,000 to $100,000, and tend towards the higher end of the scale.

With the market covered so thoroughly, it might seem there is little room for additional products. But here are two products that aim at niches that are not quite filled by anyone else.

Applied Information Group MailEasy 2.0 (Applied Information Group, 908-738-8444) attempts to straddle the gap between middle and high-end products, offering service bureau capabilities at a $595 price point. The system, built on the FoxPro database engine and coded in FoxPro and C, is aimed at small service bureaus and lettershops and large-volume in-house mail rooms. Because such operations work largely with lists that are generated elsewhere, MailEasy offers fewer features for on-line data entry and review than an ArcList or Mail Manager 2000, but more sophisticated list conversion and batch processing. MailEasy also provides reporting capabilities that are tailored to service bureau needs, including free-form analysis, such as user-defined cross tabs, and prebuilt reports for merge/purge output, Zip penetration and response analysis. Beyond its primary targets, AIG hopes the low price point will attract smaller mailers as well.

The MailEasy interface uses pull-down menus to access screens that are essentially forms: users check off lists of options to define the particular job

at hand. Once filled out, the screens can be saved as defaults for future jobs, and different screens can be associated with different projects–for example, to allow all jobs for Client A to use one set of options, while all jobs for Client B use another set. To keep things simple, basic options are generally listed on one screen, while choices for more sophisticated users are presented on a separate “Advanced” screen. The interface is quite easy and convenient, although the test version had an annoying habit of hiding error messages behind larger screens, which remained stuck until the error message was uncovered and acknowledged.

The system stores files in the standard dBASE (.DBF) format. It can user externally-generated DBF files without conversion, and provides tools to translate files in common PC formats including ASCII fixed and delimited, DIF, Paradox, Excel, Lotus 1-2-3, and several others. Mainframe tapes must be changed from EBCDIC to ASCII before the system can use them.

The system does not require that files use standard field names, although non-standard names must be “mapped” to indicate which field holds standard elements such as name, street, city, etc. During file conversion, users can import only records meeting a specified condition, add new fields, merge or split existing fields, append a list code and sequence number, change to all upper case or mixed upper/lower case, and overlay data by matching against another file. Users can preview the first fifty imported records on screen before the process is executed.

The system also provides a complete set of file parsing routines, which can split unformatted records into name and address elements (prefix, first name, middle name, last name, suffix, house number, street, etc.). This improves the results of address standardization and matching, and is driven by pattern recognition and keyword tables. The tables such provide lists as common first, last and company names; genders associated with first names; and words associated with business names and titles. They are not directly editable by the users, however, which means they cannot be modified to adapt to special circumstances.

Application of file parsing can be limited to records that meet specified conditions–for example, to exclude foreign or military addresses, which will not match standard U.S. formats. The conditions are defined with the system’s standard expression builder, a point-and-shoot tool that can apply logical, string and mathematical functions to any field in a record. Similar filters can be applied elsewhere in the system to processes such as merge/purge or record selection.

Despite its parsing abilities, MailEasy does not provide CASS-certified address standardization and Zip+4/carrier route coding. AIG plans to release a separate product for this purpose later this year. MailEasy itself does include postal tables that allow it to use Zip code to verify or replace the city and state entries in a record.

Zip-based city/state assignment can also be built into data entry screens.

Apart from this feature, though, the screens lack abilities found in programs built with data entry in mind, such limiting field entries to a specific format (say, xxx-xxx-xxxx for telephone numbers) or to codes found on a user-defined look-up table. Nor can the screens automatically copy specific fields from one record to the next–a feature commonly used when keying lists with certain data in common, such as company name and address. The screens do have the ability to copy an entire record into another, page through a file and find a record based on its contents.

The system does a better job of managing files once they are imported. It can group related files–say, everything being accumulated for a merge/purge–into “sets” that can be called up simply by specifying the “set” name. It also lets the user create a “folder” that holds list names, key codes, order quantities, panels, suppression files and other information related to a specific project. Folders provide labels for reports, help manage keycodes placed on output files, and give data for calculations.

