Since Digivate first published the resources on this page we have launched a new SEO reporting solution. We can now show our clients what’s really happening behind Not Provided keywords!
Read more about our leading-edge solutions here.
Google is increasingly using encryption and privacy as an excuse to hide the impact of SEO campaigns behind “(not provided)” keywords.
Nobody can give you an exact figure for the brand vs. not brand split of “not provided”. But we’ve found that the method used here can provide a good top level estimate. For those wanting a more sophisticated approach, we suggest trying the landing page “clustering” method.
Calculating not provided keywords with great accuracy takes a lot more work. There are some links to useful resources at the bottom of this page.
Brand Excl. Not Provided
Filter your organic search traffic in Google Analytics. Include your brand name. Exclude “(not provided)” keywords.
Non-Brand Excl. Not Provided
Filter your organic search traffic. Exclude your brand name. Exclude “(not provided)” keywords.
Filter your organic search traffic. Include “(not provided)” keywords only.
Brand Not Provided
This will reveal the visits or revenue from brand keywords previously hidden within “(not provided)”.
Non-Brand Not Provided
This will tell you the visits or revenue for non-brand keywords previously hidden within “(not provided)”.
Total Brand Incl. Not Provided
This will tell you the TOTAL brand visits or revenue including those previously hidden within “(not provided)”.
Total Non-Brand Incl. Not Provided
This will tell you the TOTAL non-brand visits or revenue including those previously hidden within “(not provided)”.
“Not Provided” Resources
“Not provided” visits to the homepage, sub-brand and checkout landing pages are branded. “Not provided” visits to other deep pages are non-branded.
This article suggests looking at user interaction (average number of page views and time spent on site) for both branded and non-banded traffic and then applying these filters to “not provided” visits. While an interesting workaround, it is more useful for gaining an estimate of “not provided” traffic only, because customers who complete a transaction have to view at least four pages – so non-brand revenue will mostly be filtered out.
This article details how to assess the brand/non-brand split of various landing pages by clustering long tail keywords and using these to gain insight into the distribution of “not provided” keywords.
May be useful, but the time involved could be immense. It certainly wouldn’t be feasible to do this for every page on a large e-commerce site.
Point four in the article discusses using impression data from Google Webmaster Tools – but, as the article concludes, this information is probably not very accurate.
This Analytics filter from Dan Barker Dan Barker’s shows you Not Provided for different landing pages to indicate which pages are attracting the most organic search visits.
This solution also employs the clustering approach mentioned previously – and recommends assessing a quarter of a site’s key landing pages.
The author claims the process takes 5 hours to setup, then 40 minutes per monthly report.
It involves intermediate to advanced Google Analytics and Excel skills (v-lookups, pivot tables etc.) to categorise landing pages.
Similar to the above. The author admits that figures will still not be close to exact.
Uses advanced Google Analytics (custom reports) to establish how “not provided” traffic performs (including conversions) then compares this to brand and non-brand performance.
It doesn’t really give a final breakdown of “not provided” brand vs. non-brand. But it does provide some very interesting insights. (Also very time consuming).
Disclaimer: none of these methods is 100% accurate! Please get in touch if you have any additional suggestions.