Welcome to the Q&A Forum
Browse the forum for helpful insights and fresh discussions about all things SEO.
Hey Brad.. firstly, posting client URLs is fine as far as SEOmoz Q&A is concerned, however, it's entirely up to you in terms of client privacy.
Magento has a setting in the back-end for solving www/non-www canonicalisation (System > Configuration > Web > URL options). You'll see a setting for redirect type ("Auto-redirect to Base URL"), and obviously you'll want to ensure it's set to 301 rather than 302.
Keyword rich anchor text refers to links containing your target keywords within the link text - as opposed to links which say "click here", "link", the domain name, the website name, and so on.
If, for example, you sold grey outdoor widgets, keyword rich anchor text would be phrases like "Buy grey outdoor widgets online", "this selection of grey outdoor widgets", etc.
Exact match anchor text - e.g. "grey outdoor widgets" - would be considered a type of keyword rich anchor text - but post-penguin, something to avoid overdoing. Instead you should be trying to get a variety of links, some keyword rich, some with no keywords at all (like "click here", the URL, etc). A smattering of exact match anchor text is still useful, though.
Whilst it's not necessarily "bad" per se, the implications are, so this kind of canonicalisation issue needs to be taken care of using URL rewrites/permanent 301 redirects.
Typically, on a Windows-based server (without any URL rewriting), a 200 (OK) status code will be returned for each version regardless of the combination of upper/lower-case letters used - giving search engines duplicate content to index, and others duplicate content to link to. This naturally dilutes rankings and link equity across the two (or more) identical pages.
There is an excellent section on solving canonicalisation issues on Windows IIS servers in this SEOmoz article by Dave Sottimano.
On a Linux server (without any URL rewriting) you will usually get a 200 for the lower-case version, and a 404 (Not Found) for versions with upper-case characters. Whilst search engines wont index the 404, you are potentially wasting link equity passed to non-existent pages, and it can be really confusing for users, too.
There is a lot of info around the web about solving Linux canonicalisation issues (here is an article from YouMoz). If your site uses a CMS like Joomla or Wordpress, most of these issues are solved using the default .htaccess file, and completely eliminated when you combine this with a well chosen extension or two.
You can help the search engines figure out which version of a page you regard as the original by using the rel="canonical" meta tag in the html . This passes link equity and rankings from duplicate versions to the main, absolute version.
Take a look at their robots.txt - http://www.dmoz.org/robots.txt
They disallow the /public and /editors subfolders. The editor pages, whilst indexed by Google, aren't crawled.. so whilst the location of the pages themselves is indexed (because of links to those pages), the contents of the pages aren't indexed. This means any links on the page too, obviously..
For this reason, I don't agree with Reload Media. For me, there's no point expending any effort promoting the page for link equity benefit.
The fact they show good authority on OSE is something of an anomoly. They can accrue authority (and indeed Google PR) from their inbound links, however, they are a bit of a dead end, due to the fact that no actual content is indexed.
The problem is that there is little to no content on these pages, aside from the the "boilerplate" stuff - header, navigation, sidebar, footer, etc - which appears on every page.
What that means to a crawler/bot, is that these pages look very similar. If there aren't plenty of unique words on the page, there's always a chance that the page will look like duplicate or thin content.
How many words is unique? Hard to say, but a general rule of thumb has always been 150-200 words for me. It's possible that the number could vary depending on the ratio of boilerplate to unique text, though these days Google is pretty good at distinguishing boilerplate from content.
The issue is that thin content can be a problem in a post-Panda world, although on the scale you're looking at Google can most likely deal with it just fine.. For more info, to help you make an informed judgement, I recommend this [great 'thin content' article from Dr Pete](. http://www.seomoz.org/blog/fat-pandas-and-thin-content).
Could you add some descriptive text on the client directory pages, and redirect the empty articles somewhere else (or fill them with content?) Another option for the client pages could be to consolidate (and redirect) all of the categories into one client page - maybe using Javascript to hide/show each category on request, to keep things tidy.
When you say Google Web Tools I'm assuming you mean Google Webmaster Tools. Google have access to their entire index, so they can see all the backlinks they show in GWT (and some that they don't). Quite often this will be a number in excess of that shown by other tools.
If I run the URL through ahrefs, I see 653 backlinks from 14 domains.
SEOmoz's metrics are based based on their own, more limited data, which prioritises sites to decide which should or shouldn't make the cut (as I understand it). It is updated once every month or so. You will probably discover that most of the sites which link back to you so far aren't themselves in the Mozscape data, so aren't showing as backlinks in SEOmoz toolbar or OSE.
An additional point which I noted whilst prodding your site, is that your website shows for both www and non-www URLs. This is an SEO nono, is considered duplicate content (across the entire site in this case), and is watering down your rankings. In addition, it can skew the backlink count shown using various tools (though this doesn't seem to be the case here).
