The ROI Revolution Blog
State Popularity: the latest addition to the GARE
February 22, 2010
We have another new addition to the Google Analytics Report Enhancer, thanks to Ophir Prusak of Google Analytics Authorized Consulting firm POP. This metric helps to interpret the significance of visit counts at the US State level. You can hear the rest of the story by reading Ophir’s excellent post on the metric.
Now that State Popularity has joined the GARE family, it’s a great time to download the latest version of the Report Enhancer. Here are the steps:
- Get Firefox
- Get Greasemonkey
- Get the GARE
In addition to the new metric, I’ve also been able to improve the way additional metrics are added to tables, including better sorting and handling of advanced segments and compare to past.
So how is State Popularity calculated anyway? I’m glad you asked!
First, the number of visits for each state is divided by the population of that state, according to the latest estimates. That’s the easy part. What you’re left with are some very small numbers that don’t really give a good idea of significance.
The obvious way out of this is to multiple each ratio by the same number to make them more human friendly. I decided that rather than picking a static number, like 1,000 or 100,000, I would multiply each ratio by a number so that the total of all adjusted ratios would be equal to the number of visits. To do this, I divided each ratio by the sum of all ratios, then multiplied by the total number of visits.
The main benefit to multiplying the ratios by this number is that you can tell whether you are getting a high number of visits even from a smaller state by comparing the State Popularity number to the number of visits.
For example, if the State Popularity is higher than the number of visits for a particular state, then you know that you’re getting higher than average visits from that state based on its population. If it’s lower, then the number of visits from that state is less significant.
The problem with this method, from a Google Analytics standpoint, is that it’s difficult to calculate the sum of all the ratios unless they’re visible, that is, at least 50 rows are showing (though usually more because of D.C. and (not set)). To handle this, I had to make my own requests for the data that included all of the rows and pre-calculate the totals. The map overlay also has its own peculiarities that make it difficult to make table additions persistent.
What makes this significant is that it in the past, I could only create new metrics based on metrics in a single table. But now, theoretically, I should be able to calculate metrics based on more than one table.
I’ll probably be able to think of some good use of this technique before too long, but if someone else has an idea of what metric they might like to see, let me know! You never know, your new metric may be the next addition to the GARE.