When you launch a new website or page, what is your biggest clue that people are searching for your keywords? That’s right…your Google PageRank. Everyone wants to be the first result on the first page of those search engine results.
Unfortunately, rank tracking isn’t as straightforward as it seems.
A 4-Way Comparison
Dr. Peter J. Meyers of Moz.com compared three common rank tracking methods to see how the results differed. He took the top 500 queries from Moz.com’s Google Webmaster Tools for one month and tracked 4 different ways:
Browser with personalization (logged into Google account)
Browser anonymously (used secure search)
Web crawler (created webcrawler from MoxCast engine)
Google Webmaster Tools (accessed data from analytics)
Browser with Personalization
Dr. Meyers logged into his own Google account and entered each query manually into his browser (he used Chrome).
Browser with Secure Search
He ran the same queries but only after completely logging out of his Google Accounts and used the “incognito mode” that Chrome provides.
Moz.com modified their MozCast application, originally designed to help track changes in the Google algorithm, and used it as a stand-alone web crawler. They then set it to crawl the 500 queries and provide analytics.
Google Webmaster Tools
Dr. Meyers retrieved data for these queries for the same date as the previous three methods were used. The actual data used was from the “average position data” in Google Webmaster Tools (GWT). The resulting list was massaged to resemble the parameters used in the other searches.
All four ranking methods correlated to each other pretty well. A reminder from high school math class: r = 1 indicates perfect correlation and as the value descends from 1.0 to 0.0, the weaker the correlation.
Here are the correlation values (r values) between pairs of methods using those 500 queries (hint: anything above 0.9 is pretty darn good):
Web Crawler to:
Obviously, these results are only for correlations between two methods, not four. But Dr. Meyers and his team did some higher math and determined that, while Google Webmaster Tools agreed less with the other three methods, the difference was not statistically significant.
To you and me that means all four methods agreed well enough for comparison purposes; quibblers may argue amongst themselves.
Since all four methods agreed well enough for government work, the experiment didn’t really show any “right” way to determine how a page would rank. But Dr. Meyers picked apart the data and found a couple of interesting things:
Personalization of search results doesn’t always cause as big of an impact as we thought. Sometimes personalization resulted in a query ranking lower than expected.
Google Webmaster Tools, in some instances, appeared* to have a great deal of bias for certain search terms, ranking them differently from the other methods.
* I say “appeared” because of the lack of transparency in some of Google’s workings; Dr. Meyer was unable to determine the reason for the difference.
Moz.com concluded that any one of the four methods will get you in the ball park of knowing where your page ranks. And that’s all you really need to start improving your rankings. Anything more precise is a form of wheel-spinning that nobody really has time for.
As Dr. Meyers observed, “…don’t put too much faith in any one number.” Take what you have and run with it.