Like folders and file sets, MailEasy’s merge/purge functions are also tailored for the multi-client, multi-job environment of a service bureau or letter shop. One merge can handle an unlimited number of files, drawing on file sets and folders to simplify management. Job setup is also simplified by allowing users to check off a long list of standard exclusion/inclusion conditions such as missing address elements, three- or five-digit Zip files, gender, personal prefixes (Mr., Miss, etc.) and prefix/suffix categories (military, professional or religious). Users can specify priorities among merged files, incorporate seed lists, assign an output format, place finder numbers on records, read keycodes from a folder, and do Nth selections to limit output (of the entire selection or a single panel) to a target quantity. The system tracks record use by automatically creating a log file of record IDs when a selection is run. The system can also place a code on the selected records themselves.

When the merge job is finally run, matches are found using one of six match code algorithms provided with the system–a relatively unsophisticated approach, although AIG customers say results are comparable with mainframe systems. Sophisticated users can incorporate their own matching algorithms if they wish.

Merge/purge reports are quite complete. These include duplicate listings; a summary showing inputs, drops, intra- and inter-file duplicates, and output by file; and a detailed report showing attributes on each file such as gender and address elements. Users could view duplicate listings on-line and manually flag which records to reject, but–since most service bureaus do not perform this type of review–the system does not provide a specific screen for the task.

Output from the merge/purge function can be sent to postal sort routines which can handle first, second and third class mail, including regular and non-profit rates, pallets, and SCF or BMC entry points. Users can choose from a list of standard package sizes, or enter the dimensions and weight of non-standard pieces. Once the sorting is complete, the system can split presort files into different streams for carrier route, Zip and residual sections, in case these will be sent to different printers. It will print bag tags, create completed postal forms, and apply bar codes to mailing labels. If split address fields (segregating street number, street name, house number, etc.) are present on the file, the system can use these instead of the less-accurate combined address line when preparing bar codes.

The system can place its final output in the same ASCII formats that it will import, or create a word processor merge document. It includes a convenient utility to split large ASCII files among several floppy disks. It can also print directly to common PC printers, including line printers, dot matrix and HP lasers. Users can select from standard label formats or create their own.

In addition to merge/purge reporting, the system provides an unusual range of additional capabilities–also driven by service bureau needs. User-defined reports can count occurrences of a single value (e.g., a specific state code), count records that meet a set of criteria defined with the expression generator, create distribution report with counts for all values that occur in a field, and give cross tabs showing combinations of values for two fields or two expressions. The cross tab report can group values into ranges and can present data from several files combined. Standard reports include a response analysis that will flag or count records that match between two different files and a Zip analysis that compares the counts by Zip in a specified file with counts of commercial or residential addresses by Zip from a database provided with the system.

MailEasy tasks can be executed individually from the screen, linked together in a DOS batch file, or built into a “job” from a special menu. The “job” menu lets the user create a stream of tasks by specifying files to use and the sequence of screens to call. It lets the user “validate” a job by checking that all the required fields and files will exist when they are needed–even if they do not exist at the start of the job but will be created during an intermediate processing step. This is particularly useful for jobs that will run unattended.

Versions of MailEasy are available for both DOS and Windows, and can be purchased for $595 each or $690 for both. Support costs $99 per year and includes toll-paid telephone support from 8 a.m. to 7 p.m. Eastern time along with quarterly updates to the postal presort tables. The system comes with a thorough users manual (lacking an index, however) and provides extensive on-line help. MailEasy was released in late 1994, and has been sold to about fifty customers. A free demo version is available from AIG.

MIRUS (MIRUS, 214-881-2424) targets the upper end of the service bureau market: firms moving over one million pieces of mail per month, who typically are using mainframe software. The system can run on a single PC, but reaches its full potential when a central file server is linked to other PCs that are dedicated to job set-up, data access, batch job execution and printing. One MIRUS file server can drive as many as 255 separate print servers, each hooked through proprietary channel interface processors to one large laser printer or up to six impact printers. MIRUS is the only PC product that can currently drive the big Xerox, Seimens and other laser printers used by large personalization houses, although Postalsoft is expected to add this capability in the near future.

