14 replies to this topic

[SEOMonkey]

I have a question for group…

My employer is getting ready to launch a new website (same domain, just a revamped site) and I’d really appreciate feedback on my migration plan.

I realize some of this is probably pretty standard, but we have pretty good placement and traffic now and I’d like to do everything I can to mitigate problems that are going to occur in the SE’s with the launch of the new site (esp. loss of links and PR).

So I’d appreciate comments on the plan, or feedback to let me know if I’m completely off-base or missing something important, or any other great ideas you may have!

And if I’m not way off-base, hopefully this will help others, as I haven’t been able to find much information online around this

General Site Info:

PR – 7 homepage
Alexa – under 1k
Pages – 40–50 +
Daily Traffic - 5 - 10 k uniques

Considerations:

• This company has been launching new websites every 1.5 or 2 years for awhile, and there are now 3 or 4 very out-dated versions of their site existing on their domain.
• The lead designer wants to use all new directory structure on the new site to better organize the architecture and navigation (which would mean no recycling of URLs)
• Wordpress Blog will remain unchanged

Migration Plan:

• Get rid of old, out-dated pages by creating 301 redirects to the most appropriate pages on the new site.
• Send email to our Affiliates to let them know of the change and get them an appropriate list of new URL’s (affiliate links will be updated automatically)
• Create updated sitemap and submit/verify
• We are scheduling a new Product Launch at the same time (Hopefully this will create a spurt of new links to the new pages quickly)
• Increase PPC for essentially the same reason as # 4

Specific Questions:

• Should I fight to recycle as many page URLs as possible? Or let the designer create new ones and just worry about the 301s?
• Is there a ‘Best’ time to do this sort of thing?
• Are there any Engine Specific details I should consider?

Thanks in advance for any help/feedback with this!

[bragadocchio]

Hi SEOMonkey,

Having a plan before making changes is always a good idea, and the plan that you have sounds like a good one.

Will you be issuing a press release with the launch of the new product line, and other marketing efforts, including offline methods, to get the word out?

An improved directory structure, and strong anchor text can't hurt.

By "sitemap" I'm guessing that you mean a Google or Yahoo sitemap or both, instead of an html sitemap (recommended anyway).

If there are ways to change links pointing to old pages from the site, you should try to address those if possible - update directory listings that you might have control over, send out emails to nonaffiliates if you can and let them know the good news of your relaunch and changed URLs, etc.

Since you have a blog, it might be a good time to work on some high quality blog posts in advance in anticipation of the change.

I'd probably try to see which URLs are the ones that are linked to by the highes quality/quantity links when considering if any are worth recycling. Balance the cost of keeping those versus changing them to fit within the structure of the new version of the site.

There are a lot of aspects to your questions that are difficult to answer based purely upon a technical approach without any knowledge of the business objectives and practices behind the site. For instance, I wouldn't want to make changes like this to an ecommerce site shortly before Christmas if I anticipated a lot of Christmas sales. Or make changes to a site that sold chocolates or flowers a couple of weeks before or after Valentine's day.

You've probably gone through sometype of Content Assessment like the one described here:

http://www.digital-w...or_dis_content/

You have the beginnings of an SEO migration plan in place. Look carefully at the present state of the pages in areas like:

What keyword phrases are bringing you traffic.
What sites are sending you visitors.
What audiences are being targeted on which pages.
Which important pages on other sites are linking to which of your pages.
What affiliates/partners/friends/advertisers are bringing you traffic.
What directory listings might you be able to change.
What directory listings might you be able to add.

In the migration, you'll want to look at:

Are you still targeting the best of the keyword phrases, and incorporating some good new ones?

Are the sites that are sending you the most visitors going to be happy with the changes, and will they be willing to point links to new URLs?

Are you still addressing your most important audiences, and can there be things that you can do to fullfill their needs even better? I don't know the details of your site, but some examples might be things like letting people purchase gift certificates, adding information that makes people feel like they are making informed purchasing decisions, providing additional benefits to people who register if registration is something people can do, etc.

