Quiddity Gives Control to End-UsersIt's easy to get excited about the latest interactive, multichannel, enterprise-wide marketing technology. But even today, most firms still are wrestling with basic campaign management -- selecting names for promotions and tracking the results. Software to address those needs also can be sophisticated, even though it lacks the novelty of newer applications.
Quiddity (May & Speh, 630/964-1501, www.mayspeh.com) illustrates the capabilities that can exist in an advanced campaign management system. Designed over several years in cooperation with May & Speh's high-volume direct marketing clients, the system integrates selection, analysis and production processes. It is intended to give end-users as much direct control over these processes as possible while functioning in an environment where multiple individuals may be involved in a single marketing campaign.
Users begin their Quiddity promotions by selecting a universe, which may be the current marketing database, a group of several historical databases or a previously saved set of records. They then build a segmentation using a tree-based interface. The system offers a full set of standard segmentation capabilities -- complex query logic, user-defined calculations, nested segments, nth or random splits and maximum quantities per cell. It also adds some less common functions, such as combining multiple tree branches into a single cell and automatically applying a single split to every cell in a tree. While these manipulations can be useful in certain circumstances, their visual representation can make a tree difficult to read.
Quiddity can generate cell counts as each new cell is added or wait until the design is complete. This is a fairly standard capability. But the system lets the user perform much more detailed analysis, including browsing the individual records within a cell, splitting the cell on designated variables and showing cross-tabs, statistical profiles and calculated values for cell contents. Users can view the results in a tabular report or a graph or export them as an Excel spreadsheet or other file. An analysis can be saved as a template and applied to other cells in a tree or even to other trees.
Once a segmentation is defined, the user specifies the components of the promotion. These can be pulled from libraries of standard components and include offers, creative treatments, time periods for multistep promotions and a custom category that could be used for variables such as medium. The system automatically creates a matrix with each possible combination of the specified components and lets the user drag segmentation cells onto the matrix. Multiple segments can be placed in the same matrix cell and the same segment can be used in multiple matrix cells -- for example, to send several promotions at different times. Cells also can have conditional filters on their contents, so that individuals receive different treatments depending on how they behave after the promotion starts.
As with segmentation cells, users can define analyses and calculations for promotion matrix cells. These include data from a table or entered directly by the user, such as a cost-per-promotion piece or estimated response rate. This lets Quiddity generate proforma profitability analyses as part of the promotion design process and calculate actual results after the promotion is complete.
The final step in the promotion process is creating the physical output files and sending them to a lettershop or telemarketing center for fulfillment. Quiddity lets authorized users set up this process as well, specifying data and file formats, physical media, shipping destination, carrier, record dumps and other details that have traditionally been handled by service bureau production staff. Keycodes can be assigned in several ways, and different cells in a promotion can be handled separately. When the process is complete, Quiddity can page a clerk to put the output in a box and ship it.
Security in the system is designed to accommodate organizations where several people may be involved in a single promotion. The person who creates a segmentation or promotion plan can define it as private, shared with a workgroup or available to all users. Users are granted different levels of authority, including read-only, read/write or read/write/execute. The system makes sure that only one user at a time is making changes to a plan. It provides an audit trail showing who made changes, and a future release will notify individuals when changes are made to a plan they share. A separate audit feature lets users see the selection processes that generated a segmentation cell or tree.
Despite its many advanced features, Quiddity still lacks a job scheduler that would allow it to execute promotions automatically on a regular schedule. The vendor plans to add this in October. Nor can users define "triggers" that perform tasks automatically when specified events occur, although a database administrator can create a trigger in the underlying database and then make it available for users to employ. The system also lacks integrated modeling capabilities.
Quiddity runs on standard relational databases. It uses separate Unix or Windows NT servers for the application and the underlying database. Workstations run Microsoft Windows and can be connected directly or by modem. Performance depends on the database size and server hardware but will typically return queries against a large database in less than 10 minutes. The largest current installation has more than 500 gigabytes of data.
The system doesn't include database-building capabilities. In fact, existing installations all use databases that are built on mainframes at May & Speh and then copied into Oracle on a May & Speh server. The firm plans to allow clients to run the system in-house in the future.
Pricing for Quiddity is based on an initial set-up fee of $230,000, plus monthly charges for data processing and storage. These range from $2,225 to $3,400 per 25 gigabytes of data per month, depending on the type of backup services provided. There are additional charges for more than six users. The system originally was introduced last year and now has seven live installations with several additional ones under way.
David M. Raab is a consultant specializing in marketing database systems and analysis. He is based near Philadelphia.