For an analyst, (not provided) is probably the most frustrating aspect of Google Analytics. We are seeing this keyword term more and more as Google protects the organic keyword search activity of its signed-in users. There may, however, be some light at the end of this particular tunnel.
The replacement of actual search terms with (not provided) was designed to protect the privacy of users signed in to Gmail, YouTube and other Google-owned websites. However, Google still releases any paid search data (i.e. AdWords) regardless of whether or not a user is signed in.
The Size Of The Problem
Originally, when Google made this change to the way we can track website usage data in October 2011, they stated that (not provided) would only affect a single-digit percentage of keyword data. The publicly available statistics, however, tell a vastly different story.
(not provided) Count has tracked the percentage of keywords not provided in Google Analytics for 60 websites since 30 October 2011 – a week after (not provided) went live. To date, 30% of the keyword traffic on these sites has been blocked and they currently project that we will have a keyword blackout by March 2019.
These numbers may, however, be an underestimate of the current impact of (not provided). A recent study by Optify sought to find a more accurate figures for just how much search traffic really is (not provided). They tracked 424 websites from November 2011 to September 2012 and saw the average rate of (not provided) keywords rise from 14% to a staggering 39% across the sites.
A whopping 81% of these websites saw over 30% of their keyword data lost to (not provided), while 17% of these websites lost over 50% of the data.
The Impact
But how does this actually impact our marketing strategies? At the very least, this hampers our ability to connect an SEO strategy to a goal conversion rate. This means that you don’t know whether an optimised keyword is useful, or how useful it actually is.
What You Can Do
Luckily for us, Avinash Kaushik, the Digital Marketing Evangelist, has written a blog post that details exactly how to shine a little light down the dark (not provided) hole. He describes and carefully demonstrates a simple five-step method of doing this and provides a downloadable custom segment to query your data.
The five steps are:
- Establish Macro Context – This is designed to allow you to track the month-on-month behaviour of (not provided) terms.
- Understand the performance profile of the (not provided) traffic – Here you track the percentage of goal completions by (not provided) traffic.
- Deep dive: Match up performance profile to Brand and Non-brand visits – Here you compare (not provided) traffic against brand and non-brand traffic to work out which profile it more closely resembles.
- Tentative conclusions. Why this seems so scary, but might not be (at least for now) – Here Kaushik explains how your non-brand tail of search queries can add up to a whole chunk of traffic.
- Additional awesomeness: Landing page keyword referral analysis – This shows you how to assess whether (not provided) traffic is entering specific pages more than your overall organic search traffic and what this means.
It seems that (not provided) is here to stay, no matter how hypocritical it may seem. There are, however, tools available that you can use to get around this frustrating keyword! Here are a few you can try, as well as some more very useful insights on the issue:
- http://www.practicalecommerce.com/articles/3798-Google-Analytics-Overcoming-Not-Provided-Keywords
- http://searchengineland.com/google-search-referrers-not-provided-139416
- http://marketingland.com/dark-google-search-terms-not-provided-one-year-later-24341
- http://searchengineland.com/googles-not-provided-impacting-more-than-just-seo-sites-120144
- http://cyrusshepard.com/7-fantastic-seo-tips-for-googles-not-provided-keywords/
- http://www.infront.com/blogs/the-infront-blog/2012/12/14/keyword-not-provided-in-google-analytics
- http://www.kaushik.net/avinash/google-secure-search-keyword-data-analysis/
- http://www.optify.net/forms/optify-not-provided-study
What are your thoughts? Can you shed your own light on the very frustrating Keyword (not provided) issue? Let us know in the 'Make a comment' block below.
For more on analytics and tracking, follow us on Twitter.







While losing some insight as to just which keywords are more effective, the other side of this coin is you'll know just how much of your audience is using (often unknown to the end user) Google products when reaching you.
This information can help you leverage other Google Services for future marketing efforts if not particualy helpful for SEO.
Good points, I like step 3 "Deep Dive" the most. Thanks.
Posted by Robert on 2013/02/18