Skip to main content
Sep 30, 1995

Technofile: Death to Overload

Filtering the info-flow

Vacation season is over, at least for us Northern Hemispherians, and it's depression city. Back to deadlines, hassling bosses and unrealised ambitions. Back to reality.

You trudge into your all-too-familiar office, fire up your PC and click onto the e-mail. The heart rate accelerates dangerously. A billion or so e-mail messages, it seems, await.

Next you log on to your business online service, through which you've set up an automated electronic clipping feature. The heart rate now reaches near epileptic speeds. In your absence, a further billion articles have been flagged. They all scream: Read Me Now.

And it's not just after holidays or the two week roadshow that we're confronted with too much electronic information demanding our attention. It has become a fact of daily life in the digital age for managers up and down the organisational groove chart, as well as for the analysts, asset managers and professional investors IROs need to reach.

Electronic information technologies were supposed to liberate us from the mundane, empowering us to be better informed, more astute and more plugged into the big picture. But if your company has an internal computer network, you know the new plague affecting us.

It's called information overload, and it's where the much ballyhooed ideals of the networked company hit the fan. What was once an avalanche of memos and trade magazines building up in the in-box has become a deluge of electronic messages and clipped articles crowding our days and diverting us from tackling our real jobs.

Unchecked, information overload would only get worse as more of us go online directly or more of our employers install computer networks. By some estimates, around 70 per cent of business PCs in the US, 60 per cent in Europe, are now connected to some type of network. Within two years some 100 mn people worldwide are forecast to be linked to some type of corporate computer network, most of which will be tied into the Internet or commercial online services. That's a lot of e-mails and articles zinging around through cyberspace.

Fortunately, the end of information overload - or at least a taming of its most glaring excesses - might be at hand. A number of information companies, from giants like Dow Jones and Reuters to smaller players targeting niche markets, are already offering easy-to-use electronic tools that automatically select what you should be seeing. Other players, like Desktop Data and Sandpoint, have built powerful services that feed corporate networks with externally gathered business and research.

Now there's a new breed of information-filtering companies emerging, offering advanced software to weed out the urgent from the merely annoying.

Essentially, your company signs a deal with one of numerous vendors which, in turn, install some sort of software or write an interface to your network's server. This scans incoming feeds from, say, newswires and online services. At some point, you've either set up your personal profile, or employed a new type of filtered information service which has already created a special information product pertinent to your industry (telecoms or law, for example) or business interest (new products or results).

In either case, when a relevant story passes through the filter, bang, it's extracted from the flow and sent on to your desktop. Your colleague down the hall, meanwhile, gets a totally different set of stories.

'We're on the verge of a new era in the network environment, one in which companies rely on automated agents not only to control the flow of information, but to determine what's relevant and what's not,' says David Hoppmann, founder of Intell.X, an Arlington, Virginia-based company developing sophisticated business information services for desktop delivery. 'Everyone knows that electronic communications have given businesses a tremendous power to do more. But the obstacle that is all too apparent to management is the problem of the information glut - too much unfocused and irrelevant data burdening the guy at the end of the information path. This extends throughout the corporation, and impacts just about every professional category within it. Our mission is to curtail that overabundance.'

Intell.X, a newly launched division of the big US business online service DataTimes, will be employing artificial intelligence and neural reasoning techniques in the filtering process. Not only will the system pluck out just what is needed, it will continually refine and adapt the type of articles selected, through new associations learned over time by the neural network capability and based on new events within your industry field.

For example, a story about a new product launch by a competitor which represents a pivotal event in the cellular industry might not mention the word 'cellular'. The artificial intelligence technology will nevertheless detect and select that story.

And that's only the first step. Sometime before the end of the year, Hoppmann says, Intell.X will unveil a service to summarise automatically articles culled from news and information sources. For competitive reasons, Hoppmann is mum on the specifics of the summarising technology, saying only that it will be 'breathtaking'.

(Reader beware: Your scribe has a relationship with Intell.'Xs parent company, separate from this column gig. You'll just have to trust me that I would anyway have deemed this sufficiently newsworthy to merit your attention.)

While Intell.'Xs big marketing push is some months off, two interesting US-based news filtering services are already available from NewsWare and CompassWare.

NewsWatch and NewsLink, both marketed by the New York City-based NewsWare, consolidate different news and financial information services - ranging from Knight Ridder to S&P Marketscope to Zacks Earnings Estimates, among numerous others - into one consolidated pipeline feeding a LAN, and then applies software to filter that flow and alert any number of end-users to new, relevant stories. If your LAN should crash for a few hours, the NewsWare system will sense that and retransmit what you missed.

The system also archives and indexes articles and reports in a NewsWare data centre, allowing individual end-users to search the stored repository.

Newswire is a unit of equity and non-equity derivatives information provider Track Data and its CEO Mort Mackof says his company is the only one to offer consolidated real-time and historical news through a filtering mechanism. 'We tell you about news half-seconds after it's out,' he says. 'It's as fresh as possible. It's a real-time, instantaneous clipping service.'

CompassWare, also headquartered in New York but with its technical work based in Jerusalem, takes a different approach. It offers its Intelligent News Filter product to corporate LAN managers to filter different information feeds for relay to end-users. And a variation goes out to publishers and others who might want to add it to, say, a site on the World Wide Web to enable searchers to scan massive archives of stored materials more quickly. That latter application might be useful for Web sites containing huge volumes of technical literature, only a portion of which would be needed at any given time.

CompassWare CEO David Waxman stresses that his company is not a news management company per se, but a search engine developer. Those software engines are marketed under the name MagnetSearchTM.

Privately owned CompassWare is a so-called Alliance Development Partner for Dow Jones' Dow Vision service, customising that news source and others to work with your LAN's software configurations and its information distribution programs, such as Lotus Notes.

End-users on a CompassWare-installed corporate network set up profiles based on both keywords and concepts (which can be changed at any time), with its search engine then continually scanning the full texts of incoming stories and reports for matches. It is the combination of their own search engine with different news sources available that makes CompassWare unique, Waxman says.

How much does all this cost? For the three companies highlighted here, as well as others in the filtering business, fees depend on the number of end-users served and the number of news sources filtered and relayed. For a large company with, say, 10,000 end-users, monthly fees might be around $20,000.

If that sounds like a lot, consider the alternatives. 'It's no secret that many corporations have invested millions in their networks but haven't yet seen substantial pay-offs. The whole arena of advanced filtering is really the next frontier for computer networks,' Hoppmann says. 'As more and more of our business communications get digitised, we'll need the means to better connect the flow with our personal job requirements and individual working styles. For information to be a real resource for a corporation, individuals will need access to what's important specifically for them. Filtering will make that happen. Otherwise, companies will just bury themselves in useless data.'

Maybe next year's return from summer bliss might be a little less painful.