Receive new blog posts immediately direct to your email inbox!
The ROI Revolution Blog
« NCAA Tournament Contest Update | Main | Top 3 Google Analytics FAQ »
Testing AdWords Ads in Google Analytics
March 26, 2007
It's easy to set up multiple ads for an Ad Group within Google AdWords, but how can you tell which ad is performing at a higher level? You may be using Conversion Tracker to make your decision, but that only tells a small part of the story using a single metric, cost per conversion.
Google Analytics allows you to get the entire story about the ads that you are testing, using multiple metrics. Here's how to get that story:
You can access ad results at the campaign, ad group, and keyword level using Google Analytics. Simply use the AdWords Analysis Report, drill down to the appropriate level, and then open up the Analysis Options menu by clicking on the little purple button next to the appropriate record.
Select Cross-Segment Performance from the list, and for the field, select Content (shown below). You will see a list of all ads, arranged by title, associated with the campaign, ad group, or keyword, along with conversion data for each one of your goals and $/Visits. This will give you a better idea of how well your ads are faring against each other.
Sounds great, right? Well there is a slight problem to be aware of when testing in this manner. If you are using auto tagging within Google AdWords, then the ads will be identified by titles. Therefore, you can only use Google Analytics to test ads that have different titles.
If you want to test something different, such as the ad copy or the display URL, you would have to either change the title slightly (capitalization, punctuation, etc.), or you would have to use manual tags for these ads. If you decide to go that route, by sure that the utm_campaign, utm_term, utm_medium, and utm_source tags match your campaigns exactly, or else you will be causing more problems than you are fixing. Regardless, you would lose the ability to track at the ad group level, as there is no manual tag for ad group. For more information about manual tags, visit the Link Tagging 101 and Link Tagging 102 articles. You are better off using Conversion Tracker for tests where you would like the ad titles to be the same.
So, in summary:
1. Create the ads you wish to test in AdWords. Make sure the titles are at least slightly different.
2. Go to the AdWords Analysis report. Drill down to the campaign, ad group, or keyword you are testing.
3. Click on the little purple button to cross-segment a record by 'Content'.
4. Use the results to select the best ad.
Remember that you can use the utm_content manual tag for non-AdWords ads, and you would have to use a different report, like Overall A/B Testing, to find these results. Also, it's bad practice to run a test on your ads at the same time you are either changing bids on your keywords or testing the corresponding landing pages. Testing either of these at the same time you are testing ads will compromise your results.
Update (3/30/07): I just found out that Robbin Steif from Lunametrics wrote a different article concerning this subject last August that you may wish to check out.
Sponcered Link: Unlock the power of Google Analytics with ROI Revolution's "Six Weeks to Google Analytics Nirvana" online course.
Posted by Shawn Purtell, Analytics and Optimization Specialist at 1:59 PM
Permalink | Comments ( 5 ) | TrackBacks ( 0 )
Filed under: Analytics Basics , Improving Conversion , Reports
TrackBack
TrackBack URL for this entry:
/mt/mt-tb.cgi/211.
Comments
Hi Shawn,
I have a question you might be able to help me with (me as a curious GA person - not a competitor :-)
Wanting to track activity from a paid Google campaign - the following problem arises when one has an ad-group with multiple ad-creatives and multiple keywords, both able to contain individual landing pages (URLs).
It becomes problematic consolidating cost data as one either has to take an ad-creative or keyword attitude and when mixed in - where do you 'place' it. One can solve this by telling users to:
A)
ONLY use destination URLs at the Ad-Creative - as one then will be able to track and report on multiple ad-creatives and their performance
B)
ONLY use destination URLs at the Keyword - as one then will be able to track and report on a keyword level, but not which creative was used.
In the perfect scenario one would of course like to report on both and have clients mix it any way they want.
QUESTION to the GA EXPERT:
Am I missing something here? - is there a definition of the 'gclid' variable somewhere? - as I assume it holds all the information. Does GA do something that others can't do because of the integration to ADWORDS. OR is it just either or and not both.
Anyway.. Super blog you have here!
Cheers
Dennis R. Mortensen, COO at IndexTools
My Web Analytics Blog
Posted by: Dennis R. Mortensen at March 26, 2007 6:21 PM
Wow, there's a lot in that comment. I'll start be mentioning that the only reason I can see the problem you describe occurring is through poor Campaign organization. Maybe I'm misunderstanding the statement, but it sounds like you have ad groups that contain keywords that probably should not be grouped together.
It's good practice to maintain your ads at the ad group level, keeping the destination URL the same for these ads (unless you are testing landing pages, but that's another full article). This way, the ad group will only contain relevant keywords that all bring traffic to the same landing page.
By doing this, as you alluded to in your (A) suggestion, you will be able to have full analysis of your ads at the campaign, ad group, and keyword level. Again, it is important to remember that the gclid will use the ad title as the utm_content tag (in a manner of speaking), so you can only get separate Analytics data for ads with different titles. (Phew)
In answer to your question, the gclid is just an easy way for Google to tag your ads so that they are tracked properly. The only information they may contain that you would not otherwise get is at the ad group level, as there currently is no manual tag for that.
The thing that GA does best is track your Google AdWords campaigns, and the reason it is so good is because it can grab your cost data directly from AdWords (not using the gclid, but another method). This allows you full ROI calculation, which in our experience, is extremely handy to have.
I hope you found this article helpful!
Posted by: Shawn Purtell at March 27, 2007 8:17 AM
Hi Shawn,
Thanks a lot for your input and it seems like we are pretty much on the same page then - and that I AM NOT :-) - missing out on some secret GA information.
The challenge is then to "convince" clients to organize their campaigns appropriately... hmm - that is some task :-)
Cheers
... and off to Emetrics in London.
Dennis R. Mortensen, COO at IndexTools
Posted by: Dennis R. Mortensen at March 28, 2007 9:59 AM
Hello,
I have a small problem with the current javascript Google ads. I'd like to know if there's a way to put them in your site dynamically (PHP, Perl, etc.) instead of Javascript -- I thought this would provide faster loading for the user.
Thanks.
Posted by: MultiZ at May 25, 2007 12:03 PM
@MultiZ:
Hmm... It sounds like your question is not about ads, but about the Google Analytics script that goes on your site's pages.
You can dynamically insert the Google Analytics script on your pages using PHP, Perl, ASP, etc. I don't think this would have any positive effect on load times, and if anything the load time would increase because you are calling an additional series of functions that writes the GA code, which in turn still needs to run.
Hopefully I didn't misunderstand the question. If you're asking if the GA cookies can get created using PHP, then the answer is no.
The urchinTracker function used by Google is a JavaScript function.
I hope this helped.