If you follow best practices for SEO, and plan carefully, you may still see some dropoff in traffic. You could possibly offset that by doing things to increase usability, persuasiveness, credibility, and conversions.

[A.N.Onym]

I have a couple of additions to the excellent post above.

Has the designer really researched the needs of the customers to install a new website structure? Can you naturally use keywords (the words your customers are most familiar with) in site architecture?

I don't think using old URLs is necessary, if you setup redirects.

I'd be mostly interested in redirecting all the pages on the old four websites to the corresponding pages of the new website.

I shouldn't delete the old pages, as it'll allow you to setup redirects individually, without having insanely large .htaccess files.

Are you sure you have only 50 pages with 10k visitors per month? What kind of a product/content/service is that?

[SEOMonkey]

Thanks for the replies and great advice.

Bill,

Thanks for the interesting article link. We certainly discussed these types of issues, but in a more informal way (I copied the article for future reference though ).

We’ll definitely be promoting the product launch in our standard ways (PR, email campaigns, webinars etc) so hopefully this will help to promote some new links!

And yes, sorry, I did mean Google/Yahoo sitemaps although the new site will also have an HTML version.

Unfortunately previous versions of the site were pretty much slapped together with too little thought for the customer and none for SEO. It looks like I’m pretty much on track as the new site is designed to improve on most of the points you have listed; with better consideration for SEO, usability, and conversions!


A.N.Onym,

The plan is to use better keywords in the architecture… currently, there are way too many really long URLs and page names with too many dashes

I hadn’t even considered using the .htacess file to do any redirecting, an interesting idea but yeah, I think it could easily become way too large. My plan was to individually redirect all the necessary pages, although, the downside to this method is one cluttered directory

There are definitely only about 50 or so pages, but 2 or 3 versions of most of them.

And the 5-10k are Daily uniques, not monthly. About 35% of the daily traffic is on the blog, the rest goes to the site.

We’re an eCommerce software provider with over 10k active merchants.



So, to summarize, assuming the new site is built properly (in regards to SEO, usability etc), using 301 redirect’s is the best way to handle the old pages, while a good marketing blitz for a new product coupled with some quality content added to the blog is good plan to help ensure links to the new site and hopefully ward off too many problems in the Search Engines?

Here’s one more question. Am I correct to assume that, once I redirect all the old pages, I wouldn’t list them in the Google/Yahoo sitemap?

Thanks again for the help!

[bragadocchio]

You're welcome.

I think that your summary is spot on.

Right, you want to list the new pages in the sitemaps for the search engines, and not the old ones.

[A.N.Onym]

I'd have to remind you about your customers, who need to know where they will get after clicking on any link. That's why you need to first address the needs and interests of your customers and only then use the keywords. Yes, sounds boring, but sometimes the most popular keywords may not be the best words for your site.

[Hawaii SEO]

Hi Bill,

I'm also in the middle of a website redesign.

We're moving over to a new MS CMS system. One of the things the new CMS does is add extra characters to the Home Page URL.

I saw that Coca Cola has a similar problem. You can't just go to "Coca- Cola.com. If you do… you get redirected to "coca-cola.com/index-b.html". No matter how hard you try... The address bar will never display the simple URL, only the URL with the extra characters.

Coca Cola is using the Robots.txt file to prevent the URL with the extra characters from being indexed. (I guess to help prevent duplicate content. I don't know.)

http://www.coca-cola.com/robots.txt

User-agent: *
Disallow: /index-b.html

Do you still receive the link juice, PR, trust, authority (or whatever) from a link that points to a page or URL that is disallowed in the robots.txt?

Are their any SEO issues that could potentially result from always displaying extra characters in a Home Page URL like this?

What if the URL always stays the same but the extra characters change from time to time in the future?

Is there a way to avoid displaying the extra characters in the URL when people visit our home page?

