27
Jun
2016

ADTECH MYTHBUSTER SERIES: DISCREPANCIES WITH GOOGLE ANALYTICS

06/27/2016

Mythbusters_NL_Google-Analytics

Google Analytics is an awesome free tool provided by Google. Anyone can use it to learn more about their web traffic. It’s popular and because it’s by Google, people love it.

That said, there are some issues with Google Analytics that are worth discussing. We often hear complaints that Simpli.fi reporting doesn’t match Google Analytics reporting. Actually, we hear that most third party ad reporting, not just Simpli.fi’s, rarely matches Google Analytics.

Think of it like this. Let’s say I’m on a diet, and my goal is to lose fifteen pounds. But my scale doesn’t measure pounds. It measures kilograms. Could I use that scale to measure if I’m losing weight? Yes, absolutely. If the number on Friday is lower than the number on Monday, I know I lost weight during the week.

But what it doesn’t show me is exactly how many pounds I lost. It can’t show me because it speaks a different language.

Google Analytics works very similarly. It measures web traffic and reports on different aspects of that web traffic. And because of that, there are ways to measure the overall impact of display. We covered this in our webinar “Using Google Analytics to Measure Display Campaign Performance”.

However, the data doesn’t usually match third party reporting because it speaks a different language. Very much like our imaginary scale, it’s measuring kilograms while we’re reporting pounds. In fact, it’s more like we’re reporting pounds and Google Analytics is measuring inches. Again, both would prove weight loss, but in very different terms.

It starts with the way they are each measured. Clicks, as reported by the Simpli.fi UI, are based on ad server logs from the ad exchanges. When we serve an ad, and it gets clicked, the exchanges let us know the click happened. That’s the number we report.

Google analytics uses cookies or page loads to determine what they call sessions, or what we would consider visits. When the ad is clicked, the browser takes the user to the advertiser’s website. At this time, the browser requests the files from the server to display the page to the user. This can include the CSS, images, video, audio and other content on the page. For Google Analytics to work, it must also include the Google Analytics JavaScript file.

A few more things have to go right for GA to measure this click as a session. The browser and/or the device and the security settings for each must support cookies, JavaScript and images. If any of those are disabled either by the device or the browser, or the security settings, GA may not be able to record the session, even though the click happened.

If all of that is working properly, the user clicks the ad, the advertiser’s page loads and the browser sends a request to the GA servers at google-analytics.com, where the session is finally recorded.

Starting to see how the stars must align for GA to record the session?

The difference between clicks and sessions.

I already mentioned that Simpli.fi and the exchanges measure and report every single click on an ad.

If a user clicks an ad, goes to the advertiser’s site, everything works as it should and the click is counted as session… and then that same user leaves the site, clicks the ad again (or even a different ad with the same referral URL) within thirty minutes, GA only registers this as one session. In fact, if that person clicked the ad five times and came to the advertiser’s site five times within a thirty-minute window, Google Analytics would still only call it one session.

What else can cause discrepancies?

User browser preferences can often prevent Google Analytics from collecting data. Little things like disabling JavaScript or images. There’s even a Google analytics opt-out browser add-on that would prevent what the ad exchanges record as a click from being counted as a session.

Then there is server latency. This happens when the user clicks the ad and then either bounces or moves on to another page within the site before the Google Analytics tracking code can execute.

If you’ve ever visited a web page and then left it before everything on that page loaded, you’ve experienced server latency. Depending on the speed of the internet and the servers where the page lives, it’s not uncommon that a click can happen, but a session isn’t recorded because the GA code simply didn’t have a chance to load. This is even more apt to happen if the GA code is placed near the bottom of the web page.

What happens if there’s something wrong with the page after the user clicks the ad? A click is counted, but a session is not. Maybe the user changes his or her mind after they click the ad, pressing the browser’s stop or back button. GA won’t count a session, but the exchange still reports the click.

Other things that can cause the tracking to be inaccurate.

Old Code: There have been a lot of versions of Google Analytics. Old versions may simply be less accurate than new versions. Make sure your advertiser is using the most current GA code.

Bad Code Location: Nine years ago, placing GA code at the bottom of the page was the way to go. Today, Google recommends that it is placed immediately after the opening body tag. Are your advertisers aware of this change?

Missing Code: This should be a no-brainer. Missing code is going to throw things off. But you’d be surprised how often code is missing from entire sites, or even just certain pages or subdomains of the page.

Incorrect Code: Sometimes the code is there but wrong.

Custom Code: If the advertiser wrote the code himself, that leaves a lot of room for error. It’s not uncommon for a single letter to be out of place in the code causing everything to go completely haywire.

Double Tracking: Maybe the advertiser put the new code on the page and forgot to remove the old code. Maybe they used “cut and paste” and accidentally placed the code on their page twice.

Accurate but confusing. 

Something else that might confuse you about Google Analytics is all that traffic you are getting from locations you don’t service. When running a display campaign, it isn’t unusual for you to see traffic from locations other than your own. Traffic from other states, or even other countries, can be the result of ad testing and fraud protection. These clicks usually come from test systems, doing their part to make sure that your ads are running properly across the exchanges. So while they aren’t clicks from humans who are interested in your business, they aren’t hurting anything either. Bottom line, traffic from New York or Asia is normal for a display campaign and nothing to worry about.

Don’t take our word for it.

If you caught our June 6th newsletter, then you saw that we shared several articles and a video on the discrepancy issues of Google Analytics, including how Facebook deals with the exact same problem that Simpli.fi encounters. Articles from third party sources should help you convince your advertisers that while it is a valuable free tool, Google Analytics isn’t the holy grail of reporting.