<strong><span style="text-decoration: underline;">Background</span></strong> In October 2011, Google announced it would be encrypting searches for users logged into Google mail. This resulted in referral data from these searches being stripped out, leaving you with the very unhelpful 'not provided' keyword.
In October 2011, Google announced it would be encrypting searches for users logged into Google mail. This resulted in referral data from these searches being stripped out, leaving you with the very unhelpful 'not provided' keyword.
The rollout of Google + meant that more people were likely to be logged into their Gmail accounts increasing the amount of encrypted searches. Luckily for search marketers, Google + wasn’t as popular outside of certain industries as it was predicted to be, thus postponing the eventual increase in not provided keywords.
Just last month, Firefox started using encrypted searches as standard for its browser search facility and webmasters saw further increases in the percentage of not provided results.
Why is this an issue?
One of the most useful methods of measuring the success of a search campaign is to measure the increase in non-brand visits to a site. Over the last few weeks of reporting you may have noticed a sharp, unseasonal increase in the amount of on-brand visits, great right!?
If you’re one of the more unscrupulous search marketers out there, you may have considered inferring that this increase as your own hard work paying off, but in reality a portion of the 'not provided' visits are brand and not directly under your control. But, that's just one out of a whole host of issues as the referral data is vital to investigating the performance of keywords as a whole, optimising landing pages, keyword research and more.
What can we do about it?
It’s not all doom and gloom! It’s still possible to get some insight into the not provided visits; all you need is a bit of basic analytics know-how and some handy tutorials!
- Brand/Non Brand Split: Find the brand/non brand percentage split of organic visits and use this percentage on the not provided amount. This is a very basic way of giving you more insight into the data, it’s not perfect - but it’s useful for a quick bit of analysis.
- Total organic visits – not provided visits
- Calculate the brand/non brand split
- Use the percentage split on the not provided keyword
- This will give you an approximate amount of which visits are brand/non brand
- Which pages are the not provided keywords landing on? Although we aren’t able to see the keyword used to enter the site we can still see which page they landed on, giving us an idea of the topic of the keyword and also whether its brand/non brand.
- This tutorial shows you how to use filters to discover which page the not provided term lands on. http://econsultancy.com/uk/blog/8342-how-to-steal-some-not-provided-data-back-from-google
- Webmaster Tools Impression Data: Webmaster tools give users access to impression data, it’s then possible to extrapolate this to give insight on keywords used to get to the site.
- This tutorial gives an in depth explanation of this method. http://searchenginewatch.com/article/2158540/Recovering-Not-Provided-Keyword-Data
The percentage of not provided visits is set to grow as the popularity of Google accounts increases and further browsers adopt encrypted search as standard. This issue is here to stay, and in a data driven industry such as ours - it's vital that search marketers are equipped to handle this problem.
What are your experiences with this issue? Do you know of any other methods to get the most from this data?