1997 Jul 01
Business to Business Database Marketing Systems
David M. Raab
Relationship Marketing Report
July, 1997

Business to business database marketing systems must do everything that consumer marketing systems do, and then some. Probably the most obvious additional requirement is a more complex data structure: where most consumer databases can work primarily at the household level, a business database usually needs account, individual, physical site and headquarters levels, and often must deal with several layers of ownership as well. Moreover, the business database must record the relationships among these levels: an individual might be associated with some accounts but not others; an account might service some sites or the entire company; an individual might work at a branch or at headquarters. While capturing this data is arguably the largest challenge in business database marketing, just creating a structure to store these relationships in a way that makes them easily accessible is surprisingly difficult.

The second challenge facing business databases is data standardization and consolidation. Some one once counted 16 ways that IBM Corporation might be entered in a database: IBM, International Business Machines, IBM Corp, etc. These variants must be identified and brought together in a way that does not create false matches–you don’t want to combine Illinois Burnoose Makers with IBM just because the initials are the same. The same company also may have different trading names, subsidiaries with totally different names, different physical sites, and different mailing, shipping and physical addresses for the same site. And of course some businesses addresses will have an individual name or department without the company name showing at all.

Software built to match consumer addresses cannot even begin to deal with all these situations. Instead, business systems must rely on specialized matching logic and reference tables. In particular, business systems need to link records that have different combinations of data in common: that is, to decide that “Frank Jones, Ray’s Pizza, 456 Main Street” really matches “Joe Smith, Ray’s Pizza, PO Box 123” because there is also “Joe Smith, Ray’s Pizza, 456 Main Street”. This typically requires two passes of the file: the first to match the records at 456 Main Street, and the second to match the two records with Joe Smith. Note that you wouldn’t want to simply link all the Ray’s Pizza’s, because the name is commonn to many independent businesses–although this mistake is often made.

But even the cleverest matching algorithms cannot know that IBM and Lotus Development are part of the same company. This requires reference to tables that contain these connections. Some firms build these tables for themselves based on their knowledge of the market. This is practical in a small industry or where a large field force is covering all customers and prospects intensively. But most business marketers rely on data compilers such as Dun & Bradstreet or American Business Information, who create massive lists of all firms in the country and who owns them. Files created by these companies contain codes for the corporate parent and ultimate owner of each business or site, which allow them to bring together records that would otherwise appear unrelated. These files also provide a base for standardization. But just because a record does not match one of the compiled lists does not mean it is invalid. Even a well-maintained customer file will probably only match 50% to 60% of its records against the compiled business files. Files with many small companies and home offices will have particular problems.

Compiled lists also account for the third major difference between consumer and business databases: business databases very often contain detailed prospect lists acquired from the compilers. Some consumer files also contain large numbers of prospects, but this is less common. Having a comprehensive prospect list allows business databases to perform penetration and market analyses that compare the current base of customers with all potential customers. Compiled lists also allow marketers to execute targeted acquisition campaigns without the expense of building lists through lead generation or by combining lists from many smaller sources.

Finally, business database marketing systems often contain a level of contact detail that is not stored on consumer databases. This information is gathered by field and telephone sales representatives who write down their notes after each conversation with a customer or prospect and may even set a date to call them back. Contacts often relate to a specific sales opportunity, which may itself generate price quotations or sales forecasts to be stored in the database. This adds still more complexity to the data structure.

With all these special requirements, business marketers clearly need their own kind of software. In fact, there are two separate classes of systems used primarily for business database marketing.

The first set of systems are built primarily for list generation and marketed by the major business data compilers. Dun & Bradstreet sells its own MarketSpectrum (201-605-6270) while American Business Information has an arrangement with B2B Inc. (770-753-0988) to sell their product, also called B2B. Both systems can handle reasonably complex marketing selections, store promotion history, provide simple response analysis and reporting, and display relationships between a given company and its corporate parents and children. They can also both be purchased with a package of services that includes loading the user’s own customers, appending information about these customers from the vendor’s compiled business database, and the enhanced customer list along with a universe of similar prospects. Such packages start at about $30,000 per year.

The other major class of business marketing software is focused on contact management rather than list generation. There are literally hundreds of products in this group, ranging from relatively simple contact managers, such as Symantec ACT! (408-253-9600), to sophisticated sales automation systems like SalesKit (800-779-7205) and MarketForce (800-766-7355). Prices range from $100 to $1,500 per user. All these systems allow users to enter data directly into the records of individual customers and prospects, to document the results of conversations and to set callback dates. The simplest products use a single large note field for such information–easily readable by humans, but difficult to access in database queries and reports. Most systems store each contact as a separate database record, which makes large volumes of data more manageable. Similarly, while the simplest systems treat each customer or prospect as an independent individual, making it difficult to do company-level marketing, most products use a multi-level file structure that can link individuals to a site and parent company. Advanced systems have even more complex structures complete with sales opportunities, relationships among individuals, and relationships between one individual and multiple companies.

The complex data structures of these systems are designed primarily for an online user who wants to examine a single relationship or project in detail. The databases are designed like conventional online transaction processing systems, making it quick to retrieve all data associated with a single entity but not necessarily fast to do analysis and queries against the entire file. In fact, the query and list selection tools provided with most contact management and sales automation systems are distinctly limited in the complexity of the queries they can generate. Users who expect to develop large-scale marketing campaigns may find they need a separate, list-oriented database system for such tasks.

* * *

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.