Today, between meetings at the office, I had a brief IM chat with Jane, whose globetrotting ways have taken her to the UK for SMX London. Jane noted that our website - www.seomoz.org - was up for review on a search engine panel today, and received generally positive feedback from the engineers. One notable exception, however, was the fact that several engines commented that we were linking to a small handful of scuzzy, no-good, spamtacular sites (and that this was potentially hurting our performance in the results).
To be honest, it's not that surprising; we link out constantly to all sorts of stuff. Over the past 4 years of operation, we've linked to tens of thousands of sites, and it's very probable that some of those have changed hands or changed content and drifted off to the dark side. The frustrating part is figuring out who so we can update (or remove) those links.
Thankfully, there's a clever solution - Live.com's linkfromdomain search. Yes, technically, I could ask Nick and Ben to deliver a list of the lowest mozTrust domains we're linking out to according to the Linkscape index, but that's cheating (since we'd really be the only ones who could do it). But, the Live.com query can be used by anyone, and it's pretty effective. I started with some searches like:
- Linkfromdomain:seomoz.org viagra cialis (which produced a lot of legitimate sites, along with a couple that didn't look so good)
- Linkfromdomain:seomoz.org viaga (which gave a nice juicy one; view source to see the problem on tenthousandcents.com - anything targeting pharmaceutical misspellings is usually trouble)
- Linkfromdomain:seomoz.org porn xxx sex casino (gambling + adult almost always gives you spam)
Since it was recommended to us, it's probably wise for you to run some similar (and possibly more extensive) spot checks on your own external-pointing links. If you're fairly liberal with linking out, this might be an easy way to help make your site a better performer in the search engines and help fight spam. And, might I recommend to the engines' inclusion of an "are you sure you want to link to these guys?" tab in their Webmaster Tools to help site owners ID potential problems.
Quick question - what did the engineers in London use to find the links to the bad sites? Same method?
I suspect they were using their internal search engine tools (aka the Matt Cutts dashboard).
Hi Rand,
I was there on the panel and it was Nathan from Live.com who spotted the bad sites you were linking to. He used the upcoming version of live's webmaster central which at the moment is not available to us but should be rolling out pretty soon (I have in my head the date 4 weeks but that might be imagined?) so pretty soon all webmasters will be able to look this data up in live webmaster central.
Also - interestingly, when Nathan looked this up on the site review panel Chewy (who works for google) looked very interested, said it was a good idea and scribbled a note furiously so perhaps we'll be seeing it in Google webmaster central soon too!! ;-)
Another nice tool in MSN is to use the ip: command to find out if you share an IP address with any other sites that are promoting spam or bad neighborhoods. A typical shared host can store thousands of sites on the same IP, many ofwhich may be less than savory.
For example, there is a nice exploited blog on the same IP as Sphinn.com (64.62.209.13) that is pushing cialis among other items.
https://search.msn.com/results.aspx?q=ip%3A64.62.209.13+finasteride&go=&form=QBRE
That's a great idea!
Thanks for the tip, Rand.
I just went and reviewed one of the sites I am working on and found out that our #1 competition actually has our linkdomain stats in their web analytics.
I am really bugged out by the fact that they would use our links as a part of their web analytics. Somehow, I feel like that's an invasion of my client's privacy, even though I know the information is public.
Nice tip Rand! Thanks!
The IP operator at MSN got a mention in the buying sites for SEO session with DaveN (on top form btw)- still a useful tool! - you guys look clean as a whistle... Unlike... Me!
PS - Here are some more useful operators:
https://help.live.com/help.aspx?mkt=en-gb&project=wl_searchv1#op_ip
Maybe an alert service that alert us by email when this type of links become live in our sites would be a great service! This type of spam is a big head pain.
What I found running this linkdomain search was mostly spammy comments on blogs we linked to. Those were nofollowed as well, but it gives further pause to think that via---gra post comments were allowed into the blogs. I emailed a few folks suggesting deletion of those comments.
nice post Rand. This will help me to keep my website clean.
we were linking to a small handful of scuzzy, no-good, spamtacular sites
I had seen lot of spam links in the blog comments and used to wonder why these links were not removed. Now i guess the time has come to remove such spam links.
Discussing about spam sites for SEO purpose is now becoming harmful now a days.
Luckily, those are all nofollowed, so I don't think they're the problem. My best guess is that it's old blog posts and articles that linked out to places that have engaged in some negative practices (our weekly roundups, or articles where we've pointed to something without carefully checking it out, etc.).
Yes, i found 1 Mattcutts blog link from the 3rd linkfromdomain: example you had given above. :P
This is actually quite neat. The only thing missing is some kind of value judgement on the link itself, which I'm not sure that we can expect Live to supply (however I've heard about this Linkscape thing from some guys on the internet somewhere...).
Obviously Live is pretty well stuck to being third (fourth? fifth?) banana right now but in some ways being the smaller guy can be helpful. For example, I'm sure it's a bit easier for them to roll out innovations like this than it is for an enterprise the size of Google. I think that if they want to differentiate themselves it wouldn't hurt to keep inventing this sort of thing as opposed to buying up troubled CPC ad networks and hunting for mergers.
The addition of these type of functionality is long overdue on webmaster tools, but this is a good technique for the time being.
It would be nice if Google could offer us this information in Webmaster central. It's not easy for a webmaster to go back to every site and check which ones have changed hands or gone spammy. Especially with multiple sites. A small note in the console would really go a long way to helping us identify these sites when they become spam.
It is coming in the LIVE webmastertools very soon.
The fact that SEOmoz had just four scummy outgoing links is quite incredible considering the time it has been online. Looks like a lot of care is taken to clean up junk posts.
How bad can this hurt your rankings? Do you have to link to dozens of sites before there is a negative effect or is one site enough?