When Great Sites Get Penalized: The VandelayDesign.com Eventual Comeback Story
Editor’s Note: This is a guest post by Joe Sinkwitz (@CygnusSEO), a veteran SEO and Chief of Revenue for CopyPress.com.
Though I don’t have all the details, I’m pretty sure when Panda 4.0 rolled out, it wasn’t heralded as a fantastic day and cause for celebration by the team at VandelayDesign.com. Why? Simply put, *BANG*…stagger…stagger…*THUD*. They got hit hard.
If you’ve run enough sites for a long enough time, you’ve come across similar feelings, where for at least a fleeting moment you have the dull dagger of uncertainty thrust into your hopes and dreams. Getting slapped by Panda is no different of a feeling when you realize what happened, but the great news for Vandelay is they can have their Rocky Balboa moment, get back up, and continue fighting.
I know what you’re thinking; who is this guy? My name is Joe Sinkwitz, a guy from CopyPress that appreciates great design and great designers. How did this come about? Twitter!
So here I am. How did I conclude Panda hit the site? Aside from Steven being smart enough to self-diagnose, this was also telling:
From SEMrush chart above, you can see the drop-off occurring in May, corresponding at least loosely to the update timeline.
Auditing the Content
I have some rules when it comes to content, specific to how I like to help people address their issues. First is my rule for addressing content is to ask a simple question: does this page I’m looking meant to answer a question (search query) that has been already answered elsewhere on the site? It may seem like an oversimplification, but consider when a page about ‘logo design’ coexists with ‘best logo design’ and ‘inspiring logo design’ – ask yourself when creating if it is necessary.
So what happens when you find that you have multiple pages that all answer the same query? The best course of action is usually:
- Pick the deepest, most engaging page and improve upon it, attempting to work in some of the meatier content from the weaker pages that also answer the same query.
- For the weaker pages: are there any inbound links to the page from external sources? If so, 301. If not, 410.
- Update internal linking to reflect the new page as the desired answer.
That process again may seem very simple, but if you’re dealing with a fairly large site that’s been hit by Panda the mental exercise can be particularly enlightening. According to a site: command in Google, Vandelay has 3,460 pages indexed; are they really all authoritative? Some are fantastic, some are thin, and some like http://www.vandelaydesign.com/screenshots-psd/ are going to need to be fixed. Next piece of the puzzle, engagement.
Without going into too many details, I’ve played with engagement metrics in the past in order to bring a site out of Panda 1.0…of course in my overzealousness I was so aggressive that I pushed it into penguin 1.0 after the recovery, but let’s put that aside for now. I highlight the above URL as a prime example for something that needs to be fixed. If I’m a user looking for a walk through on editing screenshots in PSD, I might immediately bounce when I see all the images are broken. By solely fixing broken images and broken links, a site like Vandelay can go a long way towards improving on engagement metrics.
The other fun side effect is that because it provides for a greater user experience (especially amongst designers whom are going to be the toughest visual critics out there) sales should increase on the existing traffic. How can you find all these broken images? I free tool I recommend for projects under 250,000 pages is Xenu. At the end of the crawl, sort on the red highlighted URLs and begin fixing. It is also a decent tool for pointing out another big potential issue, duplicate content, though in Xenu’s case duplicated titles.
Duplicated content and its cousin keyword stuffing is still a problem in 2014, which may surprise you. As Panda has advanced it has evolved to take into consideration the full domain in the context of analysis and as such it needs to rely less and less upon keyword usage instances. If you were to use the word ‘theme’ on http://www.vandelaydesign.com/blog/category/wp-themes/ 61 times (which is currently the case) would the penalty hit if the rest of the site were relatively unstuffed? Most likely not.
Unfortunately, that’s not currently the case. Only 1 or 2 usages of the word ‘theme’ need to be mentioned for Google to understand that the page is self-ascribing itself as being important to themes. Be creative when it comes to copywriting; yes, you’re describing a WordPress theme but when using ‘it’ or removing the word entirely from the sentence doesn’t change the overall readability of the sentence, you’ll know you’ve gone too far. A great free tool (once you sign-up for free account) you can use for looking into KW density quickly is at SEObook.
Killing the Orphans
Alright, let’s now assume that you’ve collapsed the pages that don’t need to exist, improved existing pages for better engagement, and de-optimized on-page text to where the content is still very readable but less ‘stuffy’; time to find home for all the orphans!
Orphaned pages that is. Now, there are two definitions of orphaned pages; some consultants will say orphaned pages are those pages not linked to by either the host domain or external sources and some will say orphaned pages are those not linked to by external sources. I fall into the latter camp; not linking to pages from somewhere on your own site is just bad information architecture. How are we going to find these orphans? Hat tip a long time coming to the SEO Kyle Sanders who brought Ahrefs.com to my attention so many years ago, which has a neat feature called Top Pages. In the top pages report I sorted the 4,501 found pages in ahrefs in descending order by the number of referring domains linking to a page.
Unfortunately what we find is that only the top ~1500 or so pages have some sort of external link support. This is an opportunity though, as Vandelay does outreach to other prominent design blogs and forums, it’ll find that some of the support pages are candidates for consolidation, and the others once beefed up will make for great linking assets. Note, a lot of the SEO community doesn’t mention links as it pertains to Panda, however I bring it up again for the engagement theory: all things being equal, a search engine will want to curate to the top of the list pages that are “used” with time-on-site, bounce rates, shareability, and external linking votes as mathematically usable measurements of use.
Attention Matt Cutts & John Mueller
That’s certainly not the extent of what may have occured in May though. One huge, giant, enormous reason which I’m recommending Vandelay reach out to Google (JohnMu or Matt Cutts) directly over is the site migration. Site migration you ask? Oh yes. Whenever site owners are moving thousands of pages from one domain to another there exists the possibility that something will go wrong from a technical standpoint. As a bit of history for those of you who didn’t know (which included me up until about a day ago) is in January 2014 VandealayDesign.com (the awesome blog) and VandelayPremier.com (the place where you used to buy stuff) were merged, with the thousands of pages from VandelayPremier now existing over at VandelayDesign.com/shop.
Bigger Issues with Site Migration
I’ve seen and heard of several cases now where site migrations appear to be working, only to have the new site tank faster than the career of Lindsay Lohan 4-5 months into the transition as page references get confused on Google halts trust of the transfer. In this particular case, I’m thinking that might be very much what it is occuring and the root of many of the issues. After some digging it turned out that even after the migration over 500 of the newly moved pages were still referencing the old VandelayPremier section, which would then 301 back to the proper page on VandelayDesign — thankfully that will be fixed extremely soon. With a little luck from gogobot seeing the fixed pages and a quick doublecheck in Google WMT ensure that you let them know that you’ve moved, I foresee this issue getting resolved.
I hope this has been useful to you, dear reader of the Vandelay Design blog. Should you yourself find yourself facing down the gut kick feeling of a Panda update, let this serve as your guide on how to starting the recovery process.