Raven Tools SEO Tools


Google “Not Provided” Analytics: A Workaround


October 22nd, 2012

google analytics not providedBy Samuel J. Scott

Ever since Google blocked the organic keywords in Google Analytics used by web searchers to find your websites while logged into their Google accounts a year ago, the amount of “not provided” data has increased — to the frustration of many SEO analysts. Here, I will provide two examples from this website and another of mine to show the effect on online marketers and then give a few workarounds that you might want to consider.

Of course, I write this with the caveat that nothing will return the data to being completely visible — the key is to find creative workarounds that make sense and, when applicable, are able to be explained to your company or clients. (You may know what you are doing — but if your boss or whoever signs your clients’ checks does not understand what you are doing, then it is likely all for nothing.)

First, a snapshot of this site’s organic traffic over the past three months:

google analytics not provided analysis

As we can see, 48 percent of the organic visits here have been from people who searched Google while logged into their accounts.

Next, an organic snapshot from my Buffy the Vampire Slayer Online website over the past three months:

google analytics not provided analysis

For this website, the amount was 19 percent.

It is important to consider the target audience when analyzing the effect that Google’s “not provided” keywords has been having. For reasons that are likely obvious, people searching for “seo software” (online marketers like ourselves) are much more likely to use Google’s plethora of products and be logged into their accounts all or most of the time. The teenage and twenty-something girls who make up much of Buffy’s fanbase are, well, not. If your customers are senior citizens who may still have Hotmail e-mail accounts, then “not provided” will probably be less of an issue than if you sell to website developers.

Regardless of how much this analytics change will affect you, here is a workaround that I have used. First, change the Advanced Segments setting to “non-paid search traffic”:

 

Then, go to Landing Pages in the menu:

The list you will see is the first pages on the website at which people arrived through organic search. Although the information is less precise, you can get a good idea of how the keywords each page targets are ranking and sending traffic.

Still, the issue is slightly different when you want to get a better picture of which organic keywords are sending conversions (if you have something like a sign-up form) or transactions (if you have an e-commerce site) specifically in light of the “not provided” effect. First, select “Visits with Conversions” or “Visits with Transactions” in Advanced Segments as described above. Then click Organic Traffic in the left-hand menu: Traffic Sources > Sources > Search > Organic. You will see a screen like this:

This is a method I used once for a client at my company. The client wanted to know how many total organic conversions were coming from non-branded search (people searching via terms that did not include the company name), so here is how I did the estimation (the data is based on my Buffy site as an example):

  • 336 visits with conversions — 82 not provided and 252 provided
  • “buffy the vampire slayer online” (my brand) had 5 visits
  • 5 visits out of the 252 known-keyword visits were branded — so 98% of the known traffic was non-branded
  • 98% of the 82 non-provided terms — so 80 of those were non-branded terms as well
  • 5 + 82 = an estimated 87 conversions from non-branded organic traffic

Of course, this is based on the premise that the percentage of traffic coming from branded and non-branded traffic will be the same for both available and not-available keywords. Still, it seems to be a reasonable assumption — and the client agreed.

You can do something similar for the brand name of any company or website. Say that the name is “widgets.” Calculate the percentage of organic keywords that include “widgets” (branded ones) out of the available keywords. Then apply the same percentage to the not-available keywords as described above. You’ll have another good indicator of the success of your organic SEO campaigns.

Still, there are other ways to play with Google Analytics in light of the not-available data. I found some interesting articles by Carrie Hill at Search England Land, Bluegrass, Barry Adams at State of Search, Ben Goodsell at Search Engine Watch, Mike Pantoliano at Distilled, Adam Kerr at SEOmoz, and Avinash Kaushnik. I am sure that you will want to read their thoughts as well.

Samuel J. Scott, a former journalist in Boston turned Internet marketer in Israel, is the founder and publisher of My SEO Software and Director of Digital Marketing and Communications and SEO Team Leader at The Cline Group. You can follow him at Google+, LinkedIn, Facebook, and Twitter. His views here and elsewhere do not necessarily reflect those of his company and clients.