Why Does Google Analytics Report Values That Are So Different Than Ad Servers and Site Analytics Counts?

I had a client call me today and ask, “Why do we need to use an ad server when we have Google Analytics?”  Actually, he was redirecting a client’s question and asking for ammunition, but the question was legitimate.  He wasn’t talking about ad serving per se, but reporting.


 


Google analytics enables an advertiser to measure counts like an ad server and site analytics software and yet the measurements yield very different results.  In fact, the results are always lower, enabling the client to conclude that they must be more accurate.


 


Hold the phones!  Wasn’t it just like up to four months ago that Google was guilty of 20-30% click fraud rates with its Adwords program?  Advertisers started complaining and dollars started shifting away from Google and towards MSN and Yahoo!.  When Marketing Pilgram broke the story in December 2006 the word “Click Fraud” was changed to “Invalid Clicks” and Google demonstrated that they had miraculously instituted a change that resulted in the double-digit error rates dropping to less than 2%.


 


So why are we trusting Google’s counts?  It would seem that Google has simply put some major filtering in place to cover their asses so that Advertisers aren’t getting overcharged any longer.  And now they are the more accurate source of counts? 


 


In the end, however, what Google counts and what the rest of the industry counts when it comes to banner advertising and site-side traffic has to be inherently different.  Just as it is with click-traffic.


 


According to Google: “Different web analytics products may use a variety of methods to track visits to your web site. Therefore, it is normal to see discrepancies between reports created by various products. However, we generally believe that the best way to think of metrics across different web analytics programs is to think in terms of trends, as opposed to numbers by themselves.”


 


Google presents that their tracking methods can introduce a difference in reporting values: Cookie-based tracking vs. IP + User Agent tracking.


 


Cookie-based tracking relies on a browser setting the cookie. If cookies are disabled, cookie-based analytics programs (such as Google Analytics) will not count the visit.


 


IP + User Agent tracking typically uses log file analysis for its data. Ad servers rely on this methodology. 


 


Another discrepancy that Google talks about is a resultant of first party vs. third party cookies.  “Because 3rd party cookies are set by a source other than the website being visited, they’re often blocked by browsers and security software. Google Analytics uses 1st party cookies.” 


 


Ad servers use third party cookies and therefore these may be getting blocked by Google Analytics.  That would represent a huge discrepancy between the ad server counts and the Google counts.


 


So Google and ad server and site analytics do it differently.  That makes sense.  Now back to my client’s question.  Why would you use an ad server when you have Google Analytics? 


 


For one thing, Google limits a site visit per user to one time every 30 minutes.  Ad servers, by comparison, would not filter such behavior, but would recognize the fact it is a unique visitor (using a cookie) coming to the page more than once.  So impressions would be counted separately from unique impressions.  Google would simply filter the multiple impressions out and give the unique impression.


 


My suggestion to my client was to convince their client to deploy a site analytics toolset so that there would be two third-party validations in place to offset Google.  People don’t seem to get the idea that even Google is proposing that “…the best way to think of metrics across different web analytics programs is to think in terms of trends.”  Especially when it comes to Google.  You get what you pay for … and you don’t pay for Google analytics.


 


Here is something else that we discussed.  If you synchronize your ad server and your site analytics you will get accurate – or actually identical counts.  For example, deploy a DirectServe™ Technology using First Party ad serving with a WebSideStory first party cookie and you will have a seamless pass through of data.  Impressions and clicks will go through to the site and the site will read the ad serving data – actually WWS will receive the data using the first party cookie – and the reporting will match up perfectly. 


 


Remember, DirectServe™ is a patent-pending capability of TruEffect and we are partnered with WebSideStory to implement this kind of solution so that was a plug.  But seriously if you want to put together the pieces this is how to do it.


 


The client also asked about bid optimization.  Hmm, another variable.  Love it.  Well WSS has Bid Opp and so that can easily be brought into the picture as well.  Using a first party cookie, the ad server can lay the cookie down on the user when they click on the keyword and associate the keyword and search engine with that user.  If that user is already carrying the cookie from the client, the ad server can add to the cookie the search variables that regenerated the visit.  Then the site analytics software can receive the data using the first party cookie.  Done.


 


Re-targeting is a wonderfully versatile capability.  My favorite part of the conversation was when he said, “oh…we’re already testing targeting with TACODA.”  Love it.  Obviously we discussed the event-based targeting aspects of TACODA and how it is based on anonymous occurrences.  He agreed that while the solution works well it is limited to their network and does not have the ability to leverage client data like what DirectServe™ has to offer – site agnostic, web-wide capacity that leverages client knowledge about customers for re-targeting.  He got it and agreed that we were talking about complimentary solutions … for now.


 


Anyway, back to Google.  I think it is key to understand that ad serving has all of its benefits from the perspective of campaign management.  And site analytics has all of its benefits from web site trafficking, modeling and analysis.  But what was at conflict here was ignorance of a client’s client.  If a client is going to use Google, they need to be educated as to why they are using and what they are using it for.  It’s kind of like using fuzzy glasses to read a book.  Or a better example is using your hand to feel your kid’s forehead to see if they have a fever.  It is a trending tool that gives you a relative indication, not an exact measurement. 


 


Google Analytics is great for the advertiser who wants to log in at 3am and see what’s happening.  I do that sometimes with my blog when I write a particularly contentious article – just to see if it’s triggered some reactions.  But my server logs are far more accurate than Google Analytics.  The counts are always 30%+ off.  Same with ad server reports and site analytics. 


 


Educate your client with the tools that are going to demonstrate real accuracy.  Use sales reports and revenue reports – post-click analysis – to demonstrate further discrepancies that translate into real value to the client.  It is possible to show the client where the diversion points in the direction of the ad server and site analytics favor.  Go the extra step and you will prevail.  If you have more than one client that will bring this up, prepare a document that you can use over and over again.  This problem is not going away soon.


 


Reactionary with Insight.

_uacct = “UA-980395-1”;
urchinTracker();

Advertisements