A Bad Fortune Cookie

The Information Commissioners Office (ICO) has offered advice regarding the interpretation of the European Commission’s ePrivacy Directive in relation to cookie tracking and made suggestions for how websites might attempt to gain the explicit permission of users to be tracked. Whilst we wait for the ICO to cement their thoughts, and essentially the rules of the road for cookie tracking, they have granted the industry what is affectively a 12 month amnesty. Between now and the May 2012 businesses need not fear incurring the wrath of the ICO, regarding cookie tracking. Of course, whilst it offers short-term relief, this is neither a long or even medium term resolution and we must ask what the future of web analysis will look like.

Browser developments may bring a technical solution, which will mean site owners need not make specific changes to website design; however, the ICO recommends that alternatives are investigated and prepared. An obvious solution that they suggest is an opt-in to cookie tracking as a pop-up or a header that is displayed upon landing on a site. The ICO actually implemented such an opt-in for www.ICO.gov.uk on 26^th May, 2011, in-line with the deadline for countries to accept the ePrivacy Directive.

Vicky Brock of Highland Business Research, whom I and colleague Paul Cook had the pleasure of meeting through the Web Analytics Association (WAA), had the truly fabulous idea of making a data request, ironically by the very Freedom of Information Act that the ICO administer, to the ICO for their web traffic data for the pre & post 25^th May time period. The ICO obliged and as she shows graphically here, the cliff is startling, if expected:

As Vicky says:

“If common sense does not prevail, this is extremely bad news for analytics and for website users”.

The stakes are high and, in my opinion, the European Commission is showing a real disregard for how many businesses gain intelligence on their marketing performance. As you can see, the opt-in is stripping out circa 90% of users, leaving only 10% who give their permission to be tracked.

Looking at the graph, it became clear to me I was looking at a segment. It is the segment of visitors who ‘opted-in to be tracked’. As an analyst this segment is your sample from which you will need to determine insights regarding user behaviour and outcomes. The ‘opted-in to be tracked’ segment, whilst I am sure they are very nice people, is not a robust sample of users from which to base analysis.

In the true spirit of openness and sharing Vicky made the data from the ICO available and I took a look. By simply creating a ratio between the two metrics (unique visitors & visits) provided it was possible to see the time honoured aggregated metric of ‘average visits per visitor’. Whilst a largely deplorable metric in itself, ‘average visits per visitor’ betrays the extent of the damage that a cookie opt-in could wreak.

Tracked Visits, Visitors & ‘‘Average Visits per Visitor’

Much of the online world lives and dies by conversion rates of one sort or another and conversion rates are ratios of numbers from a given sample. Whilst the cliff in visits and visitors is shocking, the impact on conversion rates and swathes of key performance indicators (KPIs) is much scarier.

As discussed in my previous post, businesses have a tough job navigating the EU ePrivacy Directive and its local implementations however; they also need to be aware of how their metrics could turn to mush very soon. The ICO, by implementing the cookie opt-in feature, to their own site, have given us a stark message from a potential future, and I don’t like it.