Richard Clayton, a security researcher at Cambridge University, knows how to spot unsolicited junk mail – but even he had to look twice at what arrived in his inbox recently.

The mail invited him to put a link to bustem.com from his website. When he looked at bustem.com, he found a collection of security articles written by people he hadn’t heard of.

As an academic and author for the security blog lightbluetouchpaper.org, Clayton has an eye for good writing – especially security writing. What he found on bustem.com didn’t make him very happy. “I had a look at it and I wasn’t very impressed,” he says. “There were spelling mistakes, ridiculous bits of grammar, words missing from sentences and so forth.”

Web credibility

He traced bustem.org to Privila, an Illinois-based company that uses unpaid interns to produce large amounts of content that it then uses on its websites. The company, which expects a minimum commitment of 150 hours or 50 articles from each intern over a 15-week period, retains ownership of the content they produce.

Bustem.com isn’t the only site in Privila’s network, according to Clayton’s colleague Steven J Murdoch, whose research last month turned up a network of almost 300 sites connected to Privila – up from 153 in April.

Privila’s public mandate is to “preserve the value and credibility of information on the web”. Executives would not comment on the business model, and interns approached by the Guardian said they were bound by a non-disclosure agreement, but the websites it operates have something in common other than identical visual design: they all carry advertisements from either Google or Yahoo!.

Google’s advertising business has two main strands. Its Adwords service is aimed at companies who want their advertisements to appear on participating web sites.

The websites that display the ads are part of a programme called Adsense. Google’s website indexing system analyses each Adsense website and tries to display advertisements that are relevant to its content.

Every time a visitor to an Adsense website clicks on an Adsense advertisement, the owner of the site gets some of the money that the advertiser pays to Google.

The internet is replete with sites that post content optimised purely to drive ad traffic rather than having any value of its own. The internet marketing community even has a term for them: made for Adsense (MFA).

But the value of online content is a subjective issue. Who is to say what is an MFA site, and what isn’t? “There’s a fine line between clever search engine optimisation techniques and sites built purely to attract search engines,” muses Paul Moore of AdsBlackList.com, which operates an MFA blacklist.

Sites (not associated with Privila) that he calls MFA include mobiledownloads4u.com, which at one point was serving up nine Google ads, in a premium position above a single page of unattributed mobile phone-related article content. He highlights Myfirstnewsite.org as another, which contains content identical to the second paragraph of text at infobuy.info/?p=107.

Some MFA sites are obvious, says Clayton, because they’re generated by machines that create gibberish solely to fool the search engines’ indexers. But many MFA sites use content that is clearly written by people. Where do you draw the line? “Most of the articles are pretty inane, and generally you get the feeling that if you click on the ads, you’ll go somewhere better,” he says.

There is a market for the bland content used by many MFA sites, says Andy Beard, an internet marketer who maintains links with what he calls the search engine optimisation “blackhat” community. Beard has dabbled in the private label rights business – a community of writers and brokers who produce, rehash and distribute articles for customers who need to fill sites with content quickly.

“People got the idea that the content doesn’t have to be 100% unique. Someone could modify an article slightly so that it still appears as unique content to the search engines.”

Sites using this content have to make at least part of them unique, because they want to avoid the same content appearing on multiple sites and decreasing their search ranking, says Beard – but search engines can often be fooled by articles that are partly the same.

Such sites will spin single articles, sometimes producing hundreds of different versions of the same article. The process is so systematic that a market for automatically spinning software and services has arisen.

Spunwrite, for example, divides up customers’ articles into snippets that the service’s hired writers then rewrite into multiple versions. The matrix of snippets can then be reassembled into many different combinations.

Software such as WebSite Content Wizard professes to help article spinners do it themselves at home, and allows them to insert relevant keywords into their text to keep their Google rankings up.

Collections of articles are sold in packages, for as little as $1 per 500-word piece. Customers spin them into thousands of articles designed to draw traffic to sites laden with adverts or other profitable payloads such as email collection forms or credit card payment systems.

Would-be writers

But who writes the seed content in the first place? Canada-based Dave Markel uses sites like Rentacoder.com or elance.com to source original articles. Via these sites, he recruits would-be writers who want to turn a quick profit.

“I started paying between $5 and $7 per 500-word article. I was using the emerging markets – Pakistan, India and Romania. But you could tell that they were taught English as a second language,” he says. “So I started looking at the US, Canada, Britain, and Australia. Now I pay between $10 and $15 per article, but I don’t have to do nearly as much editing.”

Markel says that his writers are stay-at-home mothers, students and others wanting to write for money. The same applies to companies not in the private label rights business who are trying to create a more branded market for amateur content.

Suite101.com, for example, publishes articles from its 900 writers only on its own subdomains, rather than putting it on large numbers of apparently unrelated sites, as Privila does. It also pays its authors a commission on the Google ad revenue from their articles, and relinquishes online copyright to the author after a year.

But is the content any good? Editing and review is done mostly after the content is published online, rather than before, admits editor-in-chief and former Globe and Mail writer Joy Gugeler. “Stuff slips through, but we’re trying to tighten the net.”

Helium.com also pays writers a cut of advertising revenue. It asks readers to rank articles by reading two alongside each other and voting on the best one, says CEO Mark Ranalli. eHow focuses on step-by-step how-to articles, and contributors get paid according to their ranking by readers, says general manager Gregory Boudewijn.

These sites all manage their content quality with varying degrees of success. Clayton pokes fun at examples of naive content that he found on some of them; one of the steps in an eHow.com guide to working as a reporter abroad is “go to a country”. But the sites are at least trying to deliver financial rewards to writers rather than using them for free.

Background checking

Ben Edelman, an expert investigator of spyware affiliate networks, argues that the content used in many sites designed more explicitly for MFA purposes is far from useful for advertisers. “If I were Google, I wouldn’t have a difficult time deciding what to do here.

This content is not useful. The world would be better off if these pages didn’t exist,” he says. “The issue is where the money comes from – how it is that reasonably well-respected advertisers end up paying for this stuff?”

Yahoo! didn’t reply to the Guardian, but Google responds that it is constantly tweaking its algorithms and reviewing websites to keep up the quality of Adsense participants. That wasn’t enough for Adsblacklist.com’s Moore, who removed all sites under his control from the Adsense program last week.

He cites frustration with the number of MFA sites that he believes make it through Google’s filtering system. “Because it’s becoming so difficult to identify these sites, a lot of them are allowed into programs like Adsense without enough background checking,” he says.

The debate over whether the copy is there purely to fill the space between the ads has been raging in print publishing for years. Internet publishing has taken this to a new level.

Content’s eligibility as MFA ad fodder comes down to the quality of online writing, which sits in a spectrum from insightful reporting and well-crafted, thoughtful prose at one end through to machine-generated nonsense at the other. Both of those are recognisable for what they are. But for people like Clayton and Edelman, it’s the stuff in the middle that is causing consternation.