Blog spammers insert comments that contain a bunch of keywords and hyperlinks back to their sponsor’s site. In the old days, search engines would crawl across your page, and build indexes based on those relationships. So if a thousand blog posts included comments that said "toshiba laptop battery" and included a link to http://www.buymybatteries.com, then the search engine would display http://www.buymybatteries.com near the top of the results page anytime someone searched for "toshiba laptop battery."
That’s why spammers do what they do. They don’t really care about your page. They get paid based on the results they generate…the so-called "relevancy ranking" of their sponsor’s site in search results. So placing spam is just a means to an end.
When it became apparent that blogs were an easy avenue for unscrupulous sites to increase their rankings in this way, the search engines got smart. Google first announced that they would ignore links that included a hidden attribute called rel="nofollow". Others, like Bing and Yahoo!, followed suit, and these days, pretty much all search engines employ this strategy. The spam comments are still indexed along with all the other content on your page, but they won’t impact the keyword relevancy rankings in the least. In other words, when someone searches for "toshiba laptop batteries," then http://www.buymybatteries.com won’t appear any higher in the search results than any other website.
Blogging platforms and other social sites, such as Windows Live Spaces, now quietly slip that attribute into comments behind the scenes, so there’s no way for spammers to get around it. The expectation is that once the incentive to spam blogs dries up, the spammers will eventually go away (or at least find another way to pester us). It’s the digital equivalent of Chantix.
Does it work? Yes and no. For instance, when I searched for "toshiba laptop battery" on Google and Bing, our common Windows Live offenders are nowhere near the top 50. So, yes, rel="nofollow" is getting the job done on that end. However, it doesn’t seemed to have stopped spammers from trying to inundate our blogs with link-laden comments. They’re like addicts who keep taking a drug even when they don’t get high anymore. Search engines are doing just about all they can at this point to curb the spam problem.
FWIW, I think that it’s up to the Windows Live team to do some housekeeping to further curb the problem. They should be able to run a simple query on their comment database to identify repetitive comments and those with lots of links. It should be a simple matter to then disable their accounts. But as long as nobody complains, then WL doesn’t want to bother. My advice: complain. A lot.
- nofollow, Wikipedia
- Preventing comment spam, Official Google blog
6 Replies to “Countermeasures”
This is very interesting, thanks for sharing!
Hi Greg. For a long while I was reporting these spam activities to the Live Support team, providing the spammers profile link which clearly showed that 100% of their activity was focused on spam. Live Support in most cases didn’t reply or when they did they ased me for the exact same information as I had already provided. Clearly they don’t care or don’t know what to do about the problem. It’s very sad when a company as large as Microsoft can’t correct something as simple as this comment spamming problem. It seems to me this is very easy to correct.
It would seem that eventually boredom with the activity would cure most spammers. However they may not think like you and I. Very informative pos.
Boredom would be excellent, especially for the battery bloke! Who’s the other one, Greg?
Greg: Thanks for the excellent post. As you and many others know, I’ve been fighting against spam posts in Windows Live all this year. I kept track of reported spammers and their report numbers for a month, then checked to see how many Windows Live deleted. Obviously, spam in our blog is not a problem according to Windows Live admin. Reporting them seems to be futile if admin never closes their accounts. Also, they spam with one account, then get a new "No Name" account to spam with the next time. I’m very happy to read such a great post about the problem. X