Frustrated by those (not provided) results that keep showing up in your Google Analytics search metrics and how they seem to be an increasing percentage of your search term results each month? This is an important question for anyone that is monitoring search traffic and tracking the impact of your site’s search engine optimization (SEO). Understanding the issue, what it does and does not impact, and how to adjust your analysis as a result, is the best way that a business can respond to the change, as is the case with most Google updates.
First things first: what are (not provided) results, and why do they matter? When looking in Google Analytics (GA) to see what keywords are driving organic search traffic to your site (you can do this by going to Traffic Sources > Sources > Search > Organic), most sites are seeing (not provided) at the top of the list. This is a result of an October 2011 change by Google, which encrypted search queries for anyone logged into a Google account (Gmail, Google +, YouTube etc.) while performing a search. For example, if a logged in Google user searches for “striped shirts” in Google and your site shows up in the search listings, and the user clicks onto your site, this will show as (not provided) in your organic search traffic keyword list, rather than showing as “striped shirts”. If you have worked hard on your site’s SEO and want to really understand what types of keywords people are using to get to your site, this is a frustrating twist as it limits the ability to quantify results.
Another problem is that this change makes it difficult to differentiate between branded and non-branded search results. When looking at search traffic, it is important to differentiate between organic searches where the searcher was looking for your company (I.e. they searched for “Bob’s Shirt Company” – branded searches) and those searches that are subject-based and bringing new eyes to your company (I.e. searched “striped shirts” – non-branded searches) The increase of (not provided) results makes it harder to define the ratio of branded to non branded organic search traffic to your site. Whether you are doing SEO on your own or have hired an outside firm to help, this is a critical metric. Non-branded search traffic is often full of new visitors, whereas branded search traffic is usually comprised of visitors who already know about and were looking for your specific company. Successful SEO will drive non-branded search traffic to your site, while maintaining and building on branded search traffic.
So, what should you do about the (non-provided) search results? First, it is important to understand what is and is not impacted by this change. While a subsection of your analytics are affected by the change, the actual search traffic and organic search analytics are not.
So, now it is time to think statistically. Remember ‘sample sizes’ and ‘margins of error’ from statistics class? Same rules apply – removing the (not provided) results from keyword results still leaves us with a large, very statistically relevant sample to calculate branded vs. non-branded traffic with a relatively low margin of error. We’ll spare you the spreadsheet of calculations for the purposes of this article, but, put simply, if you remove (not provided) results from your initial branded vs. non-branded visitor and, where relevant, revenue numbers, you can find your branded: non-branded ratio. Extrapolate this ratio to the (not provided) results and then merge the two for a strong approximation of your total visitor and/or revenue splits.
While no one knows Google’s next move, all signs point toward an increasing number of (not provided) results in the future. Firefox recently announced that they are on track to use “secure search” as a default, and it would not be surprising to see others follow suit. It’s just one more moving part in the world of search analytics. Successful SEO will still be successful SEO; the key is to be armed with the knowledge, tactics, and analytics to address the ever-changing landscape of search engines and SEO.