If you’re new to looking at Google Analytics (GA) for working out what is happening on your website, one of the first things you’ll have to decide is which metric to focus on. GA offers three types of traffic metric by default: users, sessions and pageviews. You can also set events manually on specific pages or interactions. Here I’ll go into the differences between the standard metrics, problems you might encounter, and how you can think about using them.
If you’re looking at signups or purchases you’ll want a metric that will tell you how many people have visited. GA’s users metric attempts to deduce this and tell you exactly how many individuals have come to your site. This will work pretty well if your user journey only consists of single pages where the user has a fixed route to reach the goal.
However if you’re looking at a group of pages, due to the way Google Analytics is built to record pageviews as its primary metric, it doesn’t work out users particularly accurately. To take an example, a holiday website wants to know how many individuals visit the country pages. On a given day they only have one user who visits the France, Spain & Italy pages. They use a regular expression to look at the aggregate of these pages in GA and return the result of three users because it counts back from pageviews and a page can’t be viewed by less than one person.
Another thing that tracking software and people doing the tracking have long wanted to solve is the same user, different devices challenge, i.e. a person visits on a mobile and a desktop in the same day but is counted as two users. Without lots of invasive tracking (which many web users are rightly sceptical about) it isn’t possible, though with their universal analytics tracking and having many Android users logged in across their devices, Google probably does know this. They don’t tell us though, so as a rule ignore the users metric in GA and take sessions or pageviews to represent traffic.
As UX designers we tend to want to focus on the user-level to get a sense of what individuals are doing on our site but I’d always take the GA users metric as indicative and not rely on it too heavily. If you want to understand user behaviour in more depth, I’d recommend using Mixpanel, which shows you user journeys and has the functionality to link up visits across different devices.
At the opposite end of the scale to users, pageviews are what GA is built to track and they are quite literally the number of times that page has been viewed, regardless of which user viewed it. So if a page gets viewed 100 times, that could be from 100 different users or just two, thus leaving it open to being skewed by one or two people. This is rare though and it generally averages out over time to be a fairly reliable metric, and becomes more reliable the more users you have visiting your site. Traditionally it was also the recognised metric for advertisers as they pay out revenue based on the number of times an ad is viewed or pageviews that page had.
At first glance pageviews might seem a bit abstract to UX designers but on sites where content is important (like magazine/news sites) it is a great measure of popularity of different pages and the best way to compare them. It is also something you should look at in combination with users or sessions. Ideally you want your pageviews to be a lot higher than either of these numbers. If the number of pageviews is the same as your sessions this suggests that people are only looking at one page per visit and they’re not very engaged.
Sitting between users and pageviews is GA’s sessions, which they use as the default measure across the interface. A session is a period of browsing regardless of length, logged by a user triggering a cookie and reset after 30 minutes of inactivity or if a user leaves and returns via a paid search campaign. So if a user visits the site at 11am and takes a journey around before leaving and returning at 2pm this would be counted as two sessions.
As the users metric isn’t particularly accurate and pageviews is a bit too granular, this is the most clearly defined person-centric metric you can get with GA. Think of a session as a period of user intent (the intent could be just to learn more or it could be to buy but generally they’re separated by session). If a user can do something meaningful in each session on your site, like save something or even buy something then it’s a very good one to focus on. It’s the metric that funnels use as well so is a good one to use if you want to be consistent with them.
This is an excerpt from my new book, Data-Driven UX Design: Metrics, which explains how you can be smarter with your design by employing a bit of quantitative data.
Sign up to my mailing list and you'll receive my guide to tools for evidence-based design AND my ecommerce UX cheat sheet.
I explain why you should track your site analytics in your own dashboard and give a step by step guide to the quick way to do this.
I explain a handy technique I have for user-centred design when there's no budget for full user research. Turn your analytics data into outline personas to keep your work focused on the user.