So you’re bidding on a keyword. You have the keyword in the title, description, display URL and destination URL. Your click through rate is over 5% and Google still says your quality score is poor? Your landing page has links to your privacy policy and your content is useful and relevant. So what’s the deal?
First of all it is important to understand that the Google AdWords quality score bot is not a person. It is a robot. It is unable to understand the context of text on a web page the way a human can.
The quality score bot does a reasonably good job most of the time, but sometimes we need to point them it in right direction. To do that, we need to structure our web pages semantically and structure our documents logically.
There are two very useful and free tools at our disposal when troubleshooting landing page quality scores. These are the W3 Semantic Extractor and the Google site related keyword tool. What better way to get information about what Google thinks your site is about then using a tool designed by Google to figure out exactly what your site is about?
What’s the problem?
Let me first start with an example of a landing page that does not correspond or relate to the keywords and ad text used.
The Keyword is: “Directory Submit”
Here is the ad:
Here is the Landing Page:
They appear to be a perfect match right? Wrong! After 1 day of running this particular keyword/ad with this particular landing page, our quality score went to “POOR”. The CTR was was 7%. The words “directory” and “submit” appear numerous times on the landing page too.
When we run the site through the semantic extractor, we can see that this page does not appear to outline anything to do with a “directory submit”.
When we run the site through the site related keyword tool, our keyword does not even show up in the list of terms Google thinks is related to the page!
So what can we do?
The above example may seem familiar to you. If you have been “Google slapped”, the first thing you might do is complain that your site is obviously relevant and that Google is just out to squeeze you for every drop of cash you are willing to part with. First of all lets look at an example of what a good site/ad/quality score is:
Keyword(s) targeted: “Search Engine Marketing”
Here is the ad:
Here is the landing page:
Here are the semantic extractor results for that page:
And here is the Google site keyword tool results:
This particular ad/keyword/landing page combination has our quality score at “GREAT” and our minimum bids at $0.01. See the difference?
We need make sure that we have our site marked up correctly. Use headings correctly , utilize at least the h1, h2 and h3 tags and make sure the content of each is related in the semantic extractor outline. So if your site sells shoes, your landing page heading tags would be made up something like:
h1 : buy red shoes
h2: best place to buy boots
h3: why red shoes are better than blue boots
This not only helps Google AdWords identify what your content is about for organic search rankings it also helps the quality score bot understand that your landing page is indeed related to the keywords you are bidding on.
While the above steps and tools are clearly not the only way to help Google understand the content of your site, they are free and a great place to start. The next time you get “Google slapped”, make sure you don’t deserve it.
Vote for this post : 2
or Buzz it at Yahoo :
Comments
22 responses so far ↓
Targ8.com on Aug 2, 2007 at 2:57 pm
What code do you need to use to mark up your pages using h1, h2 and h3 tags? Have you got an example?
Patrick Schaber on Aug 2, 2007 at 3:08 pm
Dave,
Great work on this! I have not had that problem, but I bookmarked this in case I do. I’d also be interested in the code tip as Targ8.com requested.
Thanks,
Pat
Kris from Florida on Aug 2, 2007 at 9:47 pm
Good tips, thanks for sharing . Stumbled this for future reference .
Gidseo on Aug 3, 2007 at 2:41 am
Thank you for a great post. I’m off to run a few tests…
Suki on Aug 3, 2007 at 2:55 am
Thats how a good Seo Consultant should work.. nothing new but itll help newbies stop bugging us “profesionals” with bad optimized websites.
Dave Davis on Aug 3, 2007 at 9:12 am
Thanks for the comments guys.
There is no sample code required, just standard:
Main Keyword
Main Keyword and something related to keyword
Get Main Keyword at rock bottom prices
And possibly
synonym of Main Keyword from the keyword tool
Hope this helps.
Arnold M. on Aug 3, 2007 at 10:59 am
Very good article! One thing about the heading tags: Use only one H1 tag per page, and do not break the hierarchy, H1 should be followed by H2, H2 can be followed by either H2 or H3, do not jump directly to H3 from H1.
Tomche on Aug 4, 2007 at 1:38 am
No wonder I was getting poor click rates. Got to start working!
Universos Virtuais on Aug 4, 2007 at 4:33 pm
Thank you for the nice tip. I am trying to learn about SEO, there isn´t much about SEO in portuguese, so, i am glad i find your article.
Have a great weekend!
Neu Kunden on Aug 5, 2007 at 1:32 pm
Sometimes it takes a differnt approach to discover the obvious. Thanks a lot!
