First there was Panda, and then there was Penguin, but is there now Ghost as well? Webmasters will immediately know what I'm talking about as far as Penguin and Panda are concerned: updates to the Google search results ranking algorithm designed to make the search experience more relevant for end users by lowering the rank of 'low quality' sites such as content farms and those who use black-hat SEO techniques. Ghost, however, is not only rather spooky but also remains unconfirmed by Google itself.
While anything that makes search better for the user has to be a good thing generally speaking, when you get down to the specifics of Panda there has been a lot of collateral damage to sites such as DaniWeb. There's not a lot of point covering old ground here in too much detail, save to say that a couple of years ago DaniWeb founder and CEO Dani Horowitz became something of a poster girl for the SEO world when she managed to recover from the Panda effect.
Google has already made it clear that Panda will no longer be pushed out manually, instead it will be integrated into the general Google algorithm and therefore become in effect a rolling update. That doesn't mean that sites which are primarily forum-based with mainly User Generated Content (UGC) such as DaniWeb, which have managed to rebuild themselves (in SEO terms) to remain Google friendly are in the clear. In fact, DaniWeb was hit hard by Google in November and is still to recover the lost traffic. At first it was assumed that this was a result of the Panda #22 update which hit on 21st November, but a closer examination of the analytics data reveals that actually DaniWeb traffic was decimated the day before on 20th November. This was, perhaps, no strike by the Kung-Fu Panda after all; maybe DaniWeb traffic was spirited away by the Google Ghost instead?
According to webmaster Brian Wozeniak, writing right here on DaniWeb a "bunch of forums and UGC type websites were hit hard" between November 16th and 18th. Brian continues "I believe you got affected by the same update that many of us UGC type websites got hit by. As far as I know Google has not publicly identified any update that took place during this time, but there are many of us who were hit on these dates before Panda #22 actually really occurred". Brian has a theory that 'medium quality' content was targeted by this unofficial update.
Medium quality content might be thought of as that which is still useful to readers, and isn't thin (scraped and non-original content in other words), but isn't exactly fantastic either. Medium content is par for the course at pretty much every site where the majority of content is user generated. Poor grammar and spelling mistakes from users whose first language isn't English might be considered medium quality in this context. Some forum sites actively moderate and correct (or delete) such posts, whereas DaniWeb doesn't. Penalising members for not speaking English as well as a native seems a tad harsh, especially when the actual meaning of the post is clear enough. Where the meaning isn't clear, then the community will ask for clarification and the thread will effectively be self-corrected. That's what a community of users is all about, and it works well at DaniWeb.
Sure, content that makes no sense or is just too hard to read can be problematical. Which is why a team of DaniWeb volunteer moderators spent many hours per day, for weeks on end, manually correcting no less than 80,000 posts made across a period of ten years which were incorrectly formatted using the old BBCode system that was replaced by a new Markdown system last year. The readability, and hence quality, of the DaniWeb archive has dramatically increased as a direct result. At least it has if you are human, and not an algorithm. Since March 2012, when DaniWeb moved to an entirely new platform that was coded from the ground up in-house, there has been just about no botspam, and a much reduced occurrence of any spam at all in fact.
Others have also noticed the Google Ghost effect, and a discussion is underway on the Google webmaster support forums. Of those taking part, perhaps the most telling is John Mueller, a 'Webmaster Trends Analyst' based at Google in Zurich, Switzerland who states: "One of the difficulties of running a great website that focuses on UGC is keeping the overall quality upright. Without some level of policing and evaluating the content, most sites are overrun by spam and low-quality content. This is less of a technical issue than a general quality one, and in my opinion, not something that's limited to Google's algorithms".
While not confirming, or denying for that matter, the existence of the Google Ghost update, Mueller does insist "making sure that all of the initially visible pages of your site are of the highest quality possible isn't something that a site-owner would do just for Google, it's really something that they'd probably want to do for all users" and goes on to hint that perhaps there is an algorithm policing UGC forums after all when he says that "putting your best foot forward there is something that - imo - isn't just done for the 'algorithm' but really primarily for your new users".
However, as Dani Horowitz points out "It's one thing to figure out how to recover from Panda (at least you know what you're dealing with) but it's a different animal altogether when you don't even know what you're up against". And while the Google Ghost continues to haunt UGC forums, without any actual proof that it exists, this could be impossible to counter. Even if sites such as DaniWeb manage to tweak things which, more by luck than judgement, manage to appease the Ghost algorithm there's no guarantee of a quick fix to the decreased traffic issues as Mueller confirms. "Significantly improving the quality of a website will take time" he says "on your side, as well as on our side when our algorithms work to understand the changes that you've made".
