Companies fight back against Internet attacks

Last summer, the Yahoo Inc. message boards were full of postings that insinuated that Titan Corp.’s stock was headed south.

“Very, very bad earnings surprise coming today?” said one. “[Titan] is getting nailed with huge sell orders! Jump the sinking ship,” said another, posted by someone with the screen name “CCRibber.”

If the goal was to scare investors and drive the stock price even lower, it worked. Messages like that – plus a fake analyst report criticizing Titan – sent Titan’s shares plummeting from US$44 on June 20 to US$21 on Aug. 22. It was a staggering 50 per cent loss in market value, totalling US$1.3 billion.

Then Titan got mad.

The San Diego-based high-tech company filed suit on Aug. 30, angrily charging that the posters were “unscrupulous short sellers” who conspired to depress the stock for their own profits. The company got a subpoena to “smoke out” the people behind the three dozen screen names that had torpedoed Titan’s stock.

The case hasn’t wrapped up, but it’s yet another episode in which corporate reputations have taken a real beating from Internet messages, fake press releases and “gripe sites.” Of course, critical opinions are legally protected as free speech, but when the messages are false, defamatory or attempt to manipulate stock, companies are starting to fight back.

To do that, they’re hiring monitoring firms that use software to scan the Internet to find out what’s being said about business clients. They’re also hiring private investigators to track the perpetrators.

“We get requests for that all the time,” said J. Christopher Racich, director of high-tech investigations at Kroll Associates in Washington. “But it’s not something you want to do if you’re just aggravated [about the messages], because the investigation can be very expensive” – say, US$30,000 to US$40,000.

The investigations usually turn up former employees, disgruntled insiders or stock manipulators, Racich said. The big challenge is identifying the people behind the anonymous screen names. A flurry of messages may actually be the work of only one or two people who use different handles to make it look like they’re a crowd.

One approach is to file a “John Doe” lawsuit and use subpoena power to obtain the identity of the mischief maker from his or her Internet service provider. That’s what Titan is doing, but it’s a strategy that has to be used with caution, Racich warned. “It should be a serious lawsuit, based on a cost/benefit analysis, not just a fishing expedition.”

Another technique employs “forensic psycholinguists” – the same type of folks who analyse hate mail sent to politicians – who look for signs that the messages came from the same poison keyboard. Private eyes can also engage suspects in on-line conversations to seek clues about their identities, but there’s a danger that the undercover gumshoe could tip his/her hand or cross the line into entrapment, Racich said.

Michael D. Allison, CEO of Internet Crimes Group Inc. in Princeton, N.J., said there are even better investigation tricks.

For example, perpetrators may have left some electronic footprints behind by filling out a Web site guest book with the same cybersignature they use later for derogatory messages.

Sometimes the text of a message itself provides clues. “If they say it’s snowing outside, we’ll check [weather records] to find out where on the planet it’s snowing right now, to narrow the suspect pool. If they say they have a green Jaguar and live in Rhode Island, we’ll get a database that lists every green Jaguar owner in the state,” Allison said.

Apparently, companies are willing to go to great lengths to identify Internet content that besmirches their corporate reputation or infringes on their intellectual property.

The Associated Press news wire service in Washington hired Cyveillance Inc. in Arlington, Va., to identify Web sites making unauthorized use of copyrighted articles.

Cyveillance uses both human and artificial intelligence to monitor “on-line brand abuse” for clients. First, the company’s Web crawler looks for information that meets customer-defined criteria. Then a team of e-commerce analysts studies the automated reports and recommends a plan of action. The cost ranges from US$80,000 to US$400,000 per year.

Another company, Burlingame, Calif-based NetCurrents Inc., uses artificial-intelligence technology that scans Internet message traffic in order to provide a real-time graphical display of public sentiment about a company. Positive messages show up as a green bar, and negative ones as a red bar. But NetCurrents may not stop there. Theoretically, the technology could be used to gauge customer reaction to a new product or voter reaction to a presidential debate.

“You get an ongoing perception meter,” said Irwin Meyer, chairman and CEO of NetCurrents. He’s now testing the technology on Internet discussions about TV programs, in which fans of particular shows chat about the pros and cons of plot lines, actors and even clothing.

“The key is that it’s real time,” Meyer said. “We could deliver a report the following morning, in time to alter scripts or kill off a character that nobody likes.”