As was foretold, we've added advertisements to the forums! If you have questions, or if you encounter any bugs, please visit this thread: https://forums.penny-arcade.com/discussion/240191/forum-advertisement-faq-and-reports-thread/

Help on a Google Analytics topic - bounce and conversion

YogoYogo Registered User regular
edited June 2014 in Help / Advice Forum
Hi all

Some of you might remember me making a post about BI/Big Data a couple of weeks ago. Long story short, I got an internship doing web/social media analytics and I have a question about Google Analytics.

The deal is this:

On a website where I am managing data, we want to lead the users away from the website over to other websites. We do this by having users click a button which displays three icons where bit.ly links have been embedded underneath. Users then click on one of these icons and a new page opens in a new tab. This action happens without changing the page (I assume it happens within an embedded flash object).

According to my understanding of Google Analytics, this would mean that users, who click on these links without having clicked on anything else before, would be considered to have bounced away (since we are leading them away from the site). This means that the stated bounce rate is wrong since % of users did actually do what we wanted them to. Thus the bounce rate would be lower than listed.

Since the users, who did the right thing, are included with all the users who didn't do the right thing, I set up an event tracking the clicks on these links. Thus far I have been able to calculate the true bounce rate by substracting the no. of bounced users with the amount of converted users and recalculating how many % the new number is of all sessions.

Thus far this has worked pretty well, but now I am in doubt if I am understanding my data corrrectly.

Today when I looked at the bounced session segment, the event tracker was 0 for users who clicked on the links. Which it shouldn't be if converted users are included in the overall bounce rate.

However if the act of opening a new page in a new tab doesn't constitute bouncing away from your website, then I have been calculating the true bounce rate completely wrong, and the data Google Analytics is showing me is right.

Question is if this is a case of Google Analytics using a default calculating method which doesn't account for certain types of data (for example, all "bounced" sessions are considered 00:00 in duration even if the user spends his time reading a blog and then clicking away).

If any of you have some clue or insight into what is up and down, I would appreciate it :)

Yogo on

Posts

  • ElfWordElfWord Registered User regular
    The bounce rate is the percentage of single-page sessions where users left without interacting with the page. Anything that triggers a GA event is interacting with the page, though. So by adding an event to the links, what you've actually done is fixed the bounce rate to be calculated as you want it. If even before that you had all outbound link clicks being tracked by GA (a common option with a lot of Wordpress GA plugins), then your bounce rate has been correct all along.

    I've made this mistake myself. After publishing a proud blog post about how my redesign (which included some expand / hide events I was tracking in GA) had drastically lowered my bouncerate! *facepalm*

    Star Wars fan, Battlestar crewman, Fantastic GM. Frequent lurker, occasional adventurer.
    Awesome android RPGs are made by my friends; check them out.
  • YogoYogo Registered User regular
    This would only be the case if the event is flagged to be interactive, right?

    Cause I have seen examples on the net where users accidentally created an event and got a zero bounce rate. The solution was to flag the event as non-interactive.

  • ElfWordElfWord Registered User regular
    As interactive is the default, it might be better stated as "unless the event is flagged as non-interactive", but yes. Have you flagged your events as non-interactive and are still seeing the discrepancy between your calculation and the segment view?

    Star Wars fan, Battlestar crewman, Fantastic GM. Frequent lurker, occasional adventurer.
    Awesome android RPGs are made by my friends; check them out.
  • YogoYogo Registered User regular
    ElfWord wrote: »
    As interactive is the default, it might be better stated as "unless the event is flagged as non-interactive", but yes. Have you flagged your events as non-interactive and are still seeing the discrepancy between your calculation and the segment view?

    I wouldn't be able to say since I am not the one setting up the tracking. I just read the data gathered :p

    From what I can see of the data gathered, it looks like the event is interactive though I can't trace it specifically in the website code (website is outsourced to an external freelancer who does the coding).

    Based on this, I think I can conclude that I calculated the bounce rate incorrectly to begin with. It just is that high because users either 1) don't find our content interesting or 2) they only read our content, but don't click on anything afterwards.

    In any case, I have asked our freelancer to elaborate how he sets up the tracking, so I can understand the foundation of the data.

    Thanks for the help :)

Sign In or Register to comment.