Lucid Marketing Blog – Plain & Simple Marketing

Most of the time when people worry that they’ve been removed, or deleted from the Google index, they haven’t. It’s just that their rankings are too low.

But this was different.

I really had been removed from the Google index!

How did I know? Because I did a site:www.mywebsite.com search and I was told that my search did not match any documents.

I confirmed this by checking my Google Webmasters account and, sure enough, it confirmed that, for some unspecified reason, Google were no longer indexing my website.

Time to panic?

Of course not!

I’ve been listing sites in Google since 2001 without any problems so it was unlikely that something in my standard SEO practice was fundamentally flawed. And besides, I’m one of the Lucid SEO team so I’m supposed to know what I’m doing

I calmly considered the problem.

But, to be fair, this was not so much down to confidence as it was to the fact that it was a new website that hadn’t been in the index for long anyway. It was off to a bad start but, since it wasn’t getting traffic from the search engines yet anyway, I wasn’t missing out on much.

Not yet anyway.

I knew obvious violations such as cloaked text (putting keywords on my pages that are the same colour as the background) or pages (showing Google one thing, and human visitors another), doing mass scale reciprocal linking with no thought to theme or quality, were among the more common reasons for getting booted out…

But I wasn’t doing anything like that.

A quick scan of Google’s, “Webmaster Guidelines” refers to programs that automatically submit your site to the search engines, or check rankings.

I use SEO Elite, but have been doing so for years without any problems and, of course, I have my Google API Key (back when Google was still handing those out). So it’s unlikely to be that.

There’s always duplicate content of course, but all my content was brand, spanking new, and most of the articles were long, so that also made a template issue seem unlikely.

And I certainly wasn’t doing anything with viruses, trojans, doorway pages, etc.

Design, content and technical guidelines are a bit more ambiguous. I mean we all know a “useful, information-rich site” when we see one, but it’s still somewhat subjective.

So what then?

Well let me tell you the setup and see if YOU can figure it out.

The website in questions was a wordpress blog – ThePayrollBlog.com. Newly built, customised and setup with a nice variety of plug-ins. For about two to three weeks, new content was being added at the rate of 5-10 new articles per week. Some were press releases that may appear on other websites, but these made up the minority of pages.

At this stage I had no external links pointing to the website as I wanted to make sure everything was running smoothly before I let the search engines in.

Everything seemed to be going well, so I added my first external link from the menu of another website that I own – PayPerShop.com. PayPerShop has been online since 1997 and Google currently indexes over 15,000 pages from this site, and send approximately 40,000-45,000 visits per month.

It’s also closely related thematically, so it seemed like a good place to start.

Google crawls, on average, about 900 pages a day from PayPerShop so it wasn’t long before ThePayrollBlog began appearing in Google. Obviously it wasn’t going to rank for anything significant yet, but I had to start somewhere and the main thing is that it was in the right place.

Until it disappeared.

Any ideas?

The best I could come up with, was the link in the PayPerShop menu. The menu is, of course, on every PayPerShop page so that meant that, overnight, ThePayrollBlog had THOUSANDS of incoming links.

It was certainly a possibility.

A new site getting so many links so quickly, even if they were all from one website, is not exactly normal.

Could this be the reason why Google had shown ThePayrollBlog the door?

It would be difficult to be certain, but it was a good place to being. I immediately added the “nofollow” attribute to the menu link and added a single text link from the PayPerShop homepage. That way, I kept a link in place from a good spot, but restricted it to just one.

I submitted a re-inclusion request to Google (via the Webmasters control panel), explaining what I thought the problem might be and how I had fixed it. The instructions on the submission page said it could take weeks to get to my request and they wouldn’t be able to enter into personal correspondence.

I’m not good at waiting, so I logged into Stompernet and asked the resident experts for some advice.

Dave Taylor and Don Schnure weighed in first, both suggesting that the problem was one of over-optimisation. For example, using the same keyword too many times on the same page. Don also suggested that the way the two sites were linked together could be an issue.

