Terry's Blog

URL Parameters for SEO Newbies

Wed, 29 Apr 2009 03:40:00 GMT

Ever wondered what the question marks and ampersands are for that you see in many URL’s? The fact is they can serve a lot of uses, and can also cause a lot of SEO headaches if not considered as part of your search engine optimization auditing.

A web address is the location of a physical page on the web. The following is the web address to the Google Finance page. http://www.google.com/finance

If a question mark (?) exists in a URL, then everything from the start of the question mark and onward is called the Query String. A query string is typically made up of name/value pairs, sometimes loosely called parameters.

http://www.google.com/finance?catid=66529330

In this URL, the web address is still the same, but now there is a query string (?catid=66529330) and it contains just one name/value pair. The parameter name is “catid” and the value is “66529330”.

Additional name/value pairs beyond the first set will always be separated with an ampersand (&) in the URL.

http://www.google.com/finance?start=20&num=20&catid=66529330

In all of the URL examples above, we are actually calling the same web page from Google… the Finance page. In the last two URLs however, we are passing in some extra information that Google uses to determine the type of information to display, and how to display it.

There are a five primary reasons that web developers will choose to use query strings in there URL’s.

Controlling Page Content
When dealing with a lot of content, it’s easier to build one physical web page that serves as a template, and pass in a value in the query string to control the data that fills that template. This is perhaps most common on retail websites. If you think retailers build a web page for every single product or category, think again. Typically retailers have only one category and product web page, but the content on this page changes based on the identifier that is passed into it.
example: http://www.mystore.com/product?productid=12345, http://www.mystore.com/category?id=54321

Controlling Page View State
Many pages need features like sorting, number of items, and pagination. On the product category pages of a retail site, you may wish to sort by price, view sixty items at a time, and see what is on page two.
example: http://www.mystore.com/category?id=54321&sort=price&view=60&page=2

Tracking Click Paths
Often, if you click “Add To Cart”, “Log In”, or “Register” on a web site and complete the action, the web site will send you back to the page you were last at. Developers will commonly use the URL query string as a mechanism for keeping track of where to send you back to.
example: http://www.mystore.com/login?ReturnTo=mystore.com/home

Tracking Session State
Occasionally we also still run into the website that chooses to keep track of a particular site visitor by storing the session identifier in the URL as a query string. When you first add an item to a shopping cart, pretty much all web sites have to assign you a unique identifier. Most websites will keep that identifier in a cookie behind the scenes, but some will instead put that identifier in the URL. This is how the sites are able to keep track of the items that are in your shopping cart, without you having to be logged in.
Example: http://www.mystore.com/home?sessionid=5049399928384885960

Tracking Marketing Efforts
Most analytics packages have built in support for what are called “campaign codes”. Whether it is a paid search, banner ad, email or other type of online marketing campaign, marketers can pass values using a query string to a web page in order to capture a measure of the effectiveness of that tactic.
Example: http://www.mystore.com/home?campaign=My_Yahoo_Banner

The most important SEO concept when dealing with URL’s is that search engines view every unique URL as a different web page by default.

We (humans) know that the following URL’s are the same, but just sorted differently.
http://www.mystore.com/category?id=54321
http://www.mystore.com/category?id=54321&sort=price
http://www.mystore.com/category?id=54321&sort=bestselling

Google however doesn’t by default, and if we let Google index all versions, we risk competing with our self in search results, and potentially splitting the link juice across all variations. Yes, the search engines have become better at recognizing these instances on their own, but why risk it. Typically the only types of query strings I like for search engines to index are the ones that just control the page content. Anything that controls view state, session state, click paths or contains tracking codes will cause me to put an SEO effort around keeping them out of search engines.

First, I will usually block the spiders from indexing the trouble URL’s using statements like below in the robots.txt file.
# Variable Excludes
Disallow: /*?sort
Disallow: /*?view
Disallow: /*?page
Disallow: /*?ReturnTo
Disallow: /*?SessionID
Disallow: /*?campaign

Lately, I’ve also been touting the use of the new domain canonicalization tag that is supported by major search engines as a way to reinforce the proper default URL for a given page. This tag basically tells search engine spiders that “regardless of what URL you used to get to this page, here is the correct version for the search engine index”.

Other things to watch out for:

Session ID’s in URL’s can cause spider traps that could throw search engines into endless loops and keep them from indexing fresh content on your site.
Calendars that use the URL query string to keep track of the dates currently being viewed can also cause a spider trap if end dates aren’t specified.

Free tools like Xenu Link Sleuth or Sitemapbuilder.net can help you to identify the cases where query strings are problematic.

Google CSE and DotNetNuke (DNN)

Wed, 17 Dec 2008 07:23:00 GMT

I wanted to find a way to display Google Adsense ads for the search results on one of my DotNetNuke driven sites. I got it to work at Celebrity Couples Online. I'm sure there is a more professional method, but here were the steps I took. This method is pretty safe.

