Experian Hitwise: Not Immune to Encrypted “Not Provided”

by Roland Dunn from our Killer Keyword Research panel

Google Encryption Reaches the UK:

Since October last year, we’ve all become increasingly aware of the impact of Google’s decision to encrypt search queries for users logged into Google properties (such as Gmail) – the dreaded “Not Provided” has begun to appear in all our Google Analytics data:

According to some (for example econsultancy):

“As a site which attracts US and UK traffic, we’ve seen this affect US visits, with up to 33% of referral data encrypted on certain posts.”

33% … that’s quite some loss of referral data, but that does perhaps seem particularly high – see “Media agencies report over 10% of SEO traffic hit by SSL encryption“.

But How Is This Affecting Experian Hitwise UK?

According to Experian Hitwise UK (quote obtained on March 15^th 2012):

“A lot of people have been asking about Google’s secure search and how it will affect online analytics. The truth is that although the implementation of secure search will have an impact on all analytics tools, the effect is likely to be less significant than a lot of people have forecast. Certainly at Experian Hitwise we are currently seeing a much lower proportion of traffic coming to sites via secure search than reported by eConsultancy and other publications. The reality is that only a small proportion of Internet users have a Google account and the vast majority of searches on Google will still be captured in analytics tools. For example, less than 5% of UK visits to email providers each month go to Gmail, whereas Hotmail still has over 50% market share of email visits. This may change over time as people migrate to Google accounts, but for the near future secure search is not nearly as big an issue as many are predicting. For our own analytics, we’ve built an algorithm which factors these secure searches into the data so it doesn’t change the insight delivered via our applications.”

A few of bits of information jump out:

• This encryption is affecting Experian Hitwise – good to have that confirmed in writing (it’s hard to see why it wouldn’t affect Experian Hitwise).
• Experian Hitwise’s data is being impacted less than eConsultancy, but it’s not clear by how much.
• The statistic regarding Gmail is interesting in itself – but doesn’t tell us about number of users logged into other Google properties, and doesn’t tell us how much less their data is being affected.

When I spoke to Experian Hitwise, I wanted to understand the last two lines, i.e. the statement regarding the algorithm. My understanding is this: that Experian Hitwise UK have found a way to work around the implications of the encryption in some of their tools. For example, the tool that allows you to find related search queries, e.g. find me search queries related to “blue socks”.

However, when I use Experian Hitwise UK I’m really interested in downloading the sets of referring search queries, i.e. to get hold of the actual data myself. One particularly powerful feature of Experian Hitwise is that you can put together a collection of websites that you wish to examine (known as a custom category) – such as your own website and a selection of your competitors – and download the set of search queries that successfully refer search traffic to those websites collectively. Using this technique you can pull together very large sets of such search data, for example, up to half a million search queries over a year (which you can then mine to work out what your business should be relevant to as opposed to what it currently is relevant to).

So I’m actually much more interested in this data than some of their tools – and this data export will be affected, there will surely be less such referring search queries.

The End of Referred Search Queries?

Just this evening, Wednesday 21^st March, it’s being reported that:

(see http://www.blogstorm.co.uk/firefox-to-block-all-keyword-data-using-https-by-default/)

How long before Chrome, IE, Safari do the same? Where do we then find referral search data?

Google Webmaster Tools data is something, but it’s nothing compared to the search data that we can access in our analytics at the moment.

The only company I know that does “relevance analysis” without a crucial dependence upon referred search data is http://www.bloomreach.com/. Anyone know of any others?

Roland is a partner at Refined Practice, and a natural search marketing consultant at Cloud Shapes.