In your case, it appears redirecting to non-www will probably be best, as this is the URL Google has been indexing. Also appears you are using Wordpress, hopefully on a normal Linux/Apache server. If so, you should be able to use this rule in your .htaccess file.
removes www
RewriteCond %{HTTP_HOST} ^www.crosstrainingandfitness.com [NC]
RewriteRule ^(.*) http://crosstrainingandfitness.com/$1 [L,R=301]
It should do, though don't overdo it and fill your articles with hundreds of keyword rich anchor text links.. but where relevant and useful, in context, adding internal links to the body of your own articles will certainly help the linked pages rank for the anchor text you link with.
The more overall authority your site builds up over time, the more pronounced this effect will be, as you'll have more link equity ('juice') to spread around your internal pages.
Although you say the client is interested in organic search long term, something to consider if you have the budget is using PPC in the early stages of the campaign, for two reasons.
Firstly, the client can obviously use PPC advertising to bring traffic whilst they are waiting for the organic search traffic to grow, but secondly, can be using the PPC traffic data to help shape the SEO campaign.
After initial keyword research, the best way to discover the best converting keywords is to analyse the traffic received for each one. PPC will allow them to start making these judgements more quickly, which should reduce the SEO time (and money) spent on less effective phrases.
No ideas or suggestions, I'm afraid, but I've noticed this too for some searches. Only seems to have started happening recently.
As an in-house SEO I've seen a variety of attempts to scare, trick or otherwise confuse our MD into using certain companies. It never ceases to amaze me the depths to which some firms will stoop. I understand the method you speak of is a fairly common tactic amongst IT security businesses, too. Cause a problem yourselves, then become the saviour.
The second link will count in the example you linked to, as it's to a different landing page, and in this case, with different anchor text, too. If, however, it linked to the homepage twice instead, even with different anchor text, only the first link/anchor text would count.
In my opinion you will be completely fine.
For starters, the sitewide link is nofollow, just like in the original comment thread. As nofollow links are removed from the link graph, the assumption is they are also ignored by any webspam calculations. It's a discussion I've read a few times around the SEO forums, and most people seem to concur.
In addition, even if the links were _not _nofollowed, the anchor text here is only a person's name, rather than any money keywords. This undoubtedly would make it far less likely to be a cause for concern in the context of Penguin.
edit: apologies for repeating what Irving said, I didn't see his comment until I'd finished mine.
Another key brand signal in my mind (and if it's not, it should be) is people searching for your brand name, or brand name + anchor, in Google to find you, for example SEOmoz must get a lot of traffic every day for people searching "SEOmoz", "SEOmoz rank tracker", etc.
This is obviously more difficult for Google to determine if your brand is also an exact match anchor phrase.
One way to strengthen the signal is to use a search phrase instead of a URL in your real world/print marketing. This could sometimes take a little SEO groundwork to do, but for example, if you are Electric Fan Services Ltd and you have a promo for Honeywell fans, instead of using the URL on your print marketing, add a footer saying something like, "Google 'electric fan services honeywell promo' to discover the deals".
Jane Copland did a fantastic talk on 'Golden Links' at LinkLove London this year. Her advice in this situation, like Host1 suggested, is to just email and ask, what do you have to lose? As EGOL said, you could check if they ever link out, to see if you have a hope, if not, still ask!
If you can keep the domain, and point it at the nameservers of your new 301 friendly host, you should be able to keep most of the authority if you then redirect to a new domain. If you can do that, however, why do you need to change domains at all?
I was about to edit my original post to include it, but check out this Feb 2011 article from Rand - he writes specifically about brand signals in some detail.
Lately, I've noticed some big multinationals using search phrases in TV campaigns instead of URLs. I assumed it was largely an SEO play to strengthen brand signals, although some of it could be to do with cutting costs (of creating satellite sites for a promo, maintaining hundreds of random promo-specific domains, etc).
I think you've already taken the correct course of action here. Take a look at this OSE result, comparing your Root Domain with /index.html..
The lack of internal links to your Root Domain is a result of your menu structure linking to /index.html before. The recommended thing to do would be exactly as you have done, change all the links to point to the Root Domain, and 301 /index.html to the Root Domain, too.
It's best practice to only serve a particular site one way or the other, either www or non-www, and only have one URL available for the homepage. This prevents any related duplicate content issues, and ensures you always get backlinks back to the right domain. I'd definitely recommend you add a rewrite rule to your .htaccess, using one of the following examples.
removes www
RewriteCond %{HTTP_HOST} ^www.yourdomain.com [NC]
RewriteRule ^(.*) http://yourdomain.com/$1 [L,R=301]
adds www
RewriteCond %{HTTP_HOST} ^yourdomain.com [NC]
RewriteRule ^(.*)$ http://www.yourdomain.com/$1 [L,R=301]
You can confirm this yourself.
