Digital Services manages the University’s Google Analytics implementation. We have a couple of updates for anyone in the campus community who reviews or reports analytics for their CMS website(s).
During the fall semester, we made some adjustments to our tracking code to better attribute digital advertising conversions between Google AdWords and Slate CRM.
During this upgrade, we unintentionally triggered two pageviews within Google Analytics for each real pageview on the website. As a result, pageview metrics are doubled from mid-October through early January. The issue is temporarily resolved and we will be testing a permanent solution in May, once we are no longer in the hottest portion of the recruitment cycle.
Although Digital Services has moved on from reporting pageviews as a key performance indicator (KPI) for a few years now, we know that some may still use this metric, so we want to be transparent about the issue. We use unique pageviews as one of our KPIs and this metric was unaffected. Unique pageviews are also less susceptible to inflation by internal staff, Internet bots, and spammers.
For the past few years, because we use scroll depth tracking, Google Analytics has considered any scrolling on our website to be a notable indicator that nullifies any bounce rate measurements. As a result, our bounce rates have been extremely low.
During the fall semester, we discovered that we could now exclude scrolling from being considered a notable indicator. At the same time, our duplicate pageview issue converted all bounces on our websites to exits as every visitor had at least two webpage views, so bounce rate numbers for the fall are inaccurate.
Moving forward, bounce rate metrics will be more in line with patterns from four years ago, when it took clicks to nullify bounces.
Although Digital Services has moved on from reporting bounce rates as a KPI for a few years now, we know that some may still use this metric, so we want to be transparent about the issue from this fall and the measurement change moving forward.
Avg Time on Page
Avg time on page (ATOP) is measured from the time a webpage visitor enters a webpage to the time they navigate to another page. Because navigating to another page is a requirement for Google Analytics to take its ATOP measurement, bounces do not affect ATOP.
Since our duplicate pageview issue converted all bounces on our websites to exits, our ATOPs decreased during the fall semester as a visit of zero seconds was included in every calculation. Moving forward, since we will have more bounces now that scroll depth is not a notable indicator, ATOP may decrease as well.
Session duration, which is a measurement of how long a user’s session lasts on a website is unaffected. A session is a group of user interactions with your website that take place within a given time frame. For example, a single session can contain multiple pageviews, events, etc.
Although Digital Services has replaced ATOP with session duration as a KPI for a few years now, we know that some may still use this metric, so we want to be transparent about the issue from this fall.
In addition to using unique pageviews, sessions, and session duration, we also look at new users, scroll depth tracking, and unique conversion button clicks as our KPIs to support decisions and indicate success. All these KPIs are unaffected by any issues.
We consider metrics such as pageviews, users, events, and bounce rate to be vanity metrics that don’t support decision making or indicate success.
If you have any questions or concerns about your CMS site’s analytics, you can reach us at firstname.lastname@example.org. We’re happy to help.
Director of Digital Services