I'm OK with marketing side of SEO but I suck at the technical side of things.

My gut tells me the extra characters in the Home Page URL are a very bad move but I don't know this for a fact.

Can you help me out?

I'm hoping for an answer that I can either take to the IT department so everyone will understand the potential list of likely consequences the extra characters in the Home Page URL may have. Or... Something that explains why this is a non-issue that I shouldn't be concerned about. (If it's a non-issue, maybe you can help me to stop obsessing about it.)

Thanks,
Dave.

[bragadocchio]

Hi Dave,

Before going on to your questions, I assume that your new site is on a development server. Do you have it blocked from spiders by robots.txt? If not, you might want to do that, just to avoid the possibility that an errant link somewhere points to it, and starts indexing the new pages.

[Hawaii SEO]

Thanks Bill,

Yes it is... It's on a sub-domain right now as we work out the bugs. We have the Sub-domain blocked with the robots.txt and the nofollow meta tag on every page. I use a FireFox plugin that highlights nofollow links Pink. All the links on the test website seem to be nofollow.

[bragadocchio]

In an ideal world, you would probably want to have the shortest version as your home page. But, in lieu of that, there are things that you can do - I wouldn't suggest emulating Coca-Cola's approach, though.

Microsoft published a patent application a few months ago that detailed how they may be handling situations like this, and I suspect that Google and Yahoo probably don't treat it too differently either.

The patent application is:

System and method for optimizing search results through equivalent results collapsing

In it, they discuss situations exactly like the one that you are interested in.

Here's the scenerio the patent addresses:

[0006] Each information source contains a body of content and can be referenced by a locator such as a URL. Often, search engines locate multiple locators or results that access duplicative content. If a search engine obtains ten relevant results, three of those relevant results may lead to the same content. For example, www.ymca.net and www.ymca.net/index.isp access the same content with the former redirecting to the latter. In addition, www.ymca.com and www.ymca.com/index.jsp are also mirrors of www.ymca.net. To accurately measure whether a system returns optimal results for the query "ymca", the system must determine whether these results lead to equivalent content.

[0007] As set forth above, one of the results may link directly with the content and the other results may go through a series of redirects to access the same content. Currently, most search engines will fail to detect and correct this duplication. Accordingly, users may access three different results to ultimately reach the same content three times. The failure to detect and filter out these duplicates results in frustration and time waste for the user. Search engines do a poor job of recognizing that content is repeated over and over again and of de-duplicating content from search engine results. This failure results in a sub-optimal user experience in which the user selects multiple search engine results, but receives the same content each time.

So, in that example, there are four URLs:

www.ymca.net
www.ymca.net/index.isp
www.ymca.com
www.ymca.com/index.jsp

They tell us that one of the four goes directly to the content, and the others redirect to that URL. They also tell us that search engines have a hard time with this. They describe a solution, where all URLs that are written differently, yet point to the same content, getting collected in a database that understands they are duplicates of the same page.

They may use these different URLs differently, according to the patent application, such as showing www.ymca.net in search results to searchers but sending those searchers directly to www.ymca.net/index.isp when they click on a link to the page. I've seen Matt Cutts use the phrase "the pretty URL" in talking about similar instances in Google when discussing redirects and having the shorter URL displayed while sending people to the longer URL with a redirect directly. I suspect that Google is doing some similar analysis.

Here's where they say that in the patent filing:

[0035] As set forth above, www.ymca.net and www.ymca.net/index.jsp, www.ymca.com and www.ymca.com/index.js may all lead to the same information. After the duplication detection mechanism 240 detects that these results lead to duplicate content, it will store the results in the result storage area 360, but will not display them all. The components of the result selection module 300 will operate on the detected duplicates to select the best result to display to the user as well as the result that will serve as a navigation model to most quickly access the content. The result storage 360 may maintain all of the duplicates, while the user-preferred result selection mechanism 350 will only select one of the results to show to the user and use one of the results as a navigation model.

