ICO website traffic impact of cookie opt in

    Newer Older

    This information provided by the Information Commissioner's Office under a FOI request I made, shows how traffic measured in the web analytics tool (GA) has fallen by 90% since their explicit cookie opt in request.

    Brian Clifton, DHS, GoSquared, and 9 other people added this photo to their favorites.

    1. tim.lboyce 34 months ago | reply

      That's really stunning. Thank you for getting this and sharing it, Vicky.

    2. kewald 34 months ago | reply

      Good thinking Vicky to get this under FOI - not so good news for Analytics I think...

    3. brockvicky 34 months ago | reply

      If common sense does not prevail, this is extremely bad news for analytics and for website users. Even for non profit/government & Quango sites like this, how can you improve user experience, deliver taxpayer value and meet mission policy goals if you can't measure your results?

      Having only 10% of the base to work with really is 'tip of the iceberg analytics'

    4. chinwag.com 34 months ago | reply

      Vicky, thanks for putting in the FOI request, a real portent of doom if this is the effect on comercial sites. It's already getting a bunch of comments on the blog I put up on Chinwag: Cookiepocalypse: Implementing New Law Drops Use by 90%.

    5. richardturrell 34 months ago | reply

      Apologies if I'm dampening the mood but does anyone have any stats prior to May 2011 as it is widely assumed that the ICO traffic may have peaked exponentially during May as people were trying to seek advice on how the Cookie legislation would impact upon their own website(s) so would be great to analyse the % drop of tracked visits using a larger data set?

      It also raises the question if the techniques used by the ICO to request cookie opt in (who honestly looks above the top banner?) was the best strategy to adopt.

    6. robmclaughlin1 34 months ago | reply

      The reality of this shocking piece of legislation. I hope the ICO are using this as a test to help form their approach to enforcing the directive.

      Their 'above the banner' implementation is not by any means their suggested route to conform - in fact they currently do not officially propose anything .

      Real genius on your part Vicky for requesting the data - would love to hear about that experience in fact. They must have responded very quickly given the recency of the data they provided?

    7. brockvicky 34 months ago | reply

      I don't have data before May, but looking at the raw numbers for visits and unique visitors, I'm not seeing evidence of unusual peak type activity, which tends to follow a predictable curve. (I actually expected to see a spike & rapid tail off both immediately before and after the key dates related to us, the interested cookie community - if that is what you want to call us. Actually, it looks like there are just not that many of us). The daily visits and unique visitor figures are pretty steady and the pattern very typical of government & agency clients we have worked with. The distribution of the changes across the period are also very steady, making me think the sample was a valid and robust one.

      I do agree that the "above the banner" request may be a primary candidate for the low opt-in proportions (along with the complete lack of 'what's in it for me' messaging). Its in the area Jakob Nielsen refers to as the blind zone. I think businesses that offer clear carrots & sticks for opt-in will see very different ratios.

    8. brockvicky 34 months ago | reply

      I think the ICO will be learning from this - they are after all answerable to the Ministry of Justice and ultimately the tax payer. They have fiscal responsibility to prove their financial investments are wise, including their web invesments.

      In response to questions about the FOI request - I put this request in just over a week ago, on behalf of the WAA and UK analytics community, and I got a response today with a very helpful daily breakdown of visits and unique visitors (which is what I asked for). I popped it into Excel, ran some queries to check its robustness and these two graphs were the most clear & impactful.

      I was very impressed by the ICO's response and I am sure the person who "owns" this data internally is feeling the pain that we do when we look at the numbers.

    9. richardturrell 34 months ago | reply

      Hi Vicky

      Firstly hats off to you for obtaining this data and please do not take my comments as opposing your work :-)

      I agree that from a statistical viewpoint the cookie legislation has created an additional barrier and I am sure that there are many people hoping Google challenge the legislation or the main browsers offer a combined resolution within the next 12months.

      Looking objectively at the ICO website we also have to consider why would a visitor accept the cookie opt in as far as I can see it wouldn’t enhance their experience and only benefit the ICO’s reporting capabilities!!

    10. Timgu Ltd 34 months ago | reply

      Wow. Amazing graphic. It says so much about the implications of the new rules.

      I wrote about this on my blog www.internationalmarketingguys.com/2011/06/european-commi... and linked to flickr.

      Is it also ok to reuse the image in a blog?

    11. brockvicky 34 months ago | reply

      A few people have asked me for the raw data - given it has been made publicly available under the FOI Act, I have uploaded it here: spreadsheets.google.com/spreadsheet/ccc?key=0AqwDz7xD-hGH...

      Let me know if you do something interesting with it!

    12. brockvicky 34 months ago | reply

      Please feel free to reuse the image, just credit Vicky Brock, @brockvicky

    13. robmclaughlin1 34 months ago | reply

      Thanks for the data Vicky.

