Researchers Untangle the Junk Web

[tjw61]

Researchers Untangle the Junk Web

Barry Levine, newsfactor.com 56 minutes ago

You search for something and click on one of the first returned links. But it's a junk Web page, filled with nothing but low-end ads that might or might not relate to your search.

It's more than an annoyance. It prevents users from finding what they need, and blocks valid businesses from getting their messages across. These sites are the virtual equivalent of billboards blocking your view from the highway when you're trying to find a gas station.

A new technical study released late last week by researchers from Microsoft and the University of California at Davis (UCD) indicate that there are only a handful of operators generating these pages, and that there might be technical methods to minimize their impact.

The complete study, by Yi-Min and Ming Ha from Microsoft Research and Yuan Niu and Hao Chen from UCD, will be presented in May at an international Web conference in Alberta, Canada.

'Doorway Pages'

According to the study, a small group of as few as three operators create most of the false "doorway pages" on the Internet to attract spiders from search engines, thus gaining higher rankings on search results. Search engines are continually changing tactics to avoid such search engine spam, but the spamming can still be effective. After clicking from a search engine to one of these sites, the user sees a page with lots of ad links and nothing else.

The creators of the doorway pages, according to the report, work with a few Web hosting companies and advertising systems. The authors indicate that their research shows most of the junk pages are served from just two Web hosting companies, and as many as 68 percent of the ads were from just three advertising syndicators.

The researchers found that search words such as "drugs" or "ring tone" returned results with up to 30 percent of the links being junk pages. The density of ad-only junk Web pages was 11 percent per 1,000 search keywords.

However, the researchers also found that only two blocks of IP addresses are used to attract search engines and deliver the spam ad content, which might provide a means to address the problem.

Blog-Hosting Services

"Ultimately, it is advertisers' money that is funding the search-spam industry," their paper states, "which is increasingly cluttering the Web with low-quality content and reducing web users' productivity."

They also report that some blog-hosting services are teeming with fake doorway pages. They found that blogspot.com, owned by Microsoft competitor Google, had more such pages than any other hosting service.

One of the hosting companies cited in the report, ISPrime in New York, said that the junk pages were traced to a single customer, who was violating the company's acceptable-use policy. ISPrime said that, as a result of this report, it has discontinued its relationship with the unnamed customer.

The paper, "Spam Double-Funnel: Connecting Web Spammers with Advertisers," and related information, is available at research.microsoft.com/SearchRanger/.