How is Google Detecting Link Bombing?

Posted by Miles Evans

Whoa first Wikipedia blows away all our authority links and now Google is tinkering with their algorithm to defuse link bombing. I for one have been wondering for some time why they didn’t hand edit these suckers out as I would guess many people assume Google itself thinks George Bush really is a miserable failure.

Matt Cutts official post on the Google blog :

“When we're faced with a bad search result or a relevance problem, our first instinct is to look for an automatic way to solve the problem instead of trying to fix a particular search by hand.�

So what about all of our fun SEO contests? Alright this isn’t really such a big deal as I doubt it will affect SEOs and we will dream up new contests. We can never forget that anything that taints G’s index is prime for correction. You really have to applaud the team at Google for making the change algorithmically rather than using lowly human editors. I guess the real question is ‘how on earth could this actually function?’ or 'will this effect my SEO work?'

Most of these ideas were brought up in Danny Sullivan's article when this story broke but I will list them here for posterity...

Detection. Some type of trigger must be happening when x% of all anchor text is identical for links pointing to a page. This simply has to be happening on some level so I will assume it is the means of detection. This can actually be tested as rankings fluctuate quite quickly so I imagine people will make some interesting finds in the coming months.

Link age. A link bomb campaign will normally have a huge surge of links while it’s trendy to link something up. Over time there will be less and less new incoming links. What’s more, we already know link age is a metric being actively recorded at Google so I would be surprised if this is not a factor.

Malformed referring page. This could be metadata, title tag, keyword frequency, headers, and all of the other onsite stuff a properly built page should have. I guess with this point the impression is that a malformed page would be missing some or all of these items while a properly structured safe page would not.

Trust. Large trusted portals and brands are less likely to engage in someones link bombing endeavor. Perhaps authority page links are compared to the rabble.

Phrase based indexing? Patent watcher Bill Slawski brought up this point. I am a bit fuzzy on this one but it stems from this Google patent. Paraphrasing: 'A related phrase bit vector for document URL0 eliminates the reliance of the search system on just the relationship of phrase A in URL0 pointing to URL1 as an indicator of significance or URL1 to the anchor text phrase.'. Huh? Yeah I cannot really get my head around this point but it has to do with determining relevency through phrase matching - I think.

Since all of this can be tested I would expect this stuff will surface again with some interesting data. We might even learn something new. In the mean time we will have to find other clever ways to bend the SERPs to our political agendas.

Check out the discussion at DigitalPoint and Cre8asite Forums for more.

Posted Jan 26, 2007 at 01:51 AM | Permalink | Trackback URL | Del.icio.us | DIGG!