Noise or Music? - The Insights Blog

What Google Analytics Can’t Tell You – what rubbish

October 17, 2009 / Categories: Privacy and Accuracy / Comments: 19

Share Button

I wanted to put this out there to illustrate the type of crap competitors will go to to discredit Google Analytics. The link takes you to an article by clicktale which is a rehash of a previous discredited post by Brandt Dainow last year. Take a minute to read it and the two so called flaws of Google Analytics…

Even though Clicktale sugar coat their “article” in the first paragraph, its a poor example of misinformation to promote themselves. The specific points raised were dealt with by myself and others last year – see for example, Shelby Thayer’s excellent blog Trending Upward. I summarize my original response later in this post.

However my beef with Clicktale is not that they use misinformation to promote themselves (lots of competitors do that!), but that they themselves use Google Analytics to track individually identified visitors to their own website i.e. ignoring my right to privacy and going completely against the GA Terms of Service.

How Clicktale break the Google Analytics Terms of Service

By spamming me with their email, Clicktale encouraged me to read more of their “story” by including a link with hidden tracking parameters. Nothing wrong with using tracking parameters to detect a visitor has clicked on an email link, however this must not be personally identifiable. The following is an excerpt of Clicktale’s link:

utm_campaign=What_GA_Part1&ct=enable,t(What_GA_P1_1_10_17_09),t(First Name %3D Brian),
t(Last Name %3D Clifton),t(Email Address %3D brian%40omegadm.co.uk),
t(Company %3D http%3A%2F%2Fwww.advanced-web-metrics.com%2Fblog%2F)- Show quoted text -

As you can see my name and email address are clearly contained in the link. This means that in Clicktale’s Google Analytics reports my details will show up. Unless I clear my cookies, this information will also be tracked in repeat visits.

Is that a valid use of tracking? No – who on earth wishes to be tracked at an individually personally identified level? Its fair enough to propose this if I have done business with you before or subscribed. However, even if you explicitly ask for my permission, Google’s stance on privacy is such that you cannot do this with Google tools/products.

My response to the claims by Clicktale/Brandt Dainow:

The article is factually incorrect. I wish he had reached out to those with a stronger connection to Google Analytics before taking such a strong (inaccurate) stance. Unfortunately, these days writing those type of headlines is a sure fire way of generating traffic, which was probably the intention…

To clarify what he is claiming, he is stating that for metrics such as time on page and time on site, that GA includes single page visits. The unwritten standard in the field of web analytics is to exclude single page (bounced) visits and the last page of a visit for such calculations.

The reason is that time spent on the last page is so inaccurate to measure. For example, say I visit two pages on your web site each lasting one minute, then I get distracted on the third page and leave my browser open. GA will close my session from a tracking point of view after 30 mins of inactivity. That is also an unwritten industry standard and can be adjusted if required.

Calculating the time on site for my visit will be horribly skewed if page 3 is included:
60+60+1800 = 1920 seconds

The reality is better calculated by omitting the last very inaccurate measurement:
60+60 = 120 seconds

So most web analytics tools omit the last page visited for these types of calculations. And if the visit is only one page then then the entire session is omitted. A single page visit is still tracked as a visitor, just excluded from time metrics.

Clearly that still is not 100% accurate, but it is a much better situation than including it. Further information on this can be found at:

If you agree with my sentiment, please retweet this article and spread the word so that we can have an informed debate about the strengths and weaknesses of Google Analytics and not this pap…

Share Button

Comments

  1. Hello Michael

    Thanks for your comments. I wanted to update you with my input:

    Firstly, please note my criticism was of Clicktale’s approach to marketing themselves – not their product. I simply do not comment on products directly.

    Responding to your specific points below:

    A-C) By default, GA will be tracking the Clicktale PII as it is part of the tracking URL they used. However, that was not really my point, as clarified in my response to Clicktale. My concern is that using PII to track a visitor is a no-no. Its bad for the industry and not necessary when it comes to website, web page or marketing optimisation. You may be aware of the current debate about the use of cookies in the US and how the legislators are trying to ban them – particularly from government site. These sort of incidents add fuel to the fire. I am obviously a big fan of web tracking per se(!), but if vendors start to abuse online privacy that damages this industry we enjoy so much. Hence my flaming of any company that abuses this.

