July and August were terrific months traffic wise, the best we have ever had over here at ArticleSnatch.com. It was a nice test of scale for the platform that we are running on. At the end of August we noticed something that other sites have seen – the dread Google Slap!
Google Slap
Toward the end of august we noticed a drastic reduction in the amount of traffic coming to the site from Google. Doing some research led us to see that we also were no longer ranking for several key factors. We sent out feelers to several people to see if anyone had any answers to this, of course no one had a concrete go fix A and you will result B. We heard everything from its due to interlinking to spam to buying links to just “regular algorithm changes” to read the webmaster forum, to linking to affiliate programs. We put into place a few of those things, but I still had the feeling that it wasn’t any of those, especially not the algorithm change – articlesnatch.com should rank for the term articlesnatch (rather than a million crappy your site is worth $5500352345235 sites).
After a month of trying to figure out exactly what happened and 2 reinclusion requests being submitted and processed (with very vague answers), making various changes including removing some hacked hidden links into another blog, I remembered Andy Beard and started chatting with him. Together we drummed up a plan that hopefully will remove the Google Slap and restore the ArticleSnatch back into the site. Below I will layout the plan and follow up later with what the results have been.
The Plan
- Expand the number of posts on a topic page to allow for easier finding of the articles
- Add rel=nofollow to all outbound links from articles (I know I can hear you groaning, but wait)
- Fix other redirect, canonical and indexing issues
- Improve the usability of the rss feeds
- Remove some obviously spam articles – like the bunch from a company that included numbers in each of their pen names – nice waste of 7,000 articles from 5,000 “authors” guys – Zapped. (Working on better system to catch it sooner)
- Put on a happy face! Just kidding
Changes unrelated but made at this time
- Moved js-kit.com comment script to the last line of the code, so if their servers are running slow or lagging, it shouldn’t block the page load.
- Removed another Google FriendConnect module that wasn’t doing much but taking up space
- Added related youtube videos to the bottom of each article, stick around and watch some – its pretty fun
- Improved tracking and monitoring for article spam
Of course there is a lot more to it than just that, but thats the meat and potatoes of it.
Rel=nofollow – probably one of the most debated subjects out there. Google says to build as if search engines weren’t in existance, if we did that, then everyone could freely link to other sites that they had monetary interests in and sell ad space, link space, keyword space and whatever else they wanted to. However Google also says to add nofollow to other sites you own, affiliate programs, paid links (which you shouldn’t have anyway), sites you don’t trust and a few other things as well. In June Google changed how they treat the “page rank” from passing to nofollow links, and how it affects PageRank Sculpting, so there is less of a reason to use it on internal links. Noindex might be what you want if you want something to not show up. David Leonhardt even ran a test to see which search engines did what with a nofollow.
Mixx recently got filtered (penalty) and decided to add rel=nofollow. Digg took some protective measures before being penalized and Digg has been commended by Matt Cutts for their action, so thats a good sign.
Twitter has it on almost every link across the board, these sites all being affected shows that its a pretty serious penalty to be linking to sites you know nothing about. Twitter also removed our account in July as well and has yet to respond to any email requests, not bad for a $1 billion company.
From a SEO perspective some say that you should only build followed links, others say that you should not worry about it and build all kinds of links, and others say forget about the links, write good content and others will pick it up and reference it and the links will come naturally. The good content theory is probably the most sound, it worked for some major sites, however it really only worked after they had a following and people were happy to link to them. If you had a group of people as happy to link to your company as they are to link to wikipedia, then I’m sure you wouldn’t worry about links either.
So where do we stand, as of now we have added rel=nofollow on all links in author resource box. We have also stopped filtering out the links that are in the article body, so those links are now present, but nofollowed. Since you weren’t supposed to have submitted articles with links in the body as per our Terms, that’s just a bonus. Remember if an article gets no traffic, whats the difference how the link is? Also many articles end up getting “buried” in a sites architecture and rarely see the light of day, we are working on improving this as well. We think the increased visability some of our changes make will help get more viewers to read your articles.
