I recently read a post from Mike on Jim’s site Microblogger that shows some pretty irrefutable proof that blog carnivals are dead.  Or, at least that using them for link building can have the opposite effect that you might like to have on your site.  In the article, Mike goes through the work he did, and explains how he then used the Google Disavow tool to disavow some links.  Most of those links were from blog carnivals and blog festivals.  The results he gets are pretty spectacular, really.  He’s got some nice graphs to show it, but I won’t be stealing those, so click on the link above to see them.

Hurting Your Friends with Google’s Disavow Tool

One of the first things that popped into my head when I read the article on Microblogger was that I had hosted a few blog carnivals on Beating Broke.  And what would happen if, instead of first requesting the links be removed, someone that had been included in those carnivals just decided to go straight to the disavow tool and disavow the link?  Would it negatively effect my site?  What if I followed the advice of the article and asked for links to be removed and then disavowed the rest?  Would I be negatively effecting the owners of those sites?

In short; would I be hurting my blogging friends by disavowing links from carnivals they had hosted on their sites?

Google’s Disavow Tool and Negative SEO

As is usual when I have a question about something like that, I go straight to Google and start doing some research.  What I found was somewhat reassuring.  Eric Ward wrote for Search Engine Watch about asking Google specifically whether disavowing a competitor’s links would hurt the competitor’s search rankings.  He details the answer he got, but the general gist of it was that using disavow would not hurt the competitor’s search rankings.  Eric goes on to give a little more detail and explanation as to his understanding of the answer.

The longer I think about it, the more reasonable it seems that using the disavow tool would not cause the domain of the disavowed link any real harm.  After all, think of what that would open up as far as black hat negative SEO options!  It’d be no time at all before the disavow tool would be devalued by all the excess noise and wouldn’t present any kind of reasonable signal for the algorithm.

Disavow Tool as an IQ Test

While I was reading more on the subject, I also stumbled upon this post from Danny Sullivan where he has a Q&A with Google’s Matt Cutts.  Matt is the head of the web spam team at Google.  In the Q&A, Danny specifically asks if the disavow tool could be used as a negative SEO tool.  The answer he gave is somewhat interesting, and gave me pause.

Right now, we’re using this data in the normal straightforward way, e.g. for reconsideration requests. We haven’t decided whether we’ll look at this data more broadly. Even if we did, we have plenty of other ways of determining bad sites, and we have plenty of other ways of assessing that sites are actually good.

We may do spot checks, but we’re not planning anything more broadly with this data right now. If a webmaster wants to shoot themselves in the foot and disavow high-quality links, that’s sort of like an IQ test and indicates that we wouldn’t want to give that webmaster’s disavowed links much weight anyway. It’s certainly not a scalable way to hurt another site, since you’d have to build a good site, then build up good links, then disavow those good links. Blackhats are normally lazy and don’t even get to the “build a good site” stage. 🙂

As usual, Matt is somewhat vague when he’s giving his answers.  But, I found it of note that he doesn’t directly slam the door on the disavow files to be used as a part of the algorithm.  The second part of the answer is even more interesting.  “If a webmaster wants to shoot themselves in the foot and disavow high-quality links, that’s sort of like an IQ test and indicates that we wouldn’t want to give that webmaster’s disavowed links much weight anyway.”  He goes on to call Blackhats lazy, which may or may not be true, but that’s besides the point.  What he’s really saying is that blackhats are unlikely to put all the effort into building up a good site, with good links, and then disavow those good links.  In that regard, he’s right.

But, we bloggers aren’t blackhats.  We certainly aren’t lazy.  We have put a lot of effort into building up our good sites, with good links.  Are we failing the “IQ test” by then disavowing some of those good links from carnivals?  The answer to that really depends on whether those links from carnivals are considered good links or not.

What is a good link?