First, do a Google search for site:http://www.dmoz.org/public/profile?editor=
You see the meta descriptions aren't indexed in the results? Instead, Google puts a default message, with a link to this page.. https://support.google.com/webmasters/bin/answer.py?hl=en&answer=156449 - check that out. Note the paragraph:
"While Google won't crawl or index the content of pages blocked by robots.txt, we may still index the URLs if we find them on other pages on the web. As a result, the URL of the page and, potentially, other publicly available information such as anchor text in links to the site, or the title from the Open Directory Project (www.dmoz.org), can appear in Google search results."
So whilst they may appear in Google's index (and indeed the OSE one) because of the links pointing to them, the content isn't crawled at all (by any spiders that obey robots.txt).
You lost a whole lot of low quality backlinks through Feb and March, Panda related I assume.
Appears you lost a PR5 sitewide (paid link?) this month too.
You also appear to have a whole lot of exact match footer links on completely unrelated sites, possibly the result of a template sponsorship scheme? A potential banana skin for paid link penalties and/or Penguin.
Your anchor text profile as a whole is a pretty unnatural..
820 links from 273 domains - wholesale
503 links from 156 domains - wholesale clothing
118 links from 94 domains - ebay
...then a bunch more.. then way down..
85 links from 22 domains - http://www.wholesaleclearance.co.uk/
10 links from 9 domains - wholesale clearance
It's probably more likely to be the paid link stuff that's resulted in the loss of toolbar pagerank, but that's a guess. Just wanted to highlight a few key parts of your backlink profile.
Yes, in that case, use the code that adds www. Then anyone who links to or attempts to visit the non-www, will be 301 redirected to the www version.
You might want to also check out these posts.
http://www.seomoz.org/blog/link-profile-tool-to-discover-linking-activity
http://seogadget.com/bad-backlink-checking/
SEOGadget also have a free tool you might find useful - http://tools.seogadget.co.uk/ - you can export your backlink data from OSE/GWT/Majestic/etc and feed them into the tool 200 links at a time.
Ok I see, understandable - thanks for posting the Rich Snippets data instead.
In our experience, Google isn't showing stars in the SERPs for Organisations/Corporations, but they are for Places. This makes it possible to use schema.org's LocalBusiness with your company reviews, for which Google is showing stars in the SERPs, and this is how we've seen many companies serving up their ratings.
To my mind, it's perfectly legitimate to use LocalBusiness on pages that relate to real world locations, on your homepage if you have a single location, or on outlet-specific pages for businesses with multiple locations.
My assumption would be is that Google first interpreted your aggregate rating data as relating to a product, which is why you got stars showing. They subsequently determined the data was related to your organisation instead, and removed the stars.
I have absolutely no idea whether this would have been algorithmic or manual, but I've certainly seen a bit of a cleanup in the SERPs in recent weeks, and murmurings in the SEO world to that effect too - see http://www.seroundtable.com/less-google-rich-snippet-15216.html for one example.
Hi Kate - this is nothing to worry about.
Rel Canonicals are generally shown on the SEOmoz PRO dashboard as an alert/heads up, rather than a direct warning. There is, of course, potential for real damage if they're configured incorrectly, but in your case it appears the Wordpress SEO plugin you're using is doing fine.
Your canonical URLs are showing as what you most likely would expect and want them to be, the post permalinks or the category URLs.
I experimented and changed some of the letters in the URLs to uppercase, to check for problems. Whilst the URLs still resolved with uppercase characters and didn't 301 redirect to lowercase (which would generally considered best practice), the Rel canonical URLs set remained the lowercase version even with mixed case in the browser bar - this is, again, exactly how you would want it to behave.
If you don't have access to Webmaster Tools, the most basic way to see which pages Google has indexed is obviously to do a site: search on Google itself - like "site:google.com" - to return pages of SERPs containing the pages from your site which Google has indexed.
Problem is, how do you get the data from those SERPs in a useful format to run through Screaming Frog or similar?
Enter Chris Le's Google Scraper for Google Docs
It will let scrape the first 100 results, then let you offset your search by 100 and get the next 100, etc.. slightly cumbersome, but it will achieve what you want to do.
Then you can crawl the URLs using Screaming Frog or another crawler.
Rel Canonicals are generally shown on the SEOmoz PRO dashboard as an alert/heads up, rather than a direct warning. There is, of course, potential for real damage if they're configured incorrectly, but in your case the Wordpress SEO plugin you're using is probably doing just fine.
As long as your canonical URLs are what you want them to be, i.e. the relevant post permalinks or category URLs, you're golden.
It shouldn't really matter too much either way.
The difference with a .co.uk is that Google will automatically assume you are targeting that geographical area, whereas with generic domains like .net and .org, you'll have to set it in Webmaster Tools - or you can leave it unset and have Google work out which countries you are relevant to from other signals.
There's absolutely nothing wrong with using a .net to target the UK - although if you plan for the site to target more than just the UK in future, it's probably best left undefined in GWT.