Remember our editors are only human, we unfortunately do not have millions of visitors every day clicking thumbs up and down and saying this is lame, remove it, etc, although we do have a rating system, so feel free to use it. This change is NOT permanent. We are going to be reviewing authors and looking at the sites they link to. Those are writing quality articles and linking to quality sites, we will be removing the rel=nofollow from their links.
We are also looking at implementing the Google Safe Browsing API to check for bad links when the article is submitted.
We have already started reviewing authors based on what they are writing about, quality and the sites they are linking to. We will continue to do this as well in the upcoming weeks. Authors who have been approved will have the rel=nofollow removed from their resource_box links. We have also upped the removal of undesirable content, so if you have been submitting that (doubtful your reading this), and you login to your account and can’t find things, thats probably way.
Some other changes have taken place that at first we thought caused it, these included a few bugs with the new platform causing too much non changing content between pages, titles on pages all being the same and blocking robots from the wrong things. In late July we implemented a number of changes that improved the load time of the site as well as improvements to the usability of the site. Several of these changes had been long in the making and others were after speaking to some other site owners.
These changes included
- Implemented Google Ad Manager for improved ad targeting as well as faster ad serving.
- Added a feature for authors to select a default language that you would like to view the site in, one of 47 languages
- Moved translated pages to subdomains based on language instead of folders, added more caching to reduce the number of queries for translation, to improve page load time.
- Added a feature so authors can now submit articles in any of 47 languages and they will be machine translated back to english thanks to Google Translate
- Improved page build times, reduced load on server and some other technical things you probably don’t care about
- Added memcached to reduce the number of queries that mysql has to answer, a 75% reduction.
- Several other minor changes that allow for greater customization down the road
- Improved article statistic and logging
After over a month of work and several late nights trying to get these changes, they are ready to be pushed out to the live server. We will continue to fix things as they arrise and hopefully we can get out the google penalty box – this isn’t hockey!
Any questions, comments, feedback, suggestions? Please reply below.
UPDATE:
I think based on reading some of the comments that authors maybe confused about the nofollow policy, this is temporary until we can figure out what caused the penalty as well as who is linking to bad neighborhoods. After we get all of the bad neighbors out of here – and we removed several thousand of them yesterday alone – we will be adding things back slowly to trusted authors.
Also for those who have submitted articles across various topics, including some less than scrupulous ones, may I suggest that you go into your account and delete them. We recently removed all the articles from someone based on a few bad ones, come to find out she had 4400 articles. Now hopefully she will reply our email – but hasn’t yet. Either she can login and remove the bad stuff, or wait in line while an editor works on it, and then have the non bad ones reactivated.
For those who may have noticed pen names with a number in them and them gibberish – we’ve removed 6,000 of those articles as well. anyone see anything you think should be removed – please email us a link.
Update (10/2/2009):
We have found the first hidden link – I’m ashamed to admit that it exists, but now that I have found one, the hunt can begin to look for more. This person used some elements to hide the link, we found some new things to search for thanks to a post from 2007 on mattcutts blog.
In case you are wondering who was the first one we found — anandaapothecary – no I won’t be adding their .com here nor their link.
Another one found – Spencer Hunt – linking to his profile on EzineArticles.com – some “expert” he is.
Update (10/3/2009):
The first signs of some increased traffic from google have showed (thank you)
Update (10/4/2009):
Removed a lot of articles that were linking to online pharmacies, then went a step further and removed all articles by all authors who linked to an online pharmacy — were you affected? Want the legitimate articles added back – email us so an editor can take a look.
Update (10/5/2009):
Removed more articles from people with bouncing email addresses and duplicate submissions. If you notice that you have submitted an article more than once, login and delete the duplicate. If we notice, you will have all articles disabled until you fix the problem.
Comments