Last week I participated in a web analytics panel at the Search Engine Strategies conference in San Jose, California. One of the audience members asked why campaign data reported from their search engine didn’t align with data from their web analytics tool. For example, why would Google report 1,000 clicks for the keyword “ipod”, when the web analytics package reports 800? Or why would Google report 800 clicks when the web analytics package repors 1,000?
As some of you know, this isn’t a search engine marketing phenomenon. Marketers that use email and banner ads often see similar disparities from those systems and web analytics…
So what’s the deal? As I explained at SES, it is important to understand that these two systems are measuring fundamentally different things and are affected by different measurement methodologies. In the example above, Google is measuring what I call pre-site activity. Every time a keyword is clicked on Google.com, Google registers a click — irrespective of whether the visitor ever lands on the destination website. By contrast, the analytics package measures a click when the visitor *arrives* at the website itself. In other words, the web analytics package is measuring site-side activity.
This may seem like a minor difference, but it is not. According to Google (and Yahoo) all keyword clicks are scrutinized in several ways before they are reported by those systems. For example, Google highlights that they use “3 powerful tools for protecting your clicks” — namely, detection and filtering, monitoring, and human expertise. If a click does not pass these thru these “tools”, it is not reported back to you — even if the visitor continues successfully to your site. Yet, to your web analytics package, a click is a click. If a visitor arrives at your site from a keyword search, that click is recorded.
But that’s not all. Once on your site, if a visitor hits the Back button, and then returns to your site again, that means 2 clicks in your analytics package but only 1 in Google. Hit the Refresh button, that’s another click in the analytics package, but not in Google. Bookmark the landing page on the website and return via that bookmark, that’s another click in the analytics package but not in Google.
By virtue of these two different methodologies, you can quickly get to wide disparities in keyword reporting. At the campaign and ad group level this is much less obvious than at the keyword level. I’ve seen anywhere from 5%-90% differences in click reporting from a search engine to a web analytics package at the keyword level.
There are many other factors that can also contribute to differences in reporting. To maximize your marketing success, it is imperative that you understand these factors and adjust for them accordingly. If you do not, you could easily misinterpret campaign performance and limit your marketing ROI.
In my next blog, I’ll discuss what you can actually do to reconcile these reporting differences and maximize your marketing ROI.