David Raab Article Archive

As companies do more database marketing, they place new demands on their database marketing systems. While it was once enough to execute individual promotions well, these systems must now also make it efficient to manage dozens or even hundreds of simultaneous promotions, integrate closely with operational systems such as customer service call centers, and provide a strategic overview of contacts with each market segment. The database marketing system becomes the chief operational system of the marketing department, used for planning and evaluation as well as execution.

This evolution has a number of implications. Queries, segmentations and selections, which were tightly integrated in traditional database marketing systems and typically created anew for each promotion, become increasingly distinct as unchanging customer segments become the targets of long-term, multi-promotion marketing strategies. Scheduling facilities become more important as multiple promotions may be planned to execute at different times, as individual promotions are repeated automatically, and as individual promotions contain a time-phased sequence of steps. Response analysis becomes more sophisticated and more reliant on standardized measures so different promotions can be compared consistently. The very notion of a “promotion” expands to include marketing contacts that do not originate with lists generated by the system–such as broadcast media or print advertising–and even non-marketing treatments such as alternative pricing schemes or preferred customer service. The traditional focus on query speed and complexity becomes less central, both because other requirements have been added and because there is less need for complex, on-the-fly analysis and selections in a world of semi-permanent segments and automatically executed promotions. Proprietary database systems, which are often fast but limited in other capabilities, become harder to justify, although proprietary technologies may find new uses as query accelerators within non-proprietary systems.

In fact, the relation between the new form of database marketing system and the marketing database is considerably less exclusive than in the traditional model. The new database marketing system will be connected to other corporate databases, such as budgeting, customer service and order processing, in addition to the marketing database itself. At the same time, data mining and reporting tools may access the marketing database independently of the database marketing system. In some cases, the corporate data warehouse will have replaced the conventional marketing database altogether–allowing the paradox of database marketing without a marketing database. But I digress.

Ovation (Paradigm Communications, 800-749-0228; 813-287-0028) was designed to support operations at Paradigm Communications, a marketing agency which offers advertising, fulfillment, telemarketing, broadcast fax, Yellow Pages ad placement, and other services. The system lets users build marketing programs based on predefined “tasks”, which might include letters, telemarketing scripts, surveys, e-mail, fax messages, file outputs, or other marketing actions. For telemarketing tasks, the system provides functions to create scripts and surveys, to execute these, and to place the replies directly into the marketing database.

Each task has a number of attributes. Some attributes describe the conventional information needed for any marketing program: content (such as letter text, telemarketing script, or file layout), start date, end date, unit costs, value per response, media, objective, and so on. But the attributes also include information needed for marketing operations: who will prepare the task, the time and cost of preparation, when the task is due, and current status. Ovation uses this information to prepare lists of responsibilities for individual workers, to highlight overdue tasks, and to provide managers with an overview of workload. The user can also specify an e-mail message to send automatically when task preparation is completed, and who should receive it.

In addition, attributes can include an estimated percentage of records who will complete the task. This can later be combined with input counts to calculate quantities, costs and revenues for program budgets and resource planning. Actual results can later be posted against the tasks to allow comparison against budgets. The actuals can also be searched by a “corporate memory” function to help predict the results of future promotions. Results are classified into standard categories to allow comparisons across programs.

Marketing programs are built by linking tasks in an on-screen tree. Records can pass from one task to the next task automatically, a set number of days after the prior task begins or ends, or when a certain number or percentage of records in the prior task are completed. Records can branch to different tasks depending on database queries, replies to a telemarketing script, or other responses. Programs can have as many branches and levels as desired.

Customers or prospects are assigned to the marketing program by drawing from a separate library of market segments. These segments are created with a segment builder, which uses another tree structure to split the file into different groups. Splits can be on the basis of Nth selections, on equal-sized ranked segments such as deciles, on standard queries written directly in Structured Query Language (SQL), or on queries created by Ovation’s own query builder or any other query tool. The Ovation query builder is limited to fairly simple queries–for example, it does not allow embedded calculations or subqueries–although technically skilled users can embed SQL for more complex tasks if desired. Once segments are created in the segment builder, the user can execute the segmentation to see actual counts, view the details of the selected records, and add the new segments to the segment library.

After a program has been constructed and segments are assigned, it is submitted to the Ovation server for execution. Tasks that require database interaction–such as checking for responses to a promotion–run automatically. Reports are created using a large table that mimics an Excel spreadsheet and holds all the information about each task within the program. Users can create custom reports as subsidiary spreadsheets, drawing data from the main table and using standard spreadsheet functions. The reports can be exported as live spreadsheets, text files, HTML pages, or other formats.

Ovation is written with the Delphi application development system and uses Windows 95 workstations. The system runs on a Windows 95 or Windows NT server and can be connected directly to Oracle, Sybase, Informix or SQL Server databases or indirectly to other relational databases via Borland Database Engine drivers. Ovation will create its own promotion history records and can store responses to telemarketing surveys entered through its screens, but otherwise assumes that the main marketing database is maintained independently.

The product is priced at $75,000 for a single site or $300,000 for an enterprise license, plus 12% annual maintenance. Implementation services are typically $20,000 to $40,000 if a marketing database already exists. Ovation has been used within Paradigm since 1996 and has been installed at three external clients.

David M. Raab is a Principal at Raab Associates Inc., a consultancy specializing in marketing technology and analytics. He can be reached at draab@raabassociates.com.