David Raab Article Archive

OKRA Marketing Pinnacle 2000
Customer Insight Company AnalytiX
Third Wave Network MIND
David M. Raab
DM News
April, 1994

Although desktop marketing systems are quite new to most direct marketers, banks have had them for ten years. Over that period, the products have steadily evolved–for example, Marketing Profiles, Inc. now supports Unix and can report against multiple generations of its files, while Harte-Hanks has added a very powerful direct marketing campaign module and is working on a graphical interface. But as the limitations of their original technology became increasingly apparent, some vendors also decided to create an entirely new generation of systems. This column looks at two new products from established bank marketing database vendors, plus a bank marketing system from a new entrant with a radically different approach.

Pinnacle 2000 (OKRA Marketing, 800-275-6572) is the successor to OKRA’s 1987-vintage TotalMarketer. To get adequate speed from the limited PC systems of its day, TotalMarketer used fixed length records with two levels of data (households and accounts). Although fast, this structure placed inherent limits on flexibility. OKRA and other vendors employed great ingenuity to minimize the impact of these limits.

Pinnacle solves the flexibility problem directly, by using a variable record structure that allows each system to have as many fields as necessary. The product also supports data access at three levels rather than two–typically, account, individual and household.

To meet the additional demands of a variable length record, OKRA built the Pinnacle database engine with the OS/2 operating system, which is more powerful than DOS. This allows the system to run queries at about the same speed as TotalMarketer–one to three million records per minute, depending on complexity.

OS/2 also allowed Pinnacle to employ a graphical user interface, to run several tasks at once, and to schedule tasks for deferred or automatically recurring execution. Pinnacle also lets each user or group of users to see a different set of screens, reports and data.

The Pinnacle query interface uses a point-and-shoot process of picking fields, operators and values. To make things easier, the system administrator can assign each field an “alias” that is more descriptive than the actual field name.

The administrator can also add new fields at any time and overwrite existing data by importing a file. Users cannot call up a single record and edit it, although this capability will be added by the end of 1994.

Pinnacle has a very powerful set of custom reporting tools, including distributions, columnar and household formats. Most impressive is a 50 x 100 cross tab that lets the user assign a separate query statement to each row and column, and define up to 25 fields or calculated values to appear within each cell. Report results can be graphed automatically and can be stored permanently on the database.

These features allow the system to tag multiple file segments in a single pass. However, actual selections must be done one segment at a time. These selections can create a tracking file that stores current information about the selected records. This will later be used by specialized promotion analysis reports. The system also includes a utility to set aside a control group as part of the selection process.

Although existing reports provide basic calculations, the company is adding a much-expanded set of features, such a user-defined functions that can be shared by several reports. These will be available in June, along with a powerful matrix mail/campaign management function to handle multi-cell promotions. A link to the HNC neural network modeling software was released in March.

Pinnacle files must currently be built at OKRA. A module to allow updates on the client’s mainframe or PC is also scheduled for June.

The system was released last October, and has five installations in place with several more under way. Although OKRA’s existing business is nearly all in the financial industry, the additional flexibility of Pinnacle is allowing it to target other industries including retail, telecommunications, utilities, health care, publishing and direct marketing.

Pricing on Pinnacle is higher than on TotalMarketer. A one million account system might cost about $200,000 including a year of monthly updates by OKRA, but the exact cost will vary substantially depending on file size, the number of users, conversion effort, training and consulting.

AnalytiX (Customer Insight Company, 303-790-7002) is also based on a totally new database engine. The product uses a type of inverted file structure where data for each field is gathered, compressed, and then stored separately. This means that separate files hold all first names, all last names, all street addresses, etc. Since each query reads only only the exact data it requires, performance on a Pentium PC can approach 20 million records per minute where a single numeric field is involved. A query involving several fields might run at about ten million records per minute, and even a complex report can execute at two million records per minute. This is several times faster than Customer Insight’s original product, which had to read the entire fixed-length record each time a query was run.

Since data from each field is stored independently, AnalytiX can easily mingle records from any number of different source files. But while the original data might be stored at any number of levels–customer, transaction, account, household, etc.–AnalytiX organizes them into just two levels. For reports or calculations at other levels, the system uses keys such as customer ID to group more detailed data during the processing.

AnalytiX queries are defined through a point-and-shoot interface that includes a variety of functions including complex calculations and logical programming statements. Results of a calculation can be stored on the file, and the system administrator can also add a new field and import data from an external source (so long as a matching key is present).

File creation that includes householding must be done at Customer Insight, however. The company plans to provide full client-based updates, but not for a year or more. Nor does the system currently allow direct editing of a single record.

Reporting in AnalytiX is done through a specialized version of the Excel spreadsheet. A custom tool bar allows the user to place AnalytiX data fields in the spreadsheet and to use them in standard Excel calculations and graphs. Special tools can automatically lay out distributions, ranges and percentile groups, and then replicate the defined calculations and graphs for each group. The completed report is a standard Excel spreadsheet, which can be further manipulated, stored or distributed at will.