Then Dan Thies jumped in with a short, to the point question:

“You are concerned about the link from Paypershop to ThePayrollblog, but not the massive number of links going the other way?”

Hmm… That got the cogs turning.

I had quickly thought about the site-wide link I had unthinkingly placed on PayPerShop, but I had forgotten that there was a link to PayPerShop in the author signature at ThePayrollBlog. The author signature appeared at the end of almost every post on ThePayrollBlog, which meant a handful of links on the homepage, and at least one link on every other page!

That’s a lot of links.

Does that mean Google was working on the possibility that I had set ThePayrollBlog up, purely to build links to PayPerShop?

That was a very real possibility. It certainly wasn’t the ACTUAL reason why I had created the blog; the relevance of the material meant that a lot of reciprocal links were inevitable. But it certainly wasn’t sensible to structure a new site in that way.

And then Dan, made another simple, but profound statement:

“The first place to look when a site is removed, and notified, is the site that was removed and notified.”

Obvious, right?

Well hindsight is always 20/20, but there’s a good reason why I missed it.

Creating a site-wide link to a new website isn’t sensible because it creates a large number of links in a very short space of time. Which is why I thought the site-wide link from PayPerShop to ThePayrollBlog might have been the problem.

But I wouldn’t generally worry about a site-wide link to PayPerShop because the website is so well-established, and already has lots and lots of incoming links. What would be a big jump in the number of incoming links to a new website, would only be a slight increase for PayPerShop.

And that line of reasoning had caused me to forget the simple truth that it was ThePayrollBlog that had been removed and that the most likely cause of the problem was THERE.

Think about it – if it was that easy to kill a new site, then you could hurt your newest competition simply by setting up a site-wide link from another website. That doesn’t sound very likely.

But by putting so many links to PayPerShop on ThePayrollBlog, I had effectively added a site-wide link ON THE NEW SITE ITSELF.

The last thing you want to do with a new website is do ANYTHING that looks like you’re trying to do something shifty with your SEO. You might have perfectly honourable intentions, but the Google algorithm isn’t a judge of character.

It sees in 1’s and 0’s like every other program.

Now that the penny had dropped, I went back to ThePayrollBlog and studied it carefully. And I found a number of items that MIGHT appear out of the ordinary.

Here is a complete list, along with steps I took to fix them:

1) As already mentioned, I tidied up the site-wide link from PayPerShop. The site-wide is still there, but all but one of them has the nofollow attribute.

2) I changed the author signature on ThePayrollBlog and nofollow’d some of the links to PayPerShop.com. I didn’t remove or tweak all of them, but I took care of the majority.

3) If you don’t manually insert keywords for the META Keywords tag on a Wordpress blog, then it automatically inserts a list of your categories. As I had a large number of categories, the META Keywords tag was a mess and would certainly qualify as “keyword stuffing”. Google say they don’t read the META Keywords tag, but I wasn’t taking any chances and cleaned it up anyway.

4) I removed a couple of widgets that were creating a lot of internal duplicate links. Each post is already linked to from the title and footer of each post, so I didn’t really need more.

5) I also reduced the “See Also” widget to show three related posts, instead of five. Similar thinking to #4.

Two days later, ThePayrollBlog reappeared in the Google index.

About 10-14 days later, the warning message in Google Webmasters disappeared.

About two weeks after that, ThePayrollBlog ranked #2 and #3 in Google for the search term “payroll blog” (without the quotes). My other website, PayPerShop, is #1 for the same term.

The next job will be to get more incoming links from a variety of relevant websites and try and rank for some keywords and phrases that people are actually searching on.

But I’m a little more relaxed now

In conclusion, I’d like to thank the following people for their assistance:

Dave Taylor: www.askdavetaylor.com

Don Schnure: Facebook – Don Schnure

Dan Thies: www.seofaststart.com/

And, if you’ll forgive me, I couldn’t write an article about SEO, without reminding you about Lucid SEO – we re-launch with some new suprises on 1st July 2008. http://www.lucid-seo.com

Tweet This Post