    To Clicktale’s credit, they responded by holding up their hands and apologising for the mistake of using PII – see their response on my post. So I consider this particular matter closed. However if someone wishes to start a new thread about online privacy in general, I am more than happy to take part. As you will know I write about this topic often – see: http://www.advanced-web-metrics.com/blog/category/privacy-accuracy/

    In the words of Ian Thomas while he was looking after Microsoft Analytics, “online privacy is the next big thing the industry needs to address” (SES, London, Feb 2008). He was clearly ahead of his time!

    D-E) Comments to my blog became personal and vitriolic and even threats. You wouldn’t have seen some of these as they came direct to my inbox (and still are doing). I therefore closed down the comments as the debate deteriorated into a farce. No new content was being added and there was no informed debate – just a repeat of existing points with SHOUTING and silly slurs on mine and Avinash’s credibility.

    Apologies if your comments got caught up in this – I was battling against one or two people determined to “spam” my site with repeated identical posts. I will manually add this one to the existing thread as you clearly have valid points to make.

    F) Agreed. I welcome anyone’s attempt at improving metrics and understanding in our industry. If only Clicktale had approached this in a more sensible and professional way… it does however make a great case study in online reputation management!

    Best regards, Brian

    — In webanalytics@yahoogroups.com, “mfeiner17″ wrote:
    >
    >
    >
    > Having read Brian, ClickTale and Brandt (http://blogs.imediaconnection.com/2009/10/22/Web-Analytics/Brian-Clifton-s-crusade_958.aspx) blog posts repeatedly I reached somewhat different conclusions:
    >
    > A) ClickTale do not track PII data into Google Analytics thus not breaking the Google ToS. The usage of PII appears to be internal only and legitimate (as long as it remains internal).
    >
    > B) The ClickTale post was sensationalist (too sensationalist IMO) but in no way was it factually misinformed or misleading. It described some commonly accepted limitations of the GA (and most other WA tools) measurement methodology
    >
    > C) ClickTale never claimed that GA includes single page visits for the purpose of calculating Ave. Time on Page. On the contrary, they make several written and visual references stating the opposite
    >
    > D) According to Brandt ClickTale did not plagiarise his content (see his article. It also appeared on the a now deleted comment on Brian’s blog)
    >
    > E) Brian has the right to remove comments from his blog if he thinks they are vitriolic. I thought Brandt’s second comment was on topic, fair and by no means a personal attack. I’m sorry to see it removed
    >
    > F) “I thought the ClickTale post was well written and highlights a reason to use the two services in tandem” – a direct quote from one of the most respected figures in the WA industry. I think that is the point ClickTale are trying to make (even if I don’t entirely agree with how they did it – I would have written that blog post differently!)
    >
    > I’m not affiliated with any of the parties mentioned here.
    > I hope we can continue this debate based on facts rather then rhetoric from both sides.
    >
    > Thank you.
    >
    > Michael Feiner
    > AEP Convert
    > http://www.aepconvert.com
    >

  2. Comments closed.

    Thanks for those that responded. I wrote this post for two clear reasons:

    1. To highlight the misuse of PII that damages the web measurement industry as a whole.
    2. To highlight the misinformation being spread about GA

    To Clicktale’s credit, they responded to the PII issue by holding their hands up and apologising for their mistake. For the second point, the discussion descended into a farce and I started to receive personal, vitriolic comments, both here and direct via email – even threats!

    I now realise some people have their own irrational agenda when it comes to Google and its a waste of time for me to reason with them. Lesson learned…

  3. Brandt: you published your “proof” email at iMediaconnections and Trending Upward and it was commented on by numerous people. It is certainly not any confirmation that you are correct. Why? Because, quite simply the claims you make are not correct. The original post from Avinash on this subject remain correct. I suggest you append your comments to that post for peer review if you still disagree at: http://www.kaushik.net/avinash/2008/01/standard-metrics-revisited-time-on-page-and-time-on-site.html

  4. Jen/jaCob/James Dutton: Just to clarify, my article and responses are not in any way commenting on the Clicktale product, features, methodology or pricing. If you read my blog you will notice that I never comment on other vendor products.

    My comments concern Clicktale’s approach to PII and misinformation – which I am responding to separately in this thread…

  5. jaCob says:

    What?!? ClickTale is free!!

    Don’t know where you get you information from but I’ve been using ClickTale for free for the past 7 months.

    Seriously, just check their web page before you go spouting stuff that just isn’t true!

  6. Jen says:

    It’s easy to forget that GA is an Enterprise solution that is FREE and has tremendously powerful features.

    Clicktale on the other hand is not free and also not cheap. It’s a nice product and has added value.

    However, in their enthusiasm to create a sensational blog post and sell their product I think they overshot the mark.

  7. ginko says:

    Hi,
    I applaud your commitment to giving out great information. Easy to read, engaging and spot on! Well done. Keep up the good work going.

  8. Hi – I’m the original Brandt Dainow. While I note Brian is continuing to tell the world my article about Google’s calculations is factually incorrect, he is aware (because I showed it to him) that I have an email from Google confirming that my article is correct, and explaining why they do things that way. Anyone who wants to can contact me through my site (www.thinkmetrics.com) and I’ll forward the Google email to them. It’s got a name and a phone number if you want to contact them and confirm.

  9. ClickTale says:

    Brian,

    We wanted the opportunity to set the record straight on a few details, and address some of the issues you mentioned here.

    First of all, a genuine apology. The email you received from us with the PII tracking code in the URL was an accident. We were testing a possible tagging feature inside the ClickTale system which had not yet gone public, and an email was sent to a handful of our contacts accidentally. As many of your readers pointed out, It uses our own “ct=t(**)” tag that is NOT tracked by Google (thus not breaking their terms of service), nor is it used for any of our own reports. We take privacy concerns very seriously, and do not collect or track PII as a strict rule. We are aware that many other companies (such as Yahoo! or SlideShare mentioned here) do use PII tracking in their email campaigns, but if you read our own Terms of Service, we make it very clear that we do not. We sincerely apologize again for what was an honest mistake.

    Now for the claims about our post. We stated that we use Google Analytics in the blog post, and are not trying to “Google-bash”. We’re just pointing out known limitations previously mentioned by Brandt and many others and offering a viable, free and accurate complementary solution.

    You say that are claims are “factually incorrect” but refuse to give a single example. Furthermore, if we are wrong and Google Analytics does not measure time as we say, then how can you say it is “the unwritten standard in the field of web analytics”? Either Google Analytics do or don’t do what we say. Please make up your mind.

    Finally, your claim that “time spent on the last page is so inaccurate to measure” is irrelevant. We clearly state in the blog post that we track “Engagement Time” as well as “Time on Page” meaning we have a separate counter that only records time when the visitor is actually engaging with the content!! In the case you mentioned, where someone gets distracted and leaves the browser open, we stop recording, resulting in an “Engagement Time” and “Time on Page” that is far more accurate than Google Analytics.

    We stand 100% behind the claims in our blog post, and ask you again to point out exactly where we were “factually incorrect”.

    Best Regards,
    The ClickTale Team

    • Clicktale: Thanks for your response. Coming forward and acknowledging your mistake in using PII as part of your marketing campaign is a big step in the right direction. As noted in my previous comment, my article is a criticism of your PR/marketing technique and has nothing to do with your product. My response would be exactly the same if you were acme.inc.

      Whether you were breaking the Google Analytics ToS is not really my point. Breaking my right to online privacy is – whether you use GA, Yahoo Analytics, WebTrends, Clicktale or any other web measurement tool. Putting aside your use of personally identifiable information, the headlines you use in your article are simply incorrect and misleading. For example:

      “Google Analytics cannot tell you anything about bounced visitors! “
      Really? – I have entire sections of my book dedicated to describing how to use bounce metrics from GA to optimise website content and marketing. There are also numerous other articles and blog concerning bounce rate optimisation with GA.

      “Google Analytics gives you no information about how long your visitors actually interact with your online content”
      Are you really stating that time on bounced pages is the key metric to define website engagement and GA has nothing else to offer? That’s a rhetorical question, no response required…

      If you stand by those headlines, then unfortunately we really can’t have an informed debate here.

      To summarize my point
      If you have a new innovative way of tracking visitor engagement then great, lets about hear it. Shout from the rooftops, present at eMetrics etc. But please, let it be judged on its own merits and leave the sensationalist and misleading headlines in the bin – along with invading visitor’s privacy. Jeez – I should go into management consulting :)

  10. Brian,

    I’m not here to advocate for ClickTale.

    Interpreting the tone of the blog post is ultimately a subjective thing.
    I do not see this blog post as a direct attack on GA and more importantly I don’t view it as misinformation to promote themselves.

    I think James Dutton summed it up correctly in his response above.

    Thanks again for raising this issue. It certainly a worthwhile discussion.

    Michael

  11. James Dutton says:

    Brian,

    The team over at Clicktale are quite correct in how they describe the scenario within which the proposed analysis – I think you perhaps misrepresented the message they are trying to get across: that a considerable amount of data is missed by script based solutions like GA. The typical scenario Clicktale are alluding to is that of the visitor-reading-single-blog-article, or visitor-landing-on-media-landing-page and leaving. These single page view visits are a source of frustration when using a single tool for analysis. A UX based tool such as Clicktale is a useful supplement to a GA or other tool insofar as it provides an additional dimension to analyses. They were quite clear on their message about supplementary tool in their article.

    Clicktale are not suggesting all tools should ineffectually track time on site by extending last page to full session length as you suggest, they are merely highlighting this failure to capture what they refer to as engagement data. In the same regard, how many folks are really using the time on page metric as a measure of success? Probably not many given the limitations of how tools are able to capture data – data which is an indication of a story, but not conclusive by any means when we factor into simple statistics: averages are not a good measure.

    The point you raise regarding PII data is somewhat misrepresented – while CT are capturing personal data this is through a clicktale (ct=) tag, sending data to the CT backend and having nothing to do with their use of GA for campaign tagging. Though as Adrian points out this might appear in content reports. I am not a big fan of vendors using personal data in this way, but CT are not alone in doing this: one needs only look to the Yahoo Web Analytics documentation (pp 32 of Dennis Mortensen’s book recommends using YWATracker.setMemberId(“user@email.com”) to capture a member’s email address [note to Dennis: I assume this was a mistake?]) to see recommendations for capturing email addresses in custom variables (again – not cool). It’s very commonplace, especially in email – one only needs to look to Mediapost emails to see email addresses being passed in querystrings.

    In conclusion, CT have clearly opened up an interesting dialogue here – I expect we will continue to see very polarised opinion and glad to see this debate – but at the end of the day the CT tool is a very useful interface between report driven analysis and UX driven insights.

    Cheers, J

  12. Adrian says:

    I agree with you on calling this guy out on the validity of his claims, I’m not so sure on the PII issue. It looks like you’re PII is assigned to the ‘ct’ parameter in the query string, which of course is not a GA parameter. So it’s definitely not stored in any campaign tags. It could possibly show up in the strings in the content reports, but in this scenario I would hope that they’re excluding the ct parameter. If that is even the case, than your PII is completely filtered out of the GA data. Am i missing something here?

  13. Michael: I think it is pretty obvious from the tone of the Clicktale post, and that of the original Brandt Dainow article which they plagiarised, what Clicktale are attempting to achieve – using misinformation to promote their own product.

    I have written extensively, from a vendor agnostic view, about accuracy issues of the web analytics industry (see the whitepaper I reference). Not using the last page for time on site calculations is perfectly logical and very common. As I described, including the last page is in fact less reliable. It is not a limitation, but a deliberate attempt by vendors to provide representative data to their customers.

    As for the privacy issue, I include it to illustrate the lack of credibility Clicktale have – or any company that has a total disregard/disrespect of end user privacy. The fact they break the GA ToS is simply ironic.

  14. Brian,

    I have to disagree with you.

    Can you point me out to the section/s of the ClickTale post where he (or she) state that “for metrics such as time on page and time on site, that GA includes single page visits.”

    The blogger did get it wrong with this sentence “if the user minimizes the browser, changes tabs or leaves their computer on overnight, the “Time on Page” COULD BE HOURS, when in reality the user only looked at the page for a couple of seconds.”

    But I felt they clearly indicated in the paragraph immediately prior to that sentence and in the GA experiment screen grab that GA cannot measure any time on single page visits.

    It’s obvious to me why ClickTale targeted GA. I suspect most ClickTale customers use the service in conjunction with GA.
    But to be fair they also indicate that GA measures time “Like most traditional web analytics services”. So it is a wide web analytics issue.

    They clearly indicate at the end of the post that (in their opinion) ClickTale is the “the perfect COMPLEMENT to your traditional web analytics suite.” (so not a competitor and not only GA).
    Therefore, I don’t think it is ironic that ClickTale use GA on their own site. Just because they use the tool doesn’t mean they have to accept the accuracy of all features. I love GA and use it extensively but I wouldn’t use the city-level IP-based geotargeting data for Europe.

    I do agree (yeah!) that tracking of PII is problematic and that in this case ClickTale is breaching the GA terms of service.
    I hope Google chooses to take action and ensure ClickTale refrain from using such practices (that could harm the broader community).

    However, in my opinion this issue warrants a separate blog post and shouldn’t have been mixed together with the matter of data accuracy.

    Thanks,
    Michael
    Sorry, forgot to add great post! Can’t wait to see your next post!

  15. Wow. Tracking keystrokes? Geez, why don’t we just use Morae without telling anyone? (I know … CT doesn’t collect passwords, but still … ).

    That New York Times article came to my mind immediately as well. No wonder the average user has the wrong idea about web analytics.

    It’s these types of “analytics” tools that make it so hard to be a web analytics evangelist in an industry like higher education. If I ever showed our administrators a tool like this, they’d run screaming the other way.

    The GA TOS is bookmarked and readily available because people “always” ask about it.

    Great post, Brian.

  16. Stephane: yes the irony of Clicktale using Google Analytics is not lost on me…

    You are of course correct, this type of abuse and the stealth tactics used by many sites (Flash Shared Objects) are exactly the things that can destroy the web analytics industry we know. The irony is that its the marketers and website owners doing this – the very people that benefit from web analytics when used responsibly.

    I am not sure if it is greed or stupidity – probably both!

    Would love to see (and support) the WAA on a stronger stance on this. I guess they fear being pulled into a policing role, which nobody involved in the WAA wants – including myself.

    My hope is that flaming brands that do this will be enough for them to realise the damage they are doing to their own reputation and address their ways. Then again, may be they are too stupid to know or care…

  17. S.Hamel says:

    Hi Brian,
    When I read the Clicktale post I also had a strange impression of “deja vu”… Now I know why! I was also skeptical of their claims and my reaction was to think “yet another misinformed post”…

    But the point you raise about the way they are tracking PII (Personally Identifiable Information) is not only breaking Google’s agreement, but exactly the type of abuse that makes privacy advocate people release reports such as “Two-Thirds of Americans Object to Online Tracking” (see http://www.nytimes.com/2009/09/30/business/media/30adco.html)

    I carefully read the report, and as a Web Analytics Association Board member, I wish the association would take a clear stance regarding the proper, ethical use of analytics: using 1st party cookies, never tracking PII, never sell data, full disclosure of exact use of data (like the PIPEDA law in Canada), user control, etc.

    As a closing note: isn’t it funny that Clicktale is bashing Google Analytics, yet, they are using it for campaign tracking?!

    Stéphane Hamel
    (ethical) Web Analytics Advocate!
    immeria.net

© Brian Clifton 2015
Best practice privacy statement