Like Pinnacle, AnalytiX offers security features that can limit certain users to specific functions, data, screen formats or reports. Customer reports can be set up so that a user who lacks authority to modify them can still select from a list of predefined data groups–say, geographic regions or types of customers. This enables the system to serve users with very little training.

The current version of AnalytiX is basically a data access and reporting tool. It lacks the specialized promotion selection features of a typical marketing database, such as the ability to set aside a control group, automatically create a tracking file, and run specialized promotion analysis reports. It also lacks batch files and multi-tasking abilities, although the user can open a different Windows application (say, word processing) while a report is running. Customer Insight plans to add such capabilities, but has not announced a schedule.

The system currently runs as a stand-alone package or over a Novell network. A client/server version is under consideration, but again no dates have been set.

AnalytiX is aimed primarily at users with several million records or more. A typical installation is expected to cost over a half million dollars, although exact cost is based on the numbers and types of users, update processing volumes at Customer Insight, and the amount of training and consultation. Customer Insight plans to provide several months of on-site consulting to new clients: with this included, cost might run about twice the price of the company’s old system for a similar size installation.

The product was officially released on April 1, 1994, although it has been in operation at four beta sites for some time. Like OKRA, Customer Insight is targeting both financial and non-financial industries.

MIND (Third Wave Network/Transnational Software Services, 212-344-0444) represents an entirely different species of marketing system. While Pinnacle and AnalytiX offer improved versions of the basic functions as their predecessors–that is, householding and very fast access to a proprietary, desktop database–MIND addresses a different set of problems altogether. In essence, the system is designed to coordinate all the contacts an organization has with its customers. This is intended to allow a company to gain the maximum possible benefit from all its marketing, sales and customer service efforts.

Unlike a traditional bank marketing system, MIND does not rely on a proprietary data engine, include a standard householding module, or offer a comprehensive set of data access and reporting tools. Instead, the system reads standard relational database files–DB2 on the mainframe, Non-Stop SQL on Tandem, or OS/2 Database Manager on a PC–and can be linked to external software for specialized processing functions such as householding, reporting and analysis. (Third Wave, which does custom software development and system integration consulting, can provide those solutions as part of a complete package if a customer requires.)

What the system does offer is a sophisticated set of tools to monitor and direct customer interactions. At the core is a very powerful campaign management function that allows a marketer to define sequences of marketing events–each including a mix of direct mail, telephone contacts and personal sales efforts–to be followed in different circumstances, and then to automatically execute these sequences.

The system uses an efficient graphical interface to lay out the alternate sequences that can be followed within each campaign, the tasks associated with each step in each sequence, and even the specific promotional materials used with each task. As a result, the marketer gains extremely precise control over the marketing approach used with each customer–including the ability to switch the customer to a different sequence depending on actions during the campaign.

This core functionality is surrounded by tools to gather the necessary information, select the proper sequence for each customer or prospect, produce the necessary outputs, and analyze the results. Specific capabilities include customer surveys, complaint tracking and response, predictive score cards, statistical profiles, lead distribution to different sales channels, promotion selections including test and control groups, and response analysis reports.

Although these tools provide the framework to integrate and extract the necessary data, they are not complete solutions in themselves. For example, the score card function can apply a product affinity or credit scoring algorithm to a particular customer. But the statistical modeling needed to create the algorithm would be done outside of the system. At present, the algorithm must then be entered into the system by hand, although automated interfaces with products like SAS are under consideration.

Similarly, the system provides very limited custom reporting and calculation abilities.

Query definition and promotion selection are more complete, however. They use a reasonably powerful point-and-shoot interface and will automatically create tracking files for later response analysis. Query speed will depend on the hardware in use, but–since these are standard relational databases–will rarely approach the speed of specialized engines like Pinnacle or AnalytiX.

Using a standard database engine does allow data to be updated or edited at any time, with standard system capabilities. MIND itself can also access an individual record for direct editing if needed. This ability is controlled for each user by the security system, as are the screens, functions, reports and groups of data that each user can access. The security system can be used to allow individual sales agents to define marketing campaigns that apply to their own customers only.

MIND is a true client/server system, designed for a large number of users who do different types of tasks. The server can be an IBM mainframe, Tandem computer or OS/2-based PC, while the client (that is, user workstation) is always an OS/2 PC. The system does not presently support Unix, although Third Wave is looking for an opportunity to make the conversion.

MIND was originally introduced in the United Kingdom in 1991, and now has eight installations world-wide. The product was launched in the U.S. in 1993, but there are no U.S. installations. A million-account system in the United Kingdom would cost about $600,000, plus the fees for consulting and associated custom development. U.S. pricing is still being set.

MIND’s developer, the Third Wave Network, is based in the United Kingdom. It provides software packages, custom systems development and related consulting to the financial industry and for sales and marketing projects in other industries.

* * *