Welcome to the Q&A Forum

Hi there,

While doing some research on the indexation status of a client I ran into something unexpected. I have my hypothesis on what might be happing, but would like a second opinion on this.

The query 'site:example.org inurl:index.php' returns about 18.000 results. However, when I hover my mouse of these results, no index.php shows up in the URL. So, Google seems to think these (then duplicate content) URLs still exist, but a 301 has changed the actual goal URL? A similar things happens for inurl:page. In fact, all the 'index.php' and 'page' parameters were removed over a year back, so there in fact shouldn't be any of those left in the index by now. The dates next to the search results are 2005, 2008, etc. (i.e. far before 2013). These dates accurately reflect the times these forums topic were created.

Long story short: are these ~30.000 'phantom URLs' in the index out of total of ~100.000 indexed pages hurting the search rankings in some way? What do you suggest to get them out? Submitting a 100% coverage sitemap (just a few days back) doesn't seem to have any effect on these phantom results (yet).

The website of a partner of mine was recently migrated to a new platform. Even though the content on the pages mostly stayed the same, both the HTML source (divs, meta data, headers, etc.) and URLs (removed index.php, removed capitalization, etc) changed heavily. Unfortunately, the URLs of ALL forum posts (150K+) were redirected using a 302 redirect, which was only recently discovered and swiftly changed to a 301 after the discovery. Several other important content pages (150+) weren't redirected at all at first, but most now have a 301 redirect as well. The 302 redirects and 404 content pages had been live for over 2 weeks at that point, and judging by the consistent day/day drop in organic traffic, I'm guessing Google didn't like the way this migration went.

My best guess would be that Google is currently treating all these content pages as 'new' (after all, the source code changed 50%+, most of the meta data changed, the URL changed, and a 302 redirect was used). On top of that, the large number of 404's they've encountered (40K+) probably also fueled their belief of a now non-worthy-of-traffic website.

Given that some of these pages had been online for almost a decade, I would love Google to see that these pages are actually new versions of the old page, and therefore pass on any link juice & authority. I had the idea of submitting a sitemap containing the most important URLs of the old website (as harvested from the Top Visited Pages from Google Analytics, because no old sitemap was ever generated...), thereby re-pointing Google to all these old pages, but presenting them with a nice 301 redirect this time instead, hopefully causing them to regain their rankings. To your best knowledge, would that help the problems I've outlined above? Could it hurt?

Any other tips are welcome as well.

Assuming you are absolutely sure your page has been set up correctly, you should trust that data source more. How is Google showing up the pages when you're performing a 'site:www.yoursitename.com' search?

You seem to already be logged in. You can click on the icon of the person (with the fancy lines next to it) in the top right corner to see your data. If you were logged out, a 'Log In' button would be visible.

Given Moz' usual transparency I'm sure there will either be an update within a few days, or a mention of why they have failed to release a new Mozscape update at the time they had planned to release it.

You're stating "I tend to link using Branded ...", are these links internal or coming from external websites? If they are internal, there isn't anything to worry about. When you have actively 'build' those links from other websites, I would be more cautious with pages that you're giving keyword-rich links, and indeed mix it up a little.

What you would want most is to have a natural overall profile. Not necessary a perfect profiles for each page (which, ironically, would be unnatural), but a pattern that would look perfectly diverse, skewed and chaotic in all its natural perfection.

Yes, after all, the content is 'duplicate' across these pages. It depends on the amount of duplicate content across pages whether Google will consider it a problem though.

There might be more. Is there any chance you can share the website here? Perhaps with spaces in the domain name so the search engines won't pick up on it? It would help us to have a look at what might be happening over there.

You should be looking at this from two different angles: (1) how would the hosting affect your rankings and (2) how will it affect the loading time of your website.

In terms of rankings, especially since you're considering to purchase a .info domain (side note: are you sure about this? Domains with .info isn't exactly considered a super premium TLD), I don't think it'll make much of a difference whether you'll be hosting in the USA or in Germany. Had it been a .de targeting Germany (much like the articles you're referencing) things had been different and I would've advised you to keep hosting the website in Germany.

Add the fact that it seems that fewer than 5% of your visitors are coming from Germany in the first place, I would say a move to an IP located in the US wouldn't hurt your number of visitors. Perhaps it might even increase your rankings in the USA marginally because Google now finds your IP hosted over there.

In terms of loading time you might want to consider CDN solutions to serve your website as quickly as possible to a global crowd.

What does Google show you when you perform a site: search? What pages does the SEOmoz panel list when you click on the duplicate titles entry? I suspect there are several dynamic variables that are creating more pages than you want there to be. Have you properly canonicalized your pages with rel="canonical"?

I'd say you can safely put this issue aside. As you've mentioned yourself, xml files don't need title tags, and I presume you aren't trying to have this page rank in the SERPS in the first place anyway. Perhaps one of the SEOmoz fellows can look into this?

