David Raab Article Archive

What must you know about someone to have a relationship?

The answer depends on the type of relationship you have in mind. In the case of relationship marketing, the usual conclusion has been you need detailed behavioral data to create programs powerful enough to justify their considerable expense. Since the only available source of detailed behavioral data is usually a company’s own transactions, relationship marketing has largely been applied to customers.

But this is changing. Better analytical techniques such as neural networks and genetic algorithms allow marketers to make useful distinctions based on less comprehensive information. More efficient marketing tools such as automated campaign managers and lower cost media such as the Internet have reduced the expense of relationship programs, so less powerful programs can still break even. Increasing amounts of business and consumer information are available from vendors like Acxiom, Dun & Bradstreet, Experian, and First Data.

Perhaps most important, business leaders have shown a renewed interest in acquiring new customers–having finally realized that their companies will eventually wither to nothing if they focus solely on existing buyers.

The result is a surge of interest in using external data for relationship marketing. This is good news for the data vendors, whose role shifts from providing marginally-effective enhancements of existing records to offering the basis of new marketing opportunities. In fact, vendors who work out appropriate data sharing arrangements among their customers can place themselves at the very center of the information-driven economy. This perception underlies many of the recent acquisitions that are reshaping the data industry.

It also puts these firms at the center of the privacy issue–one that the industry has long waved away like a buzzing gnat, but can no longer avoid. Today, data gathering companies are still allowed largely to set their own policies (and the good firms have thoughtful, well-enforced rules already in place). But some form of government regulation is almost certainly inevitable and could radically change the amount of data sharing that is allowed. This has two implications for relationship marketers: first, that they should become involved in defining a reasonable set of practices and supporting them in the public debate; second, that they should be certain their own programs neither violate reasonable expectations of privacy nor are based on data sources that are likely to provoke criticism. While this is not a software issue, it is one that no member of the industry can responsibly ignore.

But back to software. The basic technology for incorporating external data has been the same for years: systems break the name and address of a record into components and compare these to names and addresses in the external database. When a match is found, the new data elements are appended to the original record, which will eventually be reincorporated into the data buyer’s marketing database.

This process has traditionally run as a batch job on large mainframe computers, which are well suited for the huge volumes of data involved. The large compiled databases are still built this way.

But a mainframe batch process can take days or even weeks to cycle through the vendor’s data center. If the external data is to be incorporated in marketing programs where response time may be measured in hours or minutes–or in seconds on the Internet–then a faster method is needed. Several vendors offer alternatives:

Database America (800-223-7777; www.databaseameric.com) still offers batch processing for list enhancment, but lets users move files across the Internet itself. Once an account has been set up, the process is almost fully automated and turnaround is typically same day or overnight. Turnaround on small jobs can be quicker still. The service is available for both business and consumer data.

Acxiom (888-459-3282; www.acxiom.com/adn) offers online matching across the Internet through its Acxiom Data Network. Beyond traditional demographics based on name/address matching, Acxiom’s system can do a reverse telephone append (finding the name/address associated with a telephone number); check for address changes; search property, auto, credit and mail order purchase databases; and provide lists of neighbors. Of course, some of this is highly sensitive data and Acxiom has a set of policies to determine which elements are available to which customers. The service is now limited to Acxiom’s consumer databases although it plans to add a business database in the future.

The current version of the system is based on file transfers, with near-instant turnaround. Acxiom eventually intends to allow direct integration with internal systems such as call centers or interactive Web pages. This would let marketers gather name and address from a prospect during a telephone conversation, look up additional information about them in Acxiom’s databases while they are still on the line, and then use this information to direct the conversation accordingly. Acxiom also intends to set up a service that automatically notifies marketers when there is new data on specified individuals, such as names on a customer list.

Pitney Bowes Software Systems (800-624-5377) has partnered with Experian, a major data vendor, to offer a product that lets marketers install an in-house copy of the Experian consumer database. The product, called ReUnion, can validate addresses, provide up to three forwarding addresses, and return basic information including gender, date of birth, phone number, and Social Security Number. It can run as a batch process or be accessed by online data entry or marketing systems. While ReUnion is intended primarily as a data quality tool, it has obvious marketing applications in areas such as fraud detection and insurance. The system is housed on a Windows NT server but can accessed from nearly any platform including mainframes and PCs. The in-house database is updated every eight weeks.

iMarket (800-590-0065, www.imarketinc.com) offers Datastream, which places the Dun & Bradstreet business database on an in-house server and matches user-supplied files against it. The system returns standard business information including company size, SIC code and annual revenues. Matching is currently done in a batch process but plans include real-time processing. Datastream runs on a Windows NT server.

iMarket also has developed what it calls “Open Data Architecture”, under which it builds a master database of company information from multiple sources. Users can pick the elements they want for their database and receive this on CD (updated quarterly) or by the Internet. Once the data is moved onto the user’s computer, it can be used for prospect lists, appended to a company’s existing customer records, or matched against new customers as they are added. A metering mechanism lets users access the entire business universe for analysis and still pay only for the data they actually use.

The “Architecture” includes a standard data format and a company ID scheme based on the Dun & Bradstreet DUNS number. This will eventually let data owners prepare their own lists and make them available directly to buyers, instead of storing them on a master database at iMarket. Because of the common format and ID codes, buyers will still be able to merge the data into a custom database without any special processing.

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.