Create a new page on your site that is hidden (I called mine GSearch), add a text module to the page, and paste the code that Google provides for search results into that text module using the source input option. This will become your search results page. Now you just need to modify your Skin files to pass over the details in the parameter that the Google code needs to see.
In your /admin/skins directory, create a copy of these four files.
1. Search.ascx -> GoogleSearch.ascx
2. Search.ascx.resx -> GoogleSearch.ascx.resx
3. Search.ascx.vb -> GoogleSearch.ascx.vb
4. Search.xml -> GoogleSearch.xml
In your /admin/skins/App_LocalResources directory, create a copy of Search.ascx.resx -> GoogleSearch.ascx.resx
You will only need to edit two of the above files.To start, open up GoogleSearch.ascx.vb
1. Change Const MyFileName As String = "Search.ascx" to Const MyFileName As String = "GoogleSearch.ascx"
2. Change the ExecuteSearch function to the below, inserting your own data where needed
    If Not String.IsNullOrEmpty(searchText) Then
         dim str as string
         str = "?cx=%3A72pzbzmchja&cof=FORID%3A9&ie=UTF-8&q=" & Server.UrlEncode(searchText)
    response.redirect(str)
    End If
Second, in GoogleSearch.ascx, change CodeFile="Search.ascx.vb" to CodeFile="GoogleSearch.ascx.vb" in top line.
In your website skin files, change <%@ Register TagPrefix="dnn" TagName="SEARCH" Src="~/Admin/Skins/Search.ascx" %> to <%@ Register TagPrefix="dnn" TagName="SEARCH" Src="~/Admin/Skins/GoogleSearch.ascx" %>

Hoping I didn't leave out a step, but it seems to work for me. I'm not all that thrilled with Google CSE's indexing time however. I'm seeing it take upwards of two weeks to index new pages, even with me adding the pages to an XML Sitemap and requesting a re-index from the CSE tools. If you run into problems, reverting only the very last change puts the old search back. Took a little trial and error!

SEO Tips for Changing Website URL's

Tue, 18 Nov 2008 07:37:00 GMT

Going to change the URL structure of your entire website, but scared about losing your search engine rankings? It's a common issue that I've seen impact both the largest and smallest corporations. Here are some common exchanges with site owners or web developers as how to approach it without losing the weight your old URL's have accrued over time.

More...

Checking Paid Search URLs For Errors

Sat, 15 Nov 2008 20:44:00 GMT

One challenge that large paid search advertisers inevitably face, especially when they aren’t the same group that builds the website that the traffic is ultimately being driven to, is keeping up with landing page URL's. If the website owners move or remove a web page that you are paying to drive people to without telling anyone, it can lead to missed opportunity, extra costs and a terrible online experience for your potential customers. Unfortunately, despite all the technology that Google, Yahoo and Microsoft have, they don't make it easy to pinpoint when this is occurring.

More...

Rant: Are Spam Bots Holding Back Social Growth?

Tue, 17 Jun 2008 04:41:00 GMT

Every week I i pull up SQL on one of my hobby websites, and run the same set of statements to clear the spam from my comments and ratings sections of my pages....

delete * from comments where comment like '%http%'

Today I ran the script, and came back a half hour later to find over 20 new comments attempting (and failing) to create links to the Paxil, Viagra and Porn sites. Yeah, I know I could spend 5-10 hours decompiling the DotNetNuke feedback module I'm using and adding some sort of filter, but what a waste of my time.

Of course, on the flip side, even though the spammers haven't been able to get a link in there to render, I did a search for Google and the link weight my site has for some naughty words, strictly due to spammers attempts, is pretty hefty. I don't want Google to associate my site with that crap. The only way for me to seemingly control it, is to require that people register on the site first. Doesn't this kind of slow the overall growth of the social web 2.0? (Yes, I also know I could add one of those picture letter things. Again, don't feel like totally reimplementing the ratings module I'm using just yet)

So the latest fad seems to be using YouTube to try to pass PageRank around. Here is an obvious automated Spambot attempt I found on my comments today:

This program tries three different ways to auto-post links, using standard HTML and common forum and blog commenting formats. What was interesting was that they were all pointed at YouTube.

I wonder if they have automated bots out there for creating YouTube accounts as well. Seems like Google would have the technology to flag any accounts that have this type of setup. Each of the links in the original spam post on my site was to a unique YouTube account. Obviously the spammer is counting on the little link at the bottom of the YouTube profile passing some link love.

Oh well, I'll just waste some of my life and money shoring up my site, while also making honest users lives a little harder.

Seriously, who wins here?

Adding H1 Tags to DotNetNuke Containers

Sun, 08 Jun 2008 17:12:00 GMT

Recognizing in advance that this post has a pretty niche audience, this is one of those annoying things that took some playing around with to make work right.

