Saving Chanel, but losing the web

Much to the dismay of pretty much everyone on the web, a federal judge decided yesterday to allow luxury brand Chanel to seize and shutter 228 web domains that it argued offered counterfeit versions of its products.

The decision, and the process by which it was made, are ridiculous. A judge based in Nevada issued a death sentence for these sites without so much as a hearing. And none of these sites was given a chance to make its case before the sentence was carried out.

In an even more troubling move, the judge has ordered Google to remove these sites from its index. Search is the gateway to the Internet – the modern day access to information. The government telling a search engine which sites it may include in its index sounds like censorship to most people.

But a bad decision based on a bad process doesn’t mean that there’s not a real underlying problem here. Anytime a user can go to a search engine, search for a Chanel product and end up getting a knock-off instead, that’s a problem. And not just an annoyance. It’s a full-fledged security problem.

Clever search engine optimization professionals are experts at positioning sites of dubious quality on the first page of results on a search engine. There is no branding that allows sites of high quality to distinguish themselves at a casual glance from those of content farms, spammers or knock-off sites. Indeed, it’s the opposite. These sites are specifically designed to look legitimate. So consumers can be completely unaware they’re not on a company’s official site.

Ignoring First Amendment issues for a moment, having a federal judge curating a multibillion-page search index is a little like putting a finger in the hole of the dam without realizing there are 100 more holes draining the same lake. The web did not become a littered cesspool overnight, and it won’t be cleansed in one day by one judge issuing one arbitrary ruling.

A number of technology companies are already at work on tools and services that could solve this problem. My own company, Blekko, for example, uses human experts to curate the web and tell us which sites are shady. Our curators cull the results across hundreds of vertical topics and surface the results they think offer the best information.

Then, of course, there’s Facebook, which is itself a form of curation and one of the reasons many people think the social network will be such a huge threat in search. Facebook users don’t “like” sites that are spam, so user curation has the potential to create an incredibly high-quality index.

And now Microsoft’s Bing admits it is using its own editors to curate results. The search company has said its editors eliminated many holiday shopping and information sites because content on the sites is too thin. Google this year issued another update to its algorithm in an effort to improve its machine’s ability to recognize and kill spam.

Another approach to the problem is the new peer-to-peer search model developed by YaCy. The peer-to-peer network approach is interesting as it uses your browser history — a form of curation — as its index. Only sites that users have actually visited are returned in search results.

And there’s Quora, which skips the middleman search engine altogether and asks users directly for answers to specific questions. No doubt, human-supplied answers are going to trump algorithm-served answers every time.

It’s time to face the facts: The Web is more and more the Wild West every day. Search spam is now one of the fastest growing inroads for malicious attacks. It’s time to start thinking about how to improve security in search. If we need spam filters to protect us from malicious email attacks, then we also need something to protect us from getting ripped off by phony web sites or having our computers trashed because we click on a malicious link in search results.

Rich Skrenta is cofounder and CEO of search engine startup Blekko. A serial entrepreneur, Skrenta previously cofounded news site Topix and NewHoo, a crowdsourced web directory acquired by Netscape in 1998.

blog comments powered by Disqus