Not Signed In Join the Community
Microsoft Advertising Community
Your adCenter PPC & Web Analytics Reports Don’t Match -

Microsoft Advertising Blog

Bookmark and Share

Your adCenter PPC & Web Analytics Reports Don’t Match

posted Mon, Mar 02 2009

And Why They Never Will

 

Why don’t the numbers from adCenter and my web analytics match? It’s a common question. The underlying assumption, and anxiety, is that if the numbers don’t match, one of them must be wrong. After all, 2+2=4 no matter who’s reporting the results. That’s true as long as you’re comparing apples to apples.

In practice, you’re likely comparing
apples to persimmons.

The first thing to do is relax, peel a banana, and recognize that the web is a jungle of fuzzy metrics. Each web analytics provider deals with numerous, aggregated metrics that may be defined differently than their competitors’. The metrics themselves are also dependent upon the collection methodology. In practice, you’re likely comparing apples to persimmons.

A rule of thumb — never compare the absolute numbers from one web analytics product against another.  Instead, look for trends that mirror or contradict each other and further examine the underlying definitions.

There are several reasons why data reported from adCenter —  or any search engine, for that matter — won’t tally with your web analytics, whatever analytics you’re using.

  • Missing or improperly configured tracking URLs
  • Missing landing page tags
  • Clicks versus cookies
  • Broadly matched search keywords
  • Invalid clicks
  • Recency
  • Redirects

  

Tracking URLs

Tracking parameters, also known as name-value pairs, pass data about your adCenter campaign to your web analytics. They typically follow a syntax such as:

http://www.example.com/landingpage.html?source=adCenter

If your destination URL doesn’t include a parameter to identify the referrer as an adCenter campaign, your web analytics will likely log it as an organic referral from Live Search rather than a paid ad from adCenter. The URL can also be incorrectly coded. Coding errors become more likely as the URL gets longer. Given the number of dynamic values adCenter allows you to append to a destination URL, URLs can get very long indeed.

Your adCenter report will attribute the click to the appropriate campaign even if the tracking parameters are miscoded. Your web analytics, however, won’t.

PPC landing pages seem
to sprout like mushrooms
after a heavy rain.

Landing Page Tags

Most web analytic products use page tags to collect data. The tag is a snippet of JavaScript that fires on page load and sends data to remote collection servers connected by the Internet. It may seem self-evident but without the snippet of code, no data is collected. On large web sites it’s difficult to ensure that all pages are properly tagged. That’s especially true of PPC landing pages that seem to sprout like mushrooms after a heavy rain. Both URL parameters and page tag need to be functioning properly in order to properly attribute the referral.

Clicks vs. Cookies

When discussing tracking URLs I briefly mentioned that adCenter reports on clicks. Your web analytics, however, depends upon cookies. A cookie placed on the visitors’ web browser is typically the only way for analytics products to remember a visitor from one page to the next. But cookies can be problematic. Some visitors will set their browsers to reject cookies. Or the JavaScript that sets the cookie may never fire. Most analytic providers recommend placing the JavaScript tag at the very bottom of the page. If the visitor quickly navigates away from the page the script may never load. As well, other JavaScript on the page may trigger an error which can stop the JavaScript engine from executing additional script lower on the page. Your analytics tag may become the victim of collateral damage.

Broad Match

When you’re broadly matching keywords to search phrases, there can be a lot of latitude between the keyword you’re bidding on and the query phrase entered by the searcher. So, adCenter would record credit the click to the keyword in your campaign but your analytics might credit the actual query phrase. You can pass additional data as a parameter in your destination URL to help remove the ambiguity. The following dynamic values are available within adCenter.

QueryString: The actual search phrase entered by the searcher.
OrderItemID: The keyword ID.
Keyword: The keyword as it appears in your campaign.
MatchType: The type of match that triggered your ad (exact, phrase, broad, content).
AdID: The ad ID.

In all cases the syntax is Name={Value}. To continue our first example:

http://www.example.com/landingpage.html?Query={QueryString}&Keyword={Keyword}&KeywordID={OrderItemID}

Invalid Clicks

Microsoft protects you against invalid activity by monitoring your campaign and removing suspicious clicks from your reports and your bill. Your web analytics, however, blithely record all activity to the landing page, whether suspicious or not.

Search engine data for the current day
should always be addressed gingerly.

Recency

Web analytics data is relatively static. Once recorded, there’s not a lot of adjustment required. Traffic figures from adCenter may need to be adjusted for suspicious activity or invalid clicks. It may take several hours before the data resolves. Search engine data for the current day should always be addressed gingerly.

Redirects

Many advertisers use a third party to track their display and search ads. The destination URL leads to the third party site such as DoubleClick, Efficient Frontier, Atlas, etc. The click is registered and then redirected to the landing page on the advertiser’s site. A poorly configured destination URL can result in a failed redirect and a lost prospect. It’s similar to the tracking URL issue mentioned earlier but with additional complexity.

Regards Charles

PS: Remember to get all the latest news by becoming a fan on Facebook or following us on Twitter too!

Comments

  • Tue, Mar 03 2009 03:33AM

    so can't we track them correctly well i use google analytics for tracking clicks

  • Wed, Mar 04 2009 05:56AM

    That a great summary Charles

    For a full in depth discussion on web analytics accuracy issues (vendor agnostic), have a look at the following  whitepaper: www.advanced-web-metrics.com/accuracy-whitepaper

    I would be interested in your comments.

    Best regards, Brian

  • Wed, Mar 04 2009 11:16AM

    PPC Clicks, I wasn't suggesting that you can't track your PPC campaigns in Google Analytics or Yahoo Analytics or any of a number of excellent analytic solutions. I'm simply warning you that the stats reported by adCenter won't exactly match those reported by Google Analytics. Neither are wrong, they're just different perspectives of the data. They're not apples to apples. That's true of whatever search engine and whatever analytics you're using.

  • Wed, Mar 04 2009 12:16PM
    kva

    So how do we fix it ?

  • Thu, Mar 05 2009 06:52AM

    Brian, thanks for the link to your whitepaper. It's a great resource on the subject of data collection for web analytics and the challenges inherent in comparing data from different vendors. It also has some effective recommendations on how to deal with those challenges.

    If you're unfamiliar with Brian Clifton, he's a recognized expert in analytics and the author of "Advanced Web Metrics with Google Analtyics," a book I highly recommend if you're using Google Analytics on your web site.

  • Mon, Mar 23 2009 07:55PM

    Thanks for the great write up. Do you have an opinon about which analytics package does the best job of integrating PPC data from the big three (LIve, Yahoo and Google)? I currently recommend Google to my clients because I have found it so easy to set up.

Leave a Comment

 
[optional]