1999 Oct 01
NCR Corporation NCR Relationship Optimizer
David M. Raab
DM News
October, 1999

Every company today wants to optimize relationships with its customers. Systems designed to help fall into two broad categories. One includes touchpoint products such as call center and sales automation systems. These are real-time, online systems where dozens, hundreds or thousands of users interact with individual customers or prospects–answering questions, making offers, maybe even taking orders. The other group is campaign management products such as traditional database marketing systems. These let a smaller number of users send outbound marketing messages to large numbers of customers or prospects via batch processes. Touchpoint systems aim to handle each encounter appropriately; campaign systems execute marketing strategies over time. Never the twain shall meet.

Or shall they? Both touchpoints and strategies are clearly important, and vendors from each camp recognize the need to coordinate them. Traditionally this has been done through a loose alliance between the two types of products–the campaign manager assigns segment and campaign codes to each customer and passes these to the touchpoint system, which append the codes to its own records and uses them to guide interactions. Later, the touchpoint system sends its transactions back to the marketing database that supports the campaign manager.

There are obvious problems with this method. One is the time lag as data shuttle back and forth in batch processes. Another is the difficulty of coordinating campaigns that are defined in one system and executed in another. Still worse, this method relies on the touchpoint system to make the final decision about which message to present during an interaction–even though the campaign manager, having a more strategic perspective, is better equipped to choose. Unfortunately, the campaign manager cannot play this role so long as it lacks the real-time updates. So the touchpoint system is in charge by default.

As the connection between touchpoint and campaign systems becomes more important, vendors from both groups are attempting to encompass both. For touchpoint systems, the first step is running the different touchpoints off a common database–this is accomplished by the integrated “customer relationship management” suites like Siebel, Vantive and Pivotal. The second step is to enhance their campaign management functions, something these vendors are just beginning to address. For campaign management vendors, the usual approach is to share data and campaign logic in near-real-time with selected touchpoint products, usually starting with the Internet. Exchange Applications, Prime Response, and Recognition Systems have all made recent moves in this direction.

NCR Relationship Optimizer (NCR Corporation, 937-445-5000, www.ncr.com) represents a different approach to integrating touchpoints with campaign management. Relationship Optimizer remains committed to batch processing, but focuses on reacting to events rather than executing predefined campaigns. It also manages tasks such as call center and sales resource allocation that have traditionally been handled within the touchpoint systems. The result is to give Relationship Optimizer considerably more control over customer treatment than a traditional campaign manager, even if it does not make the real-time decisions themselves.

The primary unit of Relationship Optimizer is the promotional contact, which the system somewhat obscurely calls a “vector”. Vectors are assigned selection criteria that define which customers are eligible to receive them, schedules that specify dates they are available, cycles that control how often they are executed, priorities that determine which is chosen when a customer is eligible for several vectors, limits that specify how soon a vector can be applied after others, communication materials that hold vector contents, and definitions of what counts as a response. The primary aim of these parameters is to ensure that the most appropriate message is chosen in each situation.

Users can also specify the maximum number of times a vector can be executed in a time period, the maximum number and the mix of vectors that can be sent to a channel resource such as a call center or individual agent, the priorities assigned to different vectors within a channel, and rules for treating vectors that exceed channel capacity. These parameters seek to ensure that channel resources are used as effectively as possible. Traditional campaign managers provide little or no support for this type of resource management.

A vector is typically triggered by a specific event, such as an inquiry or a behavior captured in a transaction system. But Relationship Optimizer also lets the user select a group of customers and assign them to a vector for a conventional outbound campaign. Users can split the universe assigned to a vector into test and control groups, and can link a vector to a sequence of contacts spread out over time, with different branches based on interim behavior.

Intriguingly, Relationship Optimizer also tracks the vectors a customer was eligible to receive but did not, because of limited channel capacity or a higher priority vector was chosen. This lets it execute those vectors later if appropriate.

All these options could easily overwhelm a user who wants only to set up the sort of simple list selection that still constitutes the bulk of most firms’ database marketing. To avoid this, Relationship Optimizer lets system administrators hide capabilities that are not needed in a particular situation.

The technology underlying Relationship Optimizer is fairly standard, except that it requires use of NCR’s Teradata relational database for data stored internally. This should not be a major problem for most firms, since Teradata is compatible with standard Structured Query Language (SQL) and NCR includes a Teradata license with small installations. This lets NCR rely on standard third-party software (Cognos Impromptu and PowerPlay) for reporting and analysis. Similarly, it relies on third-party software for statistical modeling, with the option to import scoring formulas and execute them as part of a selection. Relationship Optimizer currently runs on Windows 95/98 workstations and NCR Unix servers, with a Windows NT server planned for next year.

The system was released in July, 1999 and currently has one live installation. Pricing is based on the number of customers and channels. Cost begins at $250,000 plus at least $100,000 for services, and can reach into the millions.

* * *

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.

Leave a Reply

You must be logged in to post a comment.