How to Turn the Dreaded (not provided) Into Useful Data
I was lucky enough to attend Dara’s Advanced Google Analytics (GA) training course on Thursday at #brightonSEO. The issue of Google’s (not provided) came up within the first five minutes of the hour. Everyone had a deathly, depressed look on their face. Including me.
Perhaps the biggest take out of Thursday’s session was a ‘trick’ to make the most out of your (not provided) data. For those who aren’t familiar with GA, I’ll start with what (not provided) is. Sometime around two years ago, Google announced one of the biggest updates that website owners may have heard since domain names started costing them money back in 1995. When searches are made on Google pages where the users are logged in (i.e. a secure Google page), then the keyword column in Analytics will return a (not provided). Google have encrypted this for the sake of user privacy.
This encryption is an obvious problem for those of us who use GA on a regular basis. There is a huge chunk, sometimes up to 60%, of keywords showing up as (not provided). NotProvidedCount.com shows that this trend is going nowhere but up. Agencies are unhappy as their efforts are getting increasingly difficult to measure and clients are unhappy because their data is getting increasingly ambiguous. Unfortunately, Dara couldn’t tell us how to decrypt the security set by Google. What he did do, though, is tell us how to turn (not provided) into useful, measurable data.
Setting profile filters
Personally I’m not a big fan of setting custom filters, but I’ll jump on anything that hints at getting rid of that hideous (not provided) from my analytics data. This appends every (not provided) with the landing page in the keyword column. This can then tell us more about what kind of keywords visitors are using.
The first step to take before setting the filter is using a separate GA profile. This is best practice whenever you are planning on messing with profile filters so that test data doesn’t take away or change real data. Just follow the steps in the link above to do this.
Next, you want to create a filter on the new profile view you just created. Do this by going to Admin and select Filters from the View (Profiles) column and select + New Filter as shown below.
You want to make a custom, advanced filter with the options shown below.
The third row (Output To –> Constructor) can be changed to what you prefer. The “np – $B1” tells GA to show a “np – /example-blog” if the page your visitor landed on was http://www.examplesite.com/example-blog. You can change the “np” part to anything you prefer.
You can see the added value this filter brings for websites that have large amounts of (not provided) data. Of course, this method can get a bit messy. Those who are further interested can apply another filter that replaces the “np – landing page” with the actual keyword from that landing page. If you want to be more keyword-specific, you can export the data into an excel sheet where you have all the keyword information for specific landing pages and organize it there. This way, you will have a better understanding on the value of your (not provided) data.