I’m sure there are many indicators that Google looks at to determine what is a good link or not.  I also think we can probably look at a few indicators that can give us a pretty good guess as to what Google is going to think of that carnival link.

• What domain does it originate from?  If the domain is loaded up with terribly written posts that are stuffed with keywords, bad English, and not very helpful, chances are Google doesn’t think much of it.  On the other hand, if the domain is an informative site that’s well respected in the community, Google probably knows that too. Most blog carnivals are probably hosted on pretty decent domains.
• What is the link density of the page with the link?  How many links are on the page?  In the case of a carnival, there probably are no fewer than 50.  Some of the more popular carnivals receive hundreds.  Before, this might have just devalued the amount of “juice” you got from the link, but it may be penalized further based on the results that the Microblogger post showed.
• What is the link neighborhood like?  If there are 100 links in the carnival, are they all good links?  Chances are that they aren’t.  Chances are the host of the carnival didn’t take the time to fully vet and edit the links.  Most are probably O.K., but there are going to be quite a few more nefarious links in there too.  You’ve heard the saying that “one bad apple spoils the whole lot”, and this is a case of one bad link spoiling the whole carnival.

There are many other indicators that Google is going to look at, but I think those three factors are the ones that are most relevant to the blog carnival discussion.

One other small note here; As time passes, the answers to all three of those indicators could change.  Blogs die every day.  What might have one time been a very upstanding, informative blog might get recycled by a spammer after the domain fails to get renewed.  That’s a good reason to continue to be vigilant about all of your incoming links, but especially ones where all those factors could make the carnival a very bad place to be linked in.

Disavow Carnivals or Not?

With all of that being said, should we disavow links from blog carnivals, or not?  From a strictly SEO perspective, I think the answer is a resounding yes.  Ask to have as many of them removed as you are able.  Those that aren’t removed, add to your disavow list.

I still think it’s a bit hazy from a blogger’s standpoint.  The disavow lists aren’t being used as indicators of domain quality.  Right now.  Cutts leaves that wide open for future use in his Q&A with Danny Sullivan (link above).  So, are we doing harm by adding quality sites (just not necessarily quality links) to our disavow lists?  Honestly, I think only Cutts and Google really know the right answer to this one.  Site SEO changes pretty rapidly.  At the moment, adding low quality links from high quality sites doesn’t appear to do any harm to the high quality site, so I think it’s probably fine to disavow links from those high quality sites that we think are low quality links (like carnival links).

Is Hosting a Carnival Just as Bad?

If links from carnivals pointing at your site is bad, what about being the site that hosts the carnivals?  Probably just as bad.  For all the reasons that having that link pointing to your site is bad, hosting those links is probably just as bad.  In fact, it may be even worse.  After all, having those links pointing to your site is fixable by having the links removed, or adding them to your disavow list.  If the links in a carnival that you hosted are disavowed, it (probably) doesn’t do anything negative to your site.  But, you can’t disavow your own pages…  The posts and pages you host on your site are directly tied to your site (obviously) and if there’s reason to disavow a link on one of your posts, chances are that it is being taken into account by Google for your search rankings.  Hosting bad content with bad links is always bad.

Removing Bad Links and Blog Carnivals

If you own/run a blog, you really should be using something to check your site for broken or redirected links.  Spammers like to pick up used domains that used to have a lot of quality links because there are sites out there that don’t check for broken or redirected links, and those links can then be redirected or repurposed for their spammy purposes.  I use a plugin for WordPress called Broken Link Checker.  It automatically goes through and checks for broken and redirected links and then gives me the ability to work through the queue and delete or fix those links.

With the devalued nature of Carnivals, it’s probably a good idea to just go ahead and delete any that you hosted.  Especially if they’re older, or if you didn’t take the time to edit and vet the links that were submitted to the carnival.  It’ll probably also cut down on the number of emails you get from carnival participants asking for their links to be removed.

Have you participated in blog carnivals in the past?  What steps are you going to be taking (or not)?