"99.9% of the time, if a subfolder will work, it's the best choice for all parties." (http://www.seomoz.org/blog/subdomains-subfolders-and-toplevel-domains).

Even though this is a post made in 2006, I don't know of any relevant changes that would alter this advice.

That depends a lot on what you're trying to achieve with those titles. Are you looking for maximum exposure of the title to the user? Then I'd suggest ~45 characters (which is the maximum number of characters visible in the Safari browser, my personal favorite 'Atomic' shows only 10 characters in the tabbed interface). If you're trying to have the page ranked in the iPhone version of Google, I'd advice the regularly advised title tag length of no more than 65 characters as Google appears to have the same cut-off on mobile and desktop browsers. Or perhaps there is another goal you're trying to achieve?

There isn't really a maximum or optimal length to a URL (well, technically it is 2048 characters [http://www.boutell.com/newfaq/misc/urllength.html], but I don't think that was the kind of limit you were refering too).

In general it is a good idea to keep the URL as short as possible, but at the same time try to include keywords in it that make it clear to both the user and the search engine what the article is about. As with many things in SEO: try to find the right balance.

Unfortunately you can't, as Dr. Pete explained in the following blog post: http://www.seomoz.org/blog/6-extreme-canonical-tricks

Even though I'm not a SEOmoz employee, I think I can answer this answer on behalf of them.

SEOmoz will not make your website more visible or accessible as a direct result of your payment. If you decide to take a PRO membership with the site however, you'll get access to premium resources that will enable you to take your website to the next level with regards to SEO, and keep increasing that level over time.

By using the (premium) tools that are available to PRO members (such as the campaigns that you can use to monitor weekly crawls of your selected websites), you can see very clearly which mistakes you've made on these sites and what you can do to improve your SEO.

From what I've read on the internet, it is best to "noindex,follow" all pages >1. This issue had bugged me for quite some time as well, and I've struggled to find good resources explaining why their solution was the best. Now that I've actually given the subject some thought, and finally managed to read some quality material on the matter, it all makes sense.

It's basically a checklist. Do you want search engines to

• index your paginated result pages: yes / no

• reach the items that are listed in your paginated result pages: yes / no

In most cases you don't want your paginated result pages to be indexed. With our without Panda, visitors get little value from actually viewing 'page 7' in your result pages. That actual page provides little or no value to those visitors. However, you DO want those items listed on these paginated pages to be crawled, especially when you don't have any other pages linking to them (which you should by the way). This boils down to:

• Don't nofollow your paginated links (because you want search engine spiders to reach them)

• Put "noindex,follow" in the meta robots tag for all pages >1 (thus page 2 and greater) so the engines will no index these paginated results, but will crawl on to the pages that are behind the listings

Good luck!

Some of the following will be guesswork since you didn't provide any URLs, but I'll try my best. This old (Atlanta targeted) website, has it been around for a (quite) longer time than the newer (Nashville) domain? Besides the amount of links the older domain has most likely collected, domain age appears to influence ranking on its own (even though only slightly, #10 in the ranking factors http://www.seomoz.org/article/search-ranking-factors#ranking-factors). Does the Nashville targeted website have the same amount of local (and related) backlinks as the Atlanta targeted website? You've mentioned that the Nashville website is only live for about 2 months, which I'd consider a really short time to draw any real conclusions to be honest.

With some more time and the same effort as you've put into the Atlanta targeted website, I'm sure the new one will perform in a similar fasion!

You should be looking at this from two different angles: (1) how would the hosting affect your rankings and (2) how will it affect the loading time of your website.

In terms of rankings, especially since you're considering to purchase a .info domain (side note: are you sure about this? Domains with .info isn't exactly considered a super premium TLD), I don't think it'll make much of a difference whether you'll be hosting in the USA or in Germany. Had it been a .de targeting Germany (much like the articles you're referencing) things had been different and I would've advised you to keep hosting the website in Germany.

Add the fact that it seems that fewer than 5% of your visitors are coming from Germany in the first place, I would say a move to an IP located in the US wouldn't hurt your number of visitors. Perhaps it might even increase your rankings in the USA marginally because Google now finds your IP hosted over there.

In terms of loading time you might want to consider CDN solutions to serve your website as quickly as possible to a global crowd.

You're stating "I tend to link using Branded ...", are these links internal or coming from external websites? If they are internal, there isn't anything to worry about. When you have actively 'build' those links from other websites, I would be more cautious with pages that you're giving keyword-rich links, and indeed mix it up a little.

What you would want most is to have a natural overall profile. Not necessary a perfect profiles for each page (which, ironically, would be unnatural), but a pattern that would look perfectly diverse, skewed and chaotic in all its natural perfection.