Brian Turner on Aug 7, 2007 at 3:53 pm
Thanks for the heads up on the tools - will definitely be using the semantic extractor tool when I get into PPC over the coming months. :)
John S. Britsios on Aug 7, 2007 at 5:57 pm
Great article Dave, even if I follow all above already since years. But I sure will share this excellent article on my blogs and my forums. Keep up the excellent work!
Billy on Aug 8, 2007 at 9:58 pm
This is great. One of my clients had issues with quality score where Adwords were charging $10 per click, and they didn’t know what to do.
craig on Aug 13, 2007 at 5:42 am
Hi Dave,
Wonder if you may be able to point me on the right with this.
I’m new to this and trying to get some ppc up and running. I’m having problems getting a decent quality score.
I used the semantic extractor you mentioned in your post on two pages of my blog and it came back with the folllowing code related problems.
Do you think that these are litterally code errors preventing it from completing the scan of the pages?
If so that would I assume also prevent googles bot from ranking the page?
I haven’t written the blog in code form only used the visual writting tool of wordpress .
Your thoughts would be appreciated.
semantic results:
Using org.apache.xerces.parsers.SAXParser
Exception net.sf.saxon.trans.DynamicError: org.xml.sax.SAXParseException: The element type “font” must be terminated by the matching end-tag “”.
org.xml.sax.SAXParseException: The element type “font” must be terminated by the matching end-tag “”.
and on another page.
Using org.apache.xerces.parsers.SAXParser
Exception net.sf.saxon.trans.DynamicError: org.xml.sax.SAXParseException: The element type “span” must be terminated by the matching end-tag “”.
org.xml.sax.SAXParseException: The element type “span” must be terminated by the matching end-tag “”.
BrianE on Aug 18, 2007 at 4:57 pm
Thanks for the free semantic extractor tool. I’ve always used h1 /h2 tags but can see they need improvement. I have several main keywords , some are the plural version of the main ones. Should all keywords be included/distributed among the different h1/h2/h3 tags and/or should seperate landing pages be used?
Dave Davis on Aug 19, 2007 at 4:12 am
Craig,
The error that you are getting is simply that your page is not coded semantically . Of course this does not matter to Google, but in my opinion is essential when troubleshooting for yourself.
Brian, the tool is not ours. It was created by W3C. Keywords SHOULD be included in your HX tags but not spammy. As I noted, something related. Use the Google keyword tool to find related phrases and make sure Google thinks they are related to each other. That gives you the best chance.
If possible, use a separate landing page too for each ad group.
Wendy Piersall on Aug 20, 2007 at 9:22 pm
I’m not sure I buy it. I put up a landing page with some blogging products on it. I custom wrote a tutorial for the page about how to set up and install wordpress with your hosting company. I put “WordPress installation” not only in the title tags, but in my H2 tags.
I still got hit with $5 clicks for low quality scores on the term “WordPress installation”. It was relevant to both the products I was promoting as well as the content.
Now, if you can actually explain this one to me, I will PayPal you $20, just because it would be THAT worth it to me to know why the #$% Google seems to want to milk me dry! ;)
Allan Gardyne on Aug 31, 2007 at 3:53 am
Wendy, Here’s something which may help. Jeremy Palmer has just published a 10-page report (PDF file, free) on what you need to do to get a good Quality Score - http://www.quityourdayjob.com/qualityscore.pdf
He goes into a LOT of detail.
(I’m an affiliate but that’s not an affiliate link.)
SEO Ibiza on Feb 1, 2008 at 5:01 pm
excellent info thanks so much for this. dont kn ow how we havent seen your blog before now
SEO mkguyver on Apr 3, 2008 at 12:47 pm
Nice, but using superlatives is not a good thing. The best shoes for example. Also the keyword tool reference is out dated.
Seb on Apr 6, 2008 at 3:45 pm
Your findings is what I initially thought when I got slapped, so I used Google’s keyword tool to add keywords that Google recommend from the landing URL. After adding these to in to adwords, Google then gave these keywords stupid minimum bids like the rest of them.
Been using Google adwords for some years now, and it used be quite effective tool, but without harding changing anything over the years other than minor tweaks, I’m lucky to get 10 impressions.
I feel I’m going to be deleting the adword account very soon - its just not worth the effort.
iphone on May 13, 2008 at 12:57 pm
OK I’m a newbie and have a question on usage of the Semantic Data Extractor. I put in the URL I want to analyze and I got back this error:
Using org.apache.xerces.parsers.SAXParser
Exception net.sf.saxon.trans.DynamicError: org.xml.sax.SAXParseException: Content is not allowed in prolog.
org.xml.sax.SAXParseException: Content is not allowed in prolog.
Is there a format I need to use to use this tool?
Thanks,
Leave a Comment