Columnist Janet Driscoll Miller explains why you might not be accurately accounting for organic search referrals since the upgrade to Universal Analytics — and details how to fix this issue.
When doing search engine optimization, either for clients or as an in-house marketer, we must accurately measure the impact SEO is making on our websites, leads and revenue.
Measuring our success in SEO helps us to justify the importance of its role in marketing, ensure budget allocation and solidify our roles in the organization. Tracking these statistics can literally mean the difference between keeping your job or not.
One trusty tool in our bag of resources has often been Google Analytics. Last year, BuiltWith reported that 69.5 percent of Quantcast’s Top 10,000 sites (based on traffic) were using Google Analytics. If you’re using Google Analytics, though, are you truly tracking your organic traffic accurately?
Launched in 2012, Google Universal Analytics came out of beta in April 2014. For those of us who had been using the original Google Analytics (now called Google Classic Analytics), Google first gave us the choice to upgrade. But eventually, Google decided to sunset Classic Analytics, and all sites were required to upgrade to Universal Analytics. If site owners didn’t upgrade manually, Google would automatically upgrade the sites. At this point, Google has now upgraded all properties to Universal Analytics.
How Google Universal and Classic Analytics differ with session tracking
Universal and Classic Analytics differ in how they track site sessions by default. While both versions of Google Analytics defaulted the session timeout to 30 minutes, Classic Analytics did not trigger a new session when the session arrived from a referral. So what does that mean?
Distilled wrote a great piece on their blog earlier this year about how Universal Analytics tracks a bit differently from Classic Analytics and how that can affect session tracking. A session measures activities by a visitor on your site during a visit and a period of time.
Since sessions are automatically timed to be 30 minutes (unless you change this setting), site visitors who start navigating your site but become inactive and then resume more than 30 minutes later are starting a second session. In Classic Analytics, if the visitor arrived by a certain channel (such as a organic search) and that session timed out, the visitor’s new session when he/she resumed activity would remain as the original channel (organic search):
With Universal Analytics, however, when a session times out after 30 minutes, if the visitor resumes activity on your website, he/she appears to come from the referral channel, with your domain being the source, as such:
Why is this a problem for organic search measurement?
Ultimately, many of us are using Google Analytics to track traffic and goal conversions, which helps us to measure the effectiveness of our SEO efforts. This is often based on source/medium data in Google Analytics. However, if site traffic isn’t being attributed to the correct source/medium, then you’re not working with accurate data sets.
Most likely, you’re not giving yourself all of the credit your SEO efforts deserve! In the cases I’ve seen, fixing the problem typically meant a 10- to 20-percent reallocation of source/medium data, but this will vary by site and circumstances.
Here’s a screen shot from a recent fix I made to a site and how it affected their organic search sessions:
Over a two-week period, this one site saw an increase in nearly 16 percent of their organic search traffic reporting that previously was reported as a self-referral.
How big a problem is it?
Since the automatic upgrade of all sites to Universal Analytics, I’d estimate that 95 percent of the websites I’ve worked with had not implemented the fix I’ll outline below.
In part, I think this is because, during the manual and automatic upgrades, Google did a poor job of helping webmasters understand the difference in how Universal Analytics would track versus Classic Analytics.
How to fix it
Ironically, the fix is very simple and generally takes less than five minutes to implement. All it takes is adding your domain to the referral exclusion list. To update that list, go to the Property settings under the Admin panel and choose Referral Exclusion List under the Tracking Info sub-menu:
Then add a new exclusion and make it your domain:
If the site is still using ga.js (Classic Analytics’ tag), be sure to update this to the Universal Analytics tracking code as soon as possible, as incorrect tagging can lead to multiple tracking issues in Google Analytics.
Also be sure to also annotate your Google Analytics, detailing what day you made the referral exclusion change. This way, when you run reports (or others do), you’ll clearly be able to account for why there may have been traffic spikes in organic search on a given day