With most purchased DotNetNuke skins, there isn't an easy way to take a container and make an H1 tag out of the title. This is something that I wish the skin developers would include by default. Here is how I have done it so that search engines can see the emphasized text.

First, pick one of your existing containers that you want to have an H1 heading version of, and download the ASCX and CSS files from your /Portal/_default/Containers directory onto your local computer. I usually just keep the name the same, with the exception of adding _header onto the end of the file name.

Open up container_header.ascx. Typically the title token for a container looks a little like this:
dnn:TITLE runat="server" id="dnnTITLE" Cssclass="ContainTitleCSS"

I will wrap this with an H1 tag:
H1 dnn:TITLE runat="server" id="dnnTITLE" Cssclass="ContainTitleCSS" /H1

By default, there probably isn't an H1 definition in your containers CSS file. So open up your new container_header.css, double check that this is the case, and then define a new one. Here is a sample of one of my sites:

H1 { font-size: 11px; font-family: tahoma; color: #555555; font-weight: bold; margin: 0; padding: 0; line-height: 1.5em; }

Really make sure that you pay attention to the Margin, Padding and Line-Height elements. Those were the parts I had some issues with originally.

Finally, save your two new files, and upload them back to your web server in the correct containers directory. Now you can just apply the container to any module where you would like the text to be emphasized to search spiders. If it looks any different to the regular eye then one of your normal modules, some tweaking may be needed.

You can check to see if this worked by using Firefox and the Web Developer plug-in. Browse to the page where you have added your header module. Using the plug-in, disable CSS and Images. The header module title should look big and bold.

Here is a working sample on one of my test sites: Brad Pitt and Angelina Jolie

Walt Disney World on Google Earth

Sat, 07 Jun 2008 06:40:00 GMT

David Cook, the winner of American Idol, recently announced that he was going to Walt Disney World in an effort to help promote the American Idol Experience attraction opening in Hollywood Studios. Well now you can join him… virtually.

Walt Disney World has just made it easier for visitors to plan their visits in advance by releasing its Orlando theme parks in full 3D within Google Earth. All four Disney World parks (Magic Kingdom, Epcot, Hollywood Studios and MGM) as well as 22 hotels have been included. This virtual tour, created by Disney Parks and Resorts, is the largest corporate initiative ever within the Google Earth application.

There are over 1,500 3D models that went into making this effort happen. Creating these models required over 100,000 photos from 8 different photographers. Also visible while browsing the parks in 3D are attraction information, as well as plenty of user generated content like photographs, blogs and links to videos.

This has some pretty cool implications for families that want to plan their trips in advance. It also gives potential visitors a chance to check out the theme parks, and how expansive they are, prior to deciding if a Disney vacation is the place for them.

See Disney World on Google Earth!

Why Purchase Your Brand Name?

Wed, 25 Apr 2007 19:45:00 GMT

If you work for a company that is lucky enough to have your Paid Search marketing tactics under the same roof as your Natural Search Optimization, then you probably know that there are some incredible synergies between the two mediums. For some keywords, it makes sense to lower or cancel your paid search spend when your preferred landing page rises in natural search rankings. This is especially true if paid search competition is light.

So why do so many companies continue to buy their brand names, even if they are ranking at the top of natural search listings?

To go even further, many big brands have filed trademark protection with the major search engines, effectively removing their competitors from bidding on their trademarked terms. So why do they continue to buy their name?

There are many answers to this question, and only a couple are below.

Paid search allows you to speak directly to your potential customers through ad copy. You can tell them exactly what you want them to know at that point in time, and place them exactly where you want them on your website.

SEO does not allow you the ability to control your message; only to influence it.

If a search engine user is looking up information about “Disney World” online, the home page may come up first, but what if I want that searcher to know about our special offer on “Disney World Tickets”?

There has also been plenty of research released that shows there is an incremental lift in click through rates when you obtain the top paid and natural listing on a search results page. Having your listing in both spots is a way to reinforce to search engine users that you really are the best answer to their question.

I came across a few examples of companies that have recognized these benefits recently as I was browsing the web.

Nike ranks number one in natural search, but probably noticed that their natural search listing lacked a very good call to action, other then “we are Nike”. They utilized their paid search ads to introduce searchers to their Nike+ line of shoes, directing them deeper into the website. The ad copy was “Get the new Nike+ Shoes for an Interactive Workout with your I-Pod”.

Jeep ranked number one in natural search with an “Official Home of Jeep” message. Their paid listing however was much more compelling: “See the new 2007 Jeep Collection Today”.

Pizza Hut is currently utilizing it’s paid search ads to let consumers know that they have expanded their lunch offerings. If I was interested in just learning more about the company, I would probably move on and click the natural search ad.

And let’s not forget competition as another important reason to own your brand. All of the companies above had at least 9 other marketers buying their brand names. If the company didn’t protect their brand, the risk of losing clicks, and the top spot on the search results page, increases greatly.