I don't think that there's a need to disallow the http://www.example.com/index-b.html version of the URL in robots.txt, and I believe that might be more harmful than helpful. I suspect that coca-cola is hurting themselves some by doing what they do. Interesting set of header responses when arriving at the coca-cola homepage:

300 keep alive
404 not found
302 moved to a temporary location

I suspect that confuses the search engines some. I do think that using a 404 in that instance could be harmful. I'm also not sure that the coca-cola site is a good role model in a number of other ways when it comes to SEO.

If you use a 301 redirect, from the home page to the default new page, you'll probably be fine. I don't think that there is a need to disallow either version of the page. I would suggest that when you point to "home" from other pages on the site that you use the longer version in your internal links, too. (see Matt's input on canonical URLs. The trick here is to make it easy for Google/Yahoo/Microsoft/Ask understand which URL is the canonical URL for your home page. They don't always get it correct, but a redirect and consistent internal use can help them.

[Hawaii SEO]

Thanks Bill!

Wow! Great Response!

At this time, you can use two different URL's to view our current website.

For example:
Both "shorter-url.com" & much-longer-url.com both work with our website.

If you begin with "shorter-url.com" you will continue to see that URL as you navigate deeper into the website. Same thing happens if you use "much-longer-url.com".

I thought it might be a good idea to fix the problem or prevent a posible problem from happening so I tried my best to show that exact post by Matt Cutts to the IT team a few weeks ago.

My request was something like this:

Set http://www.example.com/ as the default URL & make the webserver so that if someone requests http://example.com/, it does a 301 (permanent) redirect to http://www.example.com/.

This was the response.

"Since this is not a problem right now and has not been a problem in the past, there is likely very little danger of this situation occurring in the future." (In other words... No need)

Since I don't understand the technical issues... It's hard for me to come up with a compelling argument to counter the fact that "It's not a problem right now nor has it been in the past".

In your opinion... When search engines get this type of thing wrong... What's the usual consequence. What are the chances of it happening to a large website?

Thanks again!
Dave.

[bragadocchio]

Hi Dave,

That's a canonical issue that you should be able to fix with a few lines of code, hopefully. If the site is on its own dedicated server then something like isapi rewrite might make it fairly easy to do.

This issue is one that the search engines don't always get right, though you would hope that they did. If they don't recognize those as the same page, and they might not without the permanent redirect, then you may be splitting pagerank instead of distributing it well.

An example of a site that appears to have that issue:

http://jetblue.com/ - pagerank 6
http://www.jetblue.com/ - pagerank 7

There are others where it shows a pagerank in the toolbar for one version, and almost none in the toolbar for the other. The New York Times was showing a 9 for one version, and a 6 for the other. They may have changed their internal linking structure around to favor the higher one - now it's showing 0 for the version without the "www." If the search engine recognizes them as the same page, it would show the same pagerank in the toolbar for them. It isn't. Check your site...

Fixing this problem may make it easier for the search engines to correctly recognize what is going on with the long and short versions of the home page.

[Hawaii SEO]

Sweet!

That's the exact example & situation I was looking for.

Thanks again! This was a huge help for me!

Aloha,
Dave.

[bragadocchio]

You're welcome, Dave.

Thought you might like that example.

[SEOMonkey]

A little note on canonical issue solutions:

Matt Cutts' 2 cents


My 2 cents:

If you're running apache, you can add the following to your .htaccess file:

RewriteEngine on
RewriteCond %{HTTP_HOST} ^example.com/
RewriteRule ^(.*)$ http://www.example.com

which tells the site to take all non-www. traffic and redirect to the www. version of your site (and of course you can flip it around if you prefer the non-www. traffic).

Also, if you log into Google webmaster tools, there's now an option where Google allows you to specify which version of your site you would prefer to be indexed.

I haven't had any time to experiment with it yet, but it's there.