I heart Google Analytics. Really. There is so much benefit in using it across departments to understand gaps in your content, design, development, and user experience. However, as versatile as Google Analytics may be, in some cases it’s better used as a supplemental diagnostic tool instead of your primary diagnostic tool. Given the overabundance of articles that highlight “X Ways You Can Use Google Analytics to Improve UX,” I wanted to provide a different viewpoint – 4 Reasons You Shouldn’t Use Google Analytics UX Metrics”
You Can’t Accurately Tell Demographics
Google Analytics launched the Demographic and Interests reports in 2013. These reports allowed you to see user demographics by enabling advertising reporting features for your Google Analytics property. Google collects demographic information three ways: Third-party DoubleClick cookie, Android Advertising ID, or IDFA (for users who specifically opt in or haven’t yet upgraded to iOS 14). With the continuing deprecation of third-party cookies and the increasing privacy restrictions around mobile device identifiers, it’s important to note that this reporting only shows a subset of your users. Even in Google’s test Google Analytics property, reporting shows for 40% of all users.
The fact that you’re only receiving a portion of demographic information doesn’t seem completely detrimental to your UX analysis. However, having that extra demographic information can be beneficial. As personalization comes to the forefront of every marketer’s roadmap, understanding how user behavior differs between segments will allow you to provide more meaningful personalization and a more meaningful A/B testing roadmap. A tool specifically designed to measure UX may offer the user the ability to self-identify, giving you that granular segmentation data.
Scroll Depth Doesn’t Allow for Analysis At Scale
Scroll depth is one of GA4’s enhanced measurement metrics. Enhanced measurement means this particular reporting feature can be enabled with just the click of a button. This dimension gives marketers a layer of insight they never had before with no additional coding, but it has limitations. With enhanced measurement, the scroll depth event will only fire when a user has reached 90% of the way down a page. While this doesn’t render the feature completely useless, it doesn’t give you any indication of people actively engaged and willing to scroll 60% or 80% the way through your content.
The 90% limitation also makes UX analysis at scale a little tricky. For example, 90% of a page scrolled on desktop may only be 50% of the same page scrolled through on mobile. Even within mobile devices, content on your iPhone pro-max will look a little different than your iPhone mini. This issue gets compounded with your different content lengths. While some pieces of your content are lengthy 1,000+ word blog articles, other content pieces could be your short and succinct 250 word “Contact” page.
Before I go further, I’d like to demonstrate the context word count adds with a chart. Below are four blog posts and the percentage of users who scrolled 90% of the way down the page.
| Page Name | Percent of users who 90% Scrolled (Desktop) | Percent of users who 90% Scrolled (Mobile) |
| /blog-post-1 | 85% | 70% |
| /blog-post-2 | 30% | 15% |
| /blog-post-3 | 40% | 20% |
| /blog-post-4 | 75% | 65% |
Based on the data, you’d say that blog posts 2 and 3 are uninteresting to users. This may not necessarily be the case, though. Let’s look at the same data set with a “Word Count” column.
| Page Name | Word Count | Percent of users who 90% Scrolled (Desktop) | Percent of users who 90% Scrolled (Mobile) |
| /blog-post-1 | 250 words | 85% | 70% |
| /blog-post-2 | 1,000 words | 30% | 15% |
| /blog-post-3 | 900 words | 40% | 20% |
| /blog-post-4 | 400 words | 75% | 65% |
Unlike the first chart, it’s obvious to see that users don’t like to read more than 400 words. The length gave us additional context on the user experience not built into Google Analytics UX metrics.
Now, in order to get around the first two issues with the scroll depth event, it’s possible to build your own custom scroll event (you’d have to anyways if you were using a Single Page Application like react). It’s also possible to add a custom dimension that t-shirt sizes your articles (small, medium, large, etc.). These custom events will get you closer to understanding how people consume your content, but it still leaves a gap when analyzing the UX on your page.
What if people spend 4 minutes on paragraph five and 6 seconds on paragraph two? Where did people stop reading your content? Do they always leave the page when your Newsletter Subscription pop up appears? Are the ads in the middle of your blog article a deterrent? Unfortunately, these aggregated UX insights can’t be built into Google Analytics UX metrics.
Average Time on Page Is Unusable
I wrote another blog post comparing the measurement of time in GA3 vs. GA4. To calculate the average time on page metric in GA3, Google would subtract the time you visited page 1 from the time you visited page 2. The inherent flaw was that Google Analytics had no way of measuring the time someone spent on your exit page because there was no subsequent pageview from which to subtract. This flaw meant average time on page was always underreported or not reported at all.
While GA4 stepped up their game in how they measured time, average time on page, at the time of writing this, is still inaccessible to the average user with the GA4 interface. Advanced users who have knowledge of BigQuery can calculate the difference between pageview timestamps, but those without SQL knowledge have to use built-in GA4 metrics, like engagement rate.
In GA4, engaged users are ones who a) had a visit longer than 10 seconds b) visited 2+ pages or c) completed a conversion event. This metric doesn’t allow for more granular UX analysis, though. You’re either engaged or you’re not. A user who commits to reading your blog for 11 seconds is considered just as “engaged” as a user who reads your blog for 10 minutes. Instead of relying on time metrics in Google Analytics, a heatmapping software can provide more accurate UX metrics.
Unable to Determine Where Clicks Occurred
Within Google Tag Manager, you can fire a Google Analytics event on every single click that happens on your website. This is powerful for those analysts who get the question “How many people clicked this random button on this specific page?”. This feature is supercharged when developers standardize their code and you insert dynamic variables into your event names, such as link text, button classes, IDs, and form elements into those events.
As much data as the All Click Events tag collects, it has limitations when trying to perform aggregated reporting on UX metrics. The main limitation is that users don’t always click within specific buttons or areas of your website. This is best illustrated with an example. If you have a large image on your homepage (your website hero), it could have a standardized class of “class=background.” Because your hero is such a large image though, you won’t be able to tell the difference between a click in the top left of the hero and the bottom right of the hero.
Another example of the limitations faced with click tracking is when users try and click on text on your website. If someone is trying to click on text within a paragraph that isn’t clickable, your event tracking would only show that someone clicked on the paragraph, not that a specific word is misleading someone into thinking it’s a link. In contrast, using a heatmapping tool will give you a better picture of where people are clicking, faster, with less work.
Summary
Google Analytics is a fantastic tool, but take caution when using it to analyze your user experience. The lack of segmentation, aggregation, context, and granular insights can mislead you into thinking something is working, when it’s actually hurting the customer experience and your website conversion rate. I recommend always supplementing your Google Analytics data with a heatmapping tool like Hotjar to resolve the gaps in data mentioned above and help give you a more holistic view of user experience on your website. I’ve used both the free and business plans for over 6 years now, and I’m proud to now be part of their partner program. Sign up for Hotjar using this link and get an extended business trial. By signing up using this link, I may receive a commission at no cost to you.