The building blocks of MIRUS are purposely simple–Assembler-language programs, DOS operating system, Novell or Lantastic networks, ASCII or EBCDIC files–because these allow the most efficient use of PC resources. Other than the channel interface and a security keycard, the system requires only standard PC hardware. MIRUS can provide the hardware as well as the software for customers who want a complete installation. The system can also use more exotic technologies, such as high capacity RAID drives or tape systems, to improve performance when needed.

Most work in MIRUS is done in batch processes that are controlled through “control books”. A control book specifies the input and output files for each job, as well as the procesing steps required. Books can be created directly by a user (typically by modifying the book for a previous job) or generated automatically by filling in screen-based forms developed for standard tasks such as list fulfillment. Processing is specified by calling tasks built directly into the system or by accessing “DoAll” routines. Over fifty “DoAlls” are provided with the system, with capabilities from assigning sequence numbers to doing Nth selections to changing a value based on a look-up table. According to MIRUS, they allow nearly all tasks to be specified without any conventional programming, providing a huge cost and productivity gain.

There is no data entry component to MIRUS, although a FileScan utility can be used to view or modify small amounts of data such as test records. The file conversion routines can use “DoAlls” to add and delete fields; perform string, math and date manipulations; replace values based on tables nested up to 15 deep; parse name, title and address fields into components; append gender codes; change to upper or mixed case; assign city names based on Zip code; and append list codes to the records. Tables for parsing and gender assignment cannot be changed by the user. The system does provide CASS-certified address standardization, Zip+4, carrier route and bar coding.

The system’s merge/purge can accept up to 255 input files in a single job. Users typically place files for the same job in a single DOS subdirectory, although they can also be assigned an explicit linkage or combined into a single file before the merge begins. Very large jobs can be split by the first one or two digits of the Zip code and run separately, with results automatically recombined into a single job.

Users can set priorities among files to determine which member of a duplicate set will be kept, although MIRUS will override the priorities if one record appears more complete than the other. The user can also review questionable matches in a printed report or on screen with the FileScan utility if desired.

The system uses a proprietary matching algorithm that compares patterns found in user-selected portions of the records. This is clearly more powerful than a simple match code, but we were not able to compare it to typical mainframe software. In addition to selecting the fields, users can set the degree of tightness, whether to match with name and address or address only, and whether to look for commercial or residential matches. To improve its accuracy where possible, the system will automatically standardize addresses of records that appear to be potential matches and try the match again.

Standard merge/purge reports summarize results, list duplicates found, and provide data for a cross reference report by list. “DoAll” functions can also create basic counts and distribution results, which can be combined by the user to create response analysis and other reports.

MIRUS has several special functions for list order processing. It comes with standard screens for basic selections, can do Nth or random samples, can put different keycodes on records based on user-defined conditions, will match against a Zip selection or suppression file, and automatically creates an index of records included in each selection. The system will automatically process all jobs related to the same file in a single pass, creating a separate output file for each job.

The system includes a very complete set of postal sorting capabilities, including first, second, third and fourth class, including palletization. It creates tags and postal reports, and can send different labels for bundle types to different files or printers. Unlike many systems, it can actually parse print image files during the postal sorting process, apply Zip+4 or carrier route codes, place the records in the appropriate sequence, and then output the original print images.

Of all MIRUS’ features, its personalization and printing abilities set it apart most clearly. The system’s LetterWriter module provides full text personalization, including word wrap, proportional spacing, different text blocks and images based on data within each record, and on-the-fly font selection, rotation and size changes. The system can interface directly with mainframe and other printers, using a proprietary printer control language that is then translated to the language specific to the particular printer. The system’s print servers are fast enough that they can drive the printers directly, rather than by creating files that are then fed to the printer from a spooler. This saves a great deal of disk storage.

Prices for MIRUS are based on the modules purchased and the number of records processed by the buyer. A complete set costs about $75,000 for a shop printing several million pieces of mail per month, with support at 16% per year. Clients get seven day, twenty four hour telephone service, plus monthly product changes downloaded from a computer bulletin board. New customers get one week of training in the company’s Dallas office. The system was introduced in 1987 and now has about 50 installations.