5 Ways To Use Google Analytics for Your UX Research

Originally published at: http://www.sitepoint.com/5-ways-use-google-analytics-ux-research/

As users visit your website while you sleep, do you ever get a slight gnawing feeling – like you’ve left a small child alone while you went for milk?

Let’s face it – you really have no idea how they’re behaving. Are they wandering lost? Are they stuck somewhere? Are they doing their chores?

That’s how I feel sometimes. Your users often get misdirected, lost or simply don’t know what they are doing at your site. What’s more, they may break something or get hurt themselves by making bad decisions (e.g. delete some important data of theirs).

I’m part of the Dragdis build team, a tool designed specifically to help creative professionals collect and organize their inspirations. If you’re designing for seasoned creative pros, you need to think especially carefully about interactions, and make every usage scenario perfect.

We like to say “Your path has to have flow

The truth is, no matter how well-researched your UX decisions may be, it’s impossible to predict all possible use cases. So, we need to obtain user behavior feedback as early as possible to make the appropriate design adjustments.

Furthermore, if you’re involved with a startup, your goal is to close that feedback loop for Lean UX as rapidly as you can.

So, in this article I will cover basics how to get feedback (statistical data) about your UX from the moment you roll out your design.

Why Google Analytics (GA)

Though Google Analytics is widely considered a tool for marketing purposes, it offers a powerful feature set that may well be salvation for any UX researcher needing immediate insights without adding a new layer of tools.

Here’s why we ultimately selected GA as our main UX research tool:

  • While most UX tools gouge your budget, a basic GA account is free of charge. There’s no need to go Premium, as the free data is more than enough for what we need.
  • The flexibility GA provides is almost limitless. You can set it up to observe usage trends, actions, track errors… too many to list here. It’s really up to you to extract answers need.
  • Visualization of your data is one of best ways to understand it. GA allows to you slice and dice your data in multiple ways and then display it in a wide range of charts.
  • It’s simple to deploy. Paste in a small snippet of JavaScript and you are ready to go.

Of course, GA can’t do everything. Tasks that require off-site data such as eye tracking or observations during usability testing will need be conducted with other UX tools.

Basics on GA tracking structure

Before you begin with your user analysis with GA, you should be familiar with basic levels of GA tracking. In your GA, you can have multiple Accounts, Properties and Views:

  • The Top level is ‘Account’. You should use this to separate individual projects or websites you want to track, such as example.com or test.org
  • Properties’ offer you a way to separate concerns in a single project. Let’s say that example.com has some public pages for user acquisition and internal pages for already registered users. These should become separate properties since they have different purposes. In that way your user flow and metrics will be kept clear and distinct.
  • View’ is just another level of your data separation. It is useful if you want filter and additionally process incoming data from your website to see different contexts.

Hint: if your site has different sections for different user groups (e.g. registered users and the new visitors), you should always split them into different GA properties – otherwise you will see inaccurate conversion rates for your goals.

Things to track

When you are setting up your GA or configuring a report, first you should have a clear idea of what you want to discover. Here are the 5 simplest tracking features you’ll need for basic UX analysis.

1. Standard Pageview tracking for starters

By default GA provides you standard tracking code that observes how your user is navigating, along with some parallel data that tells some details about your users (i.e. the tech they use, how they arrived at your site). Basic tracking allows you to answer these simple questions:

  • Where did users enter and where do they leave your website?
  • What paths did they take through your site?
  • What technology did they use (mobile or desktop, screen resolution, OS, etc…) ?

Since standard tracking is more oriented towards marketing, generally you will likely only receive superficial insights from it, as it tracks only actual pages, and not the user interactions WE want to know about.

However, we can easily make it more powerful by pimping up your tracking code with virtual page hits to observe user behavior in more detail. ‘’Virtual page hits’ means you have to set GA up manually by calling a tracking function every time a user performs an action what you want to track.

For example, if you want to know if your users are opening a particular tab or loading another gallery item that does not reload your whole page, you can just add a single line of code in your JavaScript, that tells GA to track the particular interaction as a page hit.

ga('send', 'pageview', '/your-custom-pageview');

2. Behavior flow

The shiny new pearl you get when you have your pages set up is the Behavior flow (Under Behavior section). This is a GA report that tells you exactly how your users are navigating your website.

This particular report is useful to:

  • Analyze how users behave at a particular point of your flow
  • See what happens immediately before/after users take a particular step or visit a particular page
  • Investigate the precise sequence that users perform different interactions
  • Isolate whether users are taking any unnecessary steps, or returning to previous states for any reason
  • Determine which steps in your scenario fail most often (when the users doesn’t behave as you intended)
Continue reading this article on SitePoint

There’s a lot in this piece – I considered cutting it into a 2 parter but there was no natural place to cut it – but it’s a really useful read.

I think marketers own a lot of the conversation around GA, but the techniques Petras talks about here are super useful for designers and UX peeps.

There is a great deal of very useful information in this article. Some you can find in other places, but extremely helpful when deiging a site. The key is of course knowing what users are after and this post gets to the bottom of how to reveal this key information. I would like to know if there is a way to get around the HTTPS search masking of google keywords? We are seeing that most visitors to our sites have used keywords to find us but they are masked because of the SSL connection they have while logged into their Google accounts. I’ve seen ways to get around this posted on GA site but none are practical or helpful. Any techniques we can use to get the keywords in GA from https searches? Thanks.

Always split