Monday, September 13, 2010

Elsevier SciVerse: A Unified Search API Meets the Era of Apps

Twenty years ago, most professionals trading financial securities had to deal with a rat's nest of information services from competing information vendors stacked on their desks, in addition to dozens of phone lines used to chat with trading counterparties. Then along came applications programming interfaces on computers that allowed software developers to combine content from many different sources into powerful analytical applications.

It didn't take long for many major information companies to start buying and building platforms that could deliver content via APIs from their own databases and other sources into the software that their clients wanted. The result was a universe of powerful applications built by these publishers, their clients and third party developers that revolutionized productivity in financial trading.

Flash forward to today, when major scientific and technical publishers are just beginning to enter an era in which APIs are becoming a marketing necessity. For most subscribers to SciTech journals and databases, saying "it's worth it" for expensive and poorly integrated intellectual property is no longer an accepted response to their productivity needs. As in earlier years in the finance industry, market pressures require engineers and researchers to deliver profitable products more quickly than ever, placing more emphasis in information budgets on productivity tools than on information subscriptions.

It's in the light of these needs that one has to both praise Elsevier's announced SciVerse unified search platform and API and to ask the obvious question: what took you so long? SciVerse is a new initiative by Elsevier to provide unified searching of their Scopus, Scirius and Science Direct databases and their beta release of SciTopics edited research summaries. Searching is available either direcly via an Elsevier-designed SciVerse Hub search interface or via other applications consuming SciVerse-organized content using a SciVerse API. Initially the beta version of the API is available only for Elsevier's own Science Direct and Scopus content with the API being opened up to curated Web content available via Scirius as well as SciTopics content.

The demonstration of the initial SciVerse Hub search interface shows a competent search interface with a unified login to Elsevier services that demonstrates typical best practices for federated searching of multiple document databases. Unlike Elsevier's Illumin8 interface into similar content, SciVerse Hub does not try to apply a natural language layer to its "white box" search interface to try to organize content into different semantic groupings. Instead, the relatively traditional center column of search results in SciVerse is complemented by sidebar applications that provide more general but powerful post-search refinement tools.

The SciVerse Hub "Matching Sentences" tool, for example, developed by Elsevier partner NextBio, gives a highly visible count of sentences and paragraphs in matching document s that use terms related to the search query. This is very handy for getting a general idea of not just document counts but the general density of matches in specific documents. A click-through will provide a more detailed listing for further analysis. The SciVerse Hub "Most Prolific Authors" tool provides a bar chart listing of authors producing articles on the search topic listed in Scopus and how many documents that they produced match the search term. A NextBio "Methods" search, available beneath the main search box on SciVerse Hub, provides the ability to retrieve Elsevier documents in a pop-up window from a NextBio search based on general scientific topics.

As a search platform, SciVerse provides Elsevier a competent response to subscribers who have been frustrated by having to choose between Elsevier's powerful but unintegrated databases to perform searches for scientific and technical research. Finally all of the resources that Elsevier has to offer are available in one place, with useful tools to help subscribers to filter down to the most relevant documents. Integrating author content from Scopus helps to pull together views of content from multiple sources relating to those authors in powerful new ways, a big plus for researchers trying to understand how relevant and valuable the insights of specific researchers may be on a particular topic. SciVerse is not as powerful as their newly acquired Collexis platform for author analysis, but no doubt the integration of Collexis with SciVerse APIs will provide more powerful results in the future. The Methods tool integration, while appearing to be rather an afterthought in this version of SciTech Hub, is likely to provide more value in the future as part of other more sophisticated applications.

As a "day one" offering, SciVerse compares favorably to other SciTech search platforms, with much more to unfold as applications developers begin to use the SciTech API to build out additional delivery channels for Elsevier content. Elsevier claims to be able to reach more than 15 million scientific professional via their content subscriptions, so certainly there's a wide potential market for applications developers to consider. More importantly, there are already dozens of applications being developed for scientific research and engineering that could benefit from better integration of Elsevier content. For example, professionals using products such as Innovation Tools' Goldfire platform to define opportunities to apply scientific innovations to product development could get more access to more relevant content on research to make more effective product decisions more quickly.

While SciVerse will benefit in time from the ability to integrate client-supplied content and more content from competitive publishers, it represents a good start for Elsevier's first strong venturing away from product-centric approaches to content delivery and towards a wider array of client-centric solutions. The complete integration of Elsevier's own content resources will certainly help to position Elsevier more strongly during renewal negotations with clients who may have otherwise decided to choose one database from Elsevier over another. The power of SciVerse integration is in this sense a defensive move by Elsevier to ensure revenue retention for starters. But as more developers of productivity applications for researchers and engineers begin to make use of SciVerse's APIs, Elsevier can develop more ways to put their content into more valuable contexts far more quickly than they could via a handful of their own applications. Over time, they will, of course, have the opportunity to acquire those applications that are the best fit for their overall strategy.

It's product R&D on the cheap with little downside in the short run and with great potential to expand Elsevier revenues in the long run as its revenues turn away inevitably from intellectual property licensing and towards productivity enhancement. Will this mean a stronger and better Elsevier in time? It may, but it's also likely that a broader array of content sources than those used today in SciVerse will prove to be the foundation of future profitability in SciTech publishing. SciVerse begins to put in place the right kinds of plumbing to develop new generations of valuable Elsevier content services. What comes through those pipes will matter as much as the plumbing, though probably in ways that many scientific and technical publishers are not yet ready to embrace. In the meantime, though SciVerse may be a bit behind the curve when compared to financial information platforms, it's a welcome step forward for scientific and technical publishiing.
Post a Comment