With the increased use of compiled lists, primarily culled from Internet sources, marketers now have many options when building prospecting databases. The question now, however, is, how do you choose among the various list-compiler vendors and parse their relative strengths?

A new study attempts to shed some light on the subject. “Online Sources of B-to-B Data: A Comparative Analysis, 2010” is a follow-up to a study from a year ago, and reveals an evolution in how business data are collected and offered.

One finding this year dovetails with last year's study: While business data tend to be relatively accurate, the breadth of company contacts is spotty.

“The fundamental background for this study is the general lack of confidence business marketers have in publicly available prospecting files,” said Ruth P. Stevens, a customer acquisition and retention consultant, Columbia University business professor and co-author of the study.

“On the other hand, the large vendors of compiled data are probably as eager as we are to show what they can do, so that's probably why they've been willing to participate in this comparative analytics study,” Stevens said.

Compiled lists have come to dominate the database marketing conversion over the past year. According to marketing consultancy Winterberry Group, U.S. marketers will more than double their annual spending on online-derived data sources over the next two years, investing as much as $840 million by 2012 on database lists and information about digital audiences and online behaviors.

In fact, Winterberry projected that of the $8 billion U.S. marketers will spend on marketing data, $1.5 billion will be driven by such digital vendors as online data compilers and exchange platforms, representing the entirety of the growth through 2012 in this arena.

Stevens' study, conducted with Bernice Grossman, president of database consultancy DMRS Group, compared the performance of five list compilers that agreed to take part. They were asked to indicate how many companies they were able to list in 10 industries and their volume of contacts in those industries.

They also were asked to report on “complete” contacts—defined as full name, address, title, phone, fax and email—of 10 specified businesspeople each within separate industries. Those 10 served as the study's controls because their complete data was known to the researchers.

The list compilers volunteering to participate in the study were D&B Selectory, Demandbase, Infogroup, Jigsaw and NetProspex, with holding company Infogroup representing a number of list management companies and divisions.

One of the most striking results of the study was the variance in the volume of companies and names reported by the different list compilers.

In the stone, clay and glass products category, for example, D&B Selectory reported 28,630 companies, while Infogroup returned 26,853, Jigsaw listed 10,446, Demandbase found 4,114 and NetProspex offered up 852.

Count variance was similar for most other Standard Industrial Classification niches, such as chemicals (D&B Selectory with 33,852 companies and Jigsaw with 16,236); business services (D&B Selectory with 2,434,988 and Infogroup with 894,833); and communications (Jigsaw with 59,168 versus Demandbase reporting 6,072).

Count differences for actual contacts within well-known companies also varied widely. For example, when asked to provide contacts at Dell Inc., Jigsaw reported 7,061 people, while NetProspex offered up 2,409, Demandbase found 2,161, D&B Selectory returned 212 and Infogroup listed 199. Other contact counts within the same companies showed similar variances.

Compiled lists for use by direct marketers are assembled from a number of sources, including directories, contacts from trade shows, public records, social sites, credit reports and even by “scrubbing” business or other special-interest websites for information about business executives.

“We did find that data from these compilers was more accurate than expected, in terms of names, addresses, company name, ZIP codes, etc.” Stevens said. “But what we found is that there are a lot of business buyers whose records are not in these databases.

“So the way we're describing it is, the accuracy was pretty good, but the coverage was surprisingly weak,” she said, reflecting the inherent narrowness of data assembled from volunteered information, or from Web scrubbing which tends to produce names primarily of top executives.This may be a problem for marketers seeking volume.

“From an agency perspective, our customers have a voracious appetite for volumes of data,” said Kevin Kerner, managing director-U.S. for agency Mason Zimbler, Austin, Texas. “Quality is important, but many marketers are just thinking about the next name. It's the idea of "good enough' data to get them started.”

Jigsaw's model of collecting business contact data has captivated the industry, because it consists of “crowd sourced” information. In crowd sourcing, businesspeople voluntarily list and update their own information in exchange for others' data. The result is 18 million contacts from more than 3.5 million companies—all, presumably, of businesspeople interested in being found.

Other vendors, such as NetProspex and Demandbase, employ such techniques as DNS reverse look-up, where contacts who visit websites can be tracked back to their companies based on the organizations' Internet domain names. Like D&B, they also employ call centers to verify contact information.

“If you want scale, you'll go to D&B and Infogroup, and maybe Jigsaw,” Kerner said. “If you want more targeted data, based on behavior, NetProspex and Demandbase are available.”

One weakness of the compiled lists: With individual records, the researchers assumed that vendors would provide direct phone numbers, but many provided only the general company number.

“The service was to look at the data the way b2b marketers generally look at it, about how to get complete contact data, who's out there and what does it look like in a comparative fashion,” Grossman said.

To increase the likelihood that marketers get the data they want, the researchers advised them to develop a detailed list-ordering methodology. They also urged marketers to understand what vendors mean by “complete” information, a definition that can vary; to be specific about industry selections; to watch for vendor specialization by industry; and to choose between breadth of companies, or breadth of contacts or both.