• Damion Brown

    I was under the impression that Google doesn’t want us using GA to personally identify specific users of a website. By tracking email address or account logins (which is what the blurred bits on the top screenshot look like) is this against Google’s T&Cs?

  • Chris Le

    Hi Damion! No, we are NOT tracking personally identifiable information in Google Analytics. We are very aware that doing that is definitely against Google’s T&Cs and we wouldn’t do that.

    The blurred out parts are variable names that identify what textbox or pulldown people interacted with. They’re blurred out because the variable name included the name of the client or something that would have made it obvious who the client is. Any of the login parts were the names of just the services like Twitter or login with Facebook. It only identified that they used Twitter or Facebook or an email address to login, but NOT who they were.

    (edited a typo)

  • calebdonegan

    Fantastic article! Our event tracking is setup without any method to the madness. Can always understand what each pertains to, but it makes it difficult to compare across accounts. This is a great way to set it up. Thanks!

  • Scott Thomas

    I’ve set up some event tracking for small sites and I’ve been wondering how to scale it up to large ones, and thanks to your post, now I understand how to tackle it. Thanks!

  • Matt

    How about reversing the category/action names? For example, the files downloaded may be in different formats (PDF, Word, PPT, etc.), but in a category titled “Product Sheets” or “Case Studies”. I see that as more important than the file type of the download. Then, you can just insert the file name in the label (e.g. CaseStudy.PDF).

  • Chris Le

    You’re very welcome! This was exactly why we had to come up with a naming convention. Turns out, it scaled both ways really well. Good luck!

  • Chris Le

    You’re very welcome!

  • Chris Le

    You could definitely do it that way if that works better for you. First, thank you for asking this question.

    It raises a point that I wished I could have talked about more explicitly. (I didn’t to keep the post short.)

    Where did this pattern came from and why do it this way (other than for reporting)?

    The reason I would prefer “PDFs > Product Sheets > Case Study.PDF” is because CaseStudy.pdf is a specific product sheet, and product sheets are a subset of all PDFs.

    I’ll model data the same way I would structure the data in a databases or making object classes in programming. For example: Users > Person > Color preferences. You can also find the same pattern in URLs, breadcrumbs, folders on your hard drive, etc.

    Why would I decide to follow this pattern?

    I’m of the opinion that business requirements will change and that’s OK! I don’t know if filetype is going to be important two years from now. So, the engineer in me tends to go towards future scalability. The drill down pattern has been tried-and-true in computer science for longer than I’ve been alive, so I’m going to take, almost at face value, that it has some merit to it. So far, it hasn’t let me down.

    Ultimately, this post is just a way of thinking. You must change it!

    Don’t ever put a square peg in a round hole. If my method doesn’t work for you. Change it! Make it work for you! And most importantly, post about it so everyone here can learn from your experience! :)

  • Matt

    Thanks Chris, good info.

  • Matt

    Thanks – good info!

  • Zdeněk Hejl

    I name events slightly different. Categories are types of pages on which events can happen (homepage, product category pages, product details, checkout process, contact page, thank you page, …). Actions are activities which can be made on different types of pages (click on email, click on external link, click on PDF, open contact form, …). And I use labels to give some extra infomation to actions (on which link/email address user clicked, which PDF user downloaded, …)

    This type of hierarchy is better for me. When I optimize the website, I pick one page type (event category) and then I can see which actions users did on these pages. For example, on contact page I can see on which email addresses users clicked – which are important for them and which are not.

  • Josh Gates

    GREAT article! Assuming your mobile content is on separate page/subdomain, I’m curious how you would scale this naming for different devices, if at all?

    For example, using your examples above, and coding for videos watched on mobile devices; would you have different ‘Action’ of “Viewed on mobile” so as to track the interaction on mobile? thanks!

  • charliesaidthat

    Really great. Have come across far too many sites that have ever changing naming conventions for event tracking.

    Will be stealing this to scale some projects I’m working on now. Cheers guys!

  • Daniel

    What logic is used to name ‘Value’?

  • Karen Wiskin

    Genius, thank you!

  • David

    can you give an example of how you would use for an email contact e.g. customer support?

    see you have as a category “Interactions” but not sure what you have in
    mind for this. I was thinking to use category “Contact” but that
    doesn’t go with your plural noun strategy

  • Tesh Vora

    Thanks SEER – Very useful! I am going to suggest this convention.

  • Aanchal Gera

    Can I see events broken down by different dimensions like country, deviceCategory etc?