Value for the People
How does helping people quickly identify required information help content creators? One answer comes from IDEAL (International Digital Electronic Access Library), the journal licensing arm of Academic Press. An analysis of usage data proved to IDEAL that libraries saved in the neighborhood of $12 billion by accessing electronic versions of articles from journals to which they didn't subscribe rather than obtaining them by interlibrary loan. Even Dr. Pieter Bolman, Academic Press president, acknowledges this is an overestimate, since not all of the downloaded material would have been interlibrary loaned. Still, by providing a source of high-quality academic and scientific journals to the library community, IDEAL provides the service of partitioning off a type of information so that end-users do not confuse science and pseudo-science. Results from Web search engines tend to look alike. IDEAL pre-qualifies the data.
There are other attributes that contribute to value-add. From the early days of online information, adding value by indexing documents was a primary distinguishing mark among data sources. Information professionals choose to pay their money to producers with a reputation for excellent indexing. When Dick Harris, now a consultant to the information industry, created the Business & Industry database, his stated aim was to position the file as unique due to its carefully chosen, and even more carefully applied, contextual indexing terms. It's not always easy, when confronted with un-indexed text to identify articles that are about market share or market size. Similarly, advertising slogans may not be so labeled within text that concentrates on another portion of a product launch. Business & Industry's breakthrough was analyzing the content of an article and assigning concept terms (such as downsizing, industry forecasts, and trends), marketing terms (such as cobranding, global campaign, and infomercials), company names, geographic areas, and industry/product names and codes.
Another value-add was title enhancement. Harris realized that busy people frequently don't have time to read an entire article-they want a quick summary. So he had his indexers summarize the content of the piece in one sentence. This intellectual effort was rewarded as information professionals began to prefer this new database to older, more established products. One particular telemarketing organization chose Business & Industry solely on the basis of the enhanced title, recalls Harris. "Summaries beat full text in telemarketing because people are measured by the phone calls they make, not what they read." Harris adds that his company, Responsible Database Services, reached profitability in a few short years. (The company is now owned by The Thomson Corporation and has been merged into The Gale Group.)
Similarly, corporate directory producers, such as Dun & Bradstreet, Kompass, and Hoppenstedt rely on industry codes to sort out the differences among how people describe an industry. Although each uses a different coding system, the underlying philosophy is the same. Substitute a numeric symbol for words that may be ill-suited to econtent retrieval. Other value-add could include hyperlinks to related information or author contact information.
Costing Out Value-Add
Two drawbacks here: the first is search engines. To effectively get the value out of the value-added features, the search engine must recognize that they are there. For highly structured, complex databases, a template that accommodates the structure enhances the value-add. In some instances, it enables the value-add because without a search engine that reveals the nature of the value-add, that value is wasted-it's actually invisible. The cost of hiring indexers is squandered. The intellectual effort of analyzing data is thrown away. Only when a template is custom designed, such as Alacra, formerly Data Downlink, is the value realized.
The second issue is the cost detailed earlier. Put simply, value-add doesn't happen without money being expended. However, where the money is put is changing. There are a number of companies building their business models based on eliminating the human indexers who add value. By automating the indexing process, companies can save money, yet still provide the added value that encourages both information professionals and novice information users to pay for data. Quiver, for example, uses a contextual algorithm and combines manual with automated solutions. Although Quiver's QKS Classifier is aimed at internal documents, the rationale behind the product is applicable to any automation of the value-add development. Andrew Feit, Quiver's EVP for sales and marketing, thinks that human judgment is needed for accuracy, but the technology side of things adds efficiency. His solution is to run documents through a categorization engine, but have people in a feedback loop to catch any anomalies, inconsistencies, or inaccuracies emanating from the automation.