Factiva Search 2.0 – An Image’s Mirror

Public search engines on the Web have taught computer users the value of search but not of good searching. Sites like Google and Yahoo! give too many off-target results to be usable as line-of-business tools. Only specialised information delivery products and services can provide the relevant, filtered and intelligent results that users can reliably act on with confidence.

Factiva’s new Search 2.0 service is an example of this sort of tool. It continuously analyses thousands of news feeds, Web pages and blogs, giving companies a focused and up-to-date picture of relevant industries, companies, trends and, as importantly, themselves.

What we see

Using the kind of concept clustering that is becoming the norm in top-end search systems, Search 2.0 extracts results based both on direct and associative relationships among terms. It presents these in a dashboard-style display. Here is a typical screen, showing the results of a search on “Adobe Systems”, which appears in a second or two.

Near the top, in red, the system asks if we mean “Adobe Systems Inc.”, which would narrow the search (to about 1,300 items at the time shown). For now, we’re happy casting a slightly broader net.

On the left appear the first news items, ranked by relevance. Ranking by time is an option. Clicking on a headline takes us through to the text of a result. On that page would be links to other, similar items about Adobe and to similar industry classifications.

Clicking a tab at the top shows us the results from Web-only sources. In these, clicking on a headline opens that Web page in a separate window.

The search extends back 90 days and explores, says Factiva, about 6,000 sources and 18 million documents. Repeated results are suppressed on this screen but not everyone will want this. Public relations companies, for instance, will wish to see every instance of a news item, no matter how invariant, to keep track of its distribution. Factiva Search 2.0 will offer selectable levels of ‘deduping’, from excluding only complete replicas of content down to suppressing half-similar articles.

On the right of the screen is the “discovery” area. In this are bar charts showing numerical summaries of various kinds of filtering. The topmost, blue histogram is arranged by time, each bar representing a week of the calendar. The graph shows the changes in the volume of mentions over the preceding 90 days. Clicking on any bar would take us through to the results for that week.

Below are four sets of horizontal histograms, dealing respectively with the main companies, topics, industries and sources that were mentioned in the articles retrieved. Clicking on any of these leads to articles with the emphases as labelled.

Between the two sets of histograms is a section called news clusters. This lists the main ‘concepts’ the system has identified in the first 100 news stories recovered. These inferred trends, relationships and patterns provide yet another set of routes to explore. The concepts often come at a right angle to the route originally selected. They can expose sometimes-unexpected commonalties with other companies and industries—hence the ‘discovery’ label for this area of the screen.

Email alerts can be set to tell users when fresh news about a topic arrives.

The value to users

The screen shown is just one of several available through Search 2.0 and typifies the results the service delivers and how it delivers them. The objective—which I think it achieves—is to present to a relatively unskilled user a quality of result that, normally, only a trained information analyst could achieve.

The results certainly differ from those from the main public search sites. For instance, an identical search of Web content on Google (back 3 months; English language only) produces 136 million results. Although 50,000 times as many as from Factiva, these results are undifferentiated. Running a similar search on Google News, which can go back only a month, produces 828 results. Sifting through even these would take impractical amounts of time and would cover narrower territory.

The results from Web searches by MSN and Yahoo! were of similar magnitudes. Yahoo! also has a beta version of a search of nine subscription services, including Wall Street Journal and FT.com (and Factiva and LexisNexis). This gave nearly 1,800 results. When deduped, this amounted to 16 entries in a relevance-ranked list. Only subscribers to the relevant sources could read the whole articles. MSN produced only 3 million or so results, reduced to 2.3 million after tweaking its ‘graphics equalizer’ results ranking controls.

Factiva’s range of products

Factiva’s main overall targets are the sales, marketing and public relations functions within user organizations and information specialists, such as librarians and researchers. Governance and compliance overseeing is an area of current growth.

The newest of its wide range of offerings is Factiva Insight: Reputation intelligence. Reputation or sentiment management is gaining in popularity as a way of monitoring an organization’s public image [1]. This product helps users with the task, finding occurrences of an article or news release, say, and reporting where and when they arise. Factiva Insight: Media Monitor Plus aids media benchmarking and monitoring.

Factiva’s SalesWorks product helps sales teams improve prospecting and lead qualification and sales call preparation. The iWorks product provides easy access to Factiva’s content for the general office worker. It integrates with Microsoft Office or can be reached by browser alone.

Some of Factiva’s newer products can work through Windows SharePoint Services and SharePoint Portal Server. Another new product works into IBM’s Unstructured Information Management Architecture, announced in August 2005. Software development kits are available to embed information search into services and application programs.

Behind all these offerings lies the Factiva.com service. This provides access to the public and to the ‘deep’ Web, the part that most free search services cannot reach (as the old Carling lager advertisement might have put it). This includes content from Dow Jones and Reuters and from subscription services, journals, periodicals and newswires. Factiva says it references more than 9,000 sources, in 22 languages. These come in over 700 document formats but arrive at the user’s screen in just one, saving much integration work.

Company outline

Factiva is unusual in being jointly owned by competing organizations. It was set up in 1999 by Reuters and Dow Jones. It is a private company, its formal title being Dow Jones Reuters Business Interactive LLC; Factiva is its trading name. The two founder companies continue to compete in real-time information supply, such as to financial dealing-rooms, and consumer-based publishing, such as in magazines and newspapers. All their business-to-business information goes via Factiva.

Clare Hart, Factiva’s President and CEO, says revenues in 2005 totalled nearly $280 million, an on-target increase of 4 per cent. Over the last five years, revenues have grown at compound rate of around 5 per cent. This year, turnover should grow by around 7 per cent. Although modest by computer industry standards, this is better than Factiva’s competitors are achieving, Ms Hart says. More than 90 per cent of its corporate customers renew their annual subscription.

Over half the company’s income comes from the Americas, with European business adding nearly 40 per cent. This income comes from Factiva’s 1.8 million paying subscribers and from its many licensing and OEM deals. Trading partners include IBM, Microsoft, LexisNexis, Inxight, FAST, Yahoo! and Salesforce.com. Factiva partners with Feedster and Biz360 for up-to-date feeds of the content of blogs and social networks on the Web.

There are 800 employees in the company, spread among 24 countries. Roughly a tenth of them work on product research and development, a proportion that will grow this year.

Summary

Factiva’s trump card is the quality and depth of the information it can incorporate in its products. It is supplementing this with a range of delivery methods and packages aimed at bringing text mining and visualisation to non-specialist users. If they are as well designed and effective as Search 2.0, the company should enjoy substantial success with them. Naturally, they would also need to be affordably priced and well supported. (A Factiva.com subscription, which contains the Search 2.0 functions, starts at around US$2,500 a month for 15 users. Enterprise pricing is based on the number of users as well as the organization’s information needs.)

Factiva’s main online competitors are Reed Elsevier’s LexisNexis (with whom it partners) and Thomson’s Dialog. Most smaller service suppliers, such as Infonic and WebClipping, look only at Web-based material.

There are some software suppliers offering similar tools, such as Insightful (with InFact), Ultra Knowledge (with Ultra Scout), IBM (with its Public Image Monitoring Solution, designed with Nstein Technologies and Factiva) and ClearForest. Some of these companies also compete with Factiva as general news searchers.

[1] To a lawyer, reputation management mainly means protecting intellectual property, such as patents and trade marks, as well as detecting passing off. In this article, we refer to the public relations activity.