      I added a ratio of visits to visitors and it really shows how the opt-in could start to mess with the some of the key performance indicators that are widely used. flic.kr/p/9W4XSg

    14. brockvicky 34 months ago | reply

      Nice evolution Rob!

    15. moresense 34 months ago | reply

      I believe browsers will or should deal with this new law.

      When a user visits a site that is trying to send them cookies it stops and asks the user for permission and give them the potential issues if they do not accept. The "error message" could even come from a meta tag in the page header.
      Obviously IE will probably be a bit slow on doing this but the other browsers will do it quickly - especially Google Chrome as their analytics app will suffer massively if they don't do something!

      Wrote a blog post about it here The impact of the new ICO’s UK/EU Cookie Law

    16. brockvicky 34 months ago | reply

      Others have more information on this than me, but I thought that the original intent behind this was a browser level solution. It seems insane that on top of everything else, we might be trying to push businesses towards the time-wasting practice of opt-in message optimisation.

      I'm not the first to suggest it, but there has to be a better way of doing this - otherwise why bother running an ebusiness in the UK? Most commercial organisations do not operate websites out of the goodness of their hearts, they do so for commercial reasons. If you cannot measure and optimize your business and revenues in the UK, meaning profits suffer, then why not shift your web business to a place where you can?

      Then you have all the government, information and non profit sites who are measuring purely to improve user experience, demonstrate stakeholder value and to comply with the EU e-accessibility and e-governance rulings. I'm not sure how removing their ability to do this helps anyone at all.

    17. tony_sprague 34 months ago | reply

      Great post Vicky.With browser settings the problem still lies with the website owner. How does a website owner know that there has been a specific opt-in to cookie acceptance. The options are for the browser to send a code to the website such as the "Do-not-track" initiative on Firefox. but that opens up the argument - what should I not track, nothing or just bits that I think are necessary. opt in needs to be cookie by cookie with explanations for the use of each. My main concern is pushing unscrupulous websites (are there many of those?) toward tracking by stealth. using browser footprints - over which you have no control. Also I envisage cookie cartels springing up which use an opt in gained by dubious means. cookies will become a currency. its a heap of mess.

    18. brockvicky 34 months ago | reply

      I don't think it is just unscrupulous websites that will get pushed to measuring by stealth. Since day two of the Internet, we've been measuring it. As the old cliche goes "you can't manage what you can't measure." Personally, I think if you effectively destroy one type of mainstream (and I would argue safe, innocuous and anonymous) measurement then the analyst has two broad choices.

      1) Go back to server logfile analytics - this was never great and is way more problematic now logs are vastly greater and fewer people have direct access to them. There is no end user control over logfile entries (it is not measuring users, bur servers) & if analysts are needing to get into ecommerce logs etc, there is much more risk of accidentally exposing personally identifiable information to those who don't need to see it.

      2) Customize page tagging technologies - maybe to flip cookies (or should they be tiddlywinks or frisbees) that grab browser or other parameters into hosted cookie jars. Maybe to write easier to manage virtual logfiles. And as Tony says, now the user has no control over reviewing or deleting these.

      Of course, this excludes the other option of completely ignoring the ruling and waiting to see if the fines outweigh the benefits.

      3rd party cookie cartels/exchanges already exist, they already have value to the behavioural targeters.

      But this has never been where the analyst is working - anonymous, first party cookies served from a site you own is a world away from the third party cookie exchanges used in BT.

      Its just that by taking a steam roller to crack a nut, we've got caught up in the legislation and - bizarrely, are being seen as the problem when the real offenders are barely acknowledged or understood

      I read on Sam's ChinWag post people saying that businesses would never measure you like this in a physical shop - um, yes they do. Every step you take, every pause you make, everywhere your head, hands and eyes point. But the main difference - at least for user experience analytics - is that if in my real world shop I accidently block the aisle so no one can get to the bread, I can see this quickly and the customers will tell me. On the web, they just click away in frustration, maybe thinking we don't sell bread.

      What ever technique we are forced to, or opt to use - we still need to make those measurements for the very reason that we have no other way of seeing what is going on in our online property. Stupid legislation will just fuel creative improvisation, like Tony suggests, it isn't going to solve anything.

    19. Op2mise 23 months ago | reply

      Great information guys, I wish i could give my clients a straight forward answer on what to do with their websites, but alas I feel at the SME end of the scale their businesses would be damaged by using pop ups and warnings, which more than 90% of people wouldnt understand and would click away quickly in fear of a virus or some other spam activity. This legislation wants sacking!

    keyboard shortcuts: previous photo next photo L view in light box F favorite < scroll film strip left > scroll film strip right ? show all shortcuts