Random Post: The Final Beginning
RSS 2.0
  • Home
  • About
  • MBA Guide
  • Print Ad Blog
  •  

    Google Search History Update

    September 9th, 2008

    In my post You Are Being Tracked: E-mail Style there was some discussion/confusion about when I said:

    Hopefully you know that Google keeps track of everything you have searched for.  Ever.

    Well this it true and false depending on how you read it.  GHamilton noted that Google does not keep everything you searched for.  Rather they keep it for 18 months.  The key word here is “YOU”.  In a recent post on the Google Blog Google announced:

    Today, we’re announcing a new logs retention policy: we’ll anonymize IP addresses on our server logs after 9 months. We’re significantly shortening our previous 18-month retention policy to address regulatory concerns and to take another step to improve privacy for our users.

    You may see where I am going here.  GHamilton is correct in that from and individual IP perspective after 18-months, or rather 9 months now, Google no longer has history on you specifically.  I am correct in that Google really does have “a history of everything you have searched for.  Ever.” with the caveat that they no longer know that you were the one that searched for it after 9 months.  If you are interested CNET does a great job of getting into the nitty-gritty and explains why the ACLU is so critical of Google’s privacy policy.

    Hopefully that helps clear things up a bit for people.  Let me know your thoughts.  Do you care that Google keeps your data for 9-months?  Should it be longer/shorter?  Should they keep anonymous search history forever?



    You Are Being Tracked: E-Mail Style

    September 6th, 2008

    Most people probably already know that they are being tracked.  There are all sorts of programs and ways to do this at all sorts of levels.  For instance your ISP may track you and give (sell) your data to a company like Hitwise – privacy policy can be found here.  I actually saw this in a newscast last week.   They interviewed some guy about what popular search terms are and tried to make it sound creepy.  Amazing! People search for weird stuff on the internet like “how to make bombs” and *gasp* “porn”.  This guy must be some sort of genius!  And he looks at historical data! Brilliant!

    Hopefully you know that Google keeps track of everything you have searched for.  Ever.  Anyway, the part that people probably don’t know as much about is how individual sites track you.  One way a site can track you is by tagging you when you click through on an e-mail they send you – the focus of this post.  Think of tags as dated stamps in your passport book.  Interestingly enough, some of this tagging can be easily found in the address bar of your browser.

    When you see something in the address bar that looks like emid=584783 that is telling the website that your internal – meaning site specific- e-mail address ID is 584783.  This value is unique to a single e-mail address. Each e-mail sent to that e-mail address will have their unique emid attached to all links in the e-mail. This also allows a site to build a history of that e-mail address – not only for activity, but for response rate as well.  Now every time you click through an e-mail for that site they have more history.  Note that larger sites rarely look at individual behavior but instead classify a behavior and then analyze that group.  Still, the information is there.

    In addition to an e-mail ID, there is usually a campaign variable such as cid=Sep08FreeShipping.  This allows the site to report on everything with Sep08FreeShipping stored in the cid variable. All of this information is contained within the link that you click from the e-mail. If you get the e-mail and directly load their site, not through the e-mail, the activity will not be tracked because in a direct load no value would have been assigned to cid.

    These variables do not have to remain in the web address the entire time.  They are stored in the background after the initial click. So when you no longer see emid or cid in the address bar, but originally arrived at the site through the e-mail, you and your activity is still being tracked.

    Look for at least one more installment of how you are tracked. There I will focus more on how a site tracks internal campaigns. Hope this helped give some people a better understanding of how websites track you.


    What A Picture is Worth

    August 25th, 2008

    It is hard to value a picture of your family on vacation, of your friends at a party, or of your website 3 weeks ago.  As touching as the first two options are, this post will focus on the last of the three: a picture of your website.

    Now for many people, this is not a big deal.  If you are running a blog or some other site that has a primary focus of serving content your website may not change much in appearance over the course of a few weeks.  But for the people that work with sites that have a focus of selling things with crisp pictures and captivating copy, things are  changing all the time – and analysts may or may not know about it.

    As a beginner in web analytics, and in fact to marketing in general, it has become clear very early on that knowing why trends are changing in your data is nearly as important as the data itself.  What people want from an analyst is not “what is changing” but rather “why is that changing”.  Answering the first part will get their attention but very few people are content knowing that traffic is up 15% without knowing why.

    Keeping a log of changes either manually or by writing a simple script – currently beyond my abilities, but a co-worker of mine whipped something up – is of great value.  Now we have a copy of what pages looked like on any day we want.  If the creative changes or a sub-zone on a division tab moves we have that change recorded.  With copies saved off, you are the owner.  Better yet, when someone asks for some analysis on your site’s homepage or division page you can not only show them the data and a cool graph, but an image of a pre-change and post-change look at the site.

    It would be great if these things just happened or if the web analysts always knew what changes were happening when, but the reality is that that is not always the case.  If you manage the copies saved off you can always be sure you will capture the change, no matter if you are informed about it ahead of time or not.  If there is a question about content change, you can now look for yourself and deliver your analysis in a more timely and – more importantly – a more informed manner.

    I’m not sure what a picture is worth, but a picture can turn data into information and information is priceless.

    UPDATE: Just found a great example of this in an article written by Bryan Eisenberg.  You can find it here.  It deals with how people interact with Google search and how it has changed overtime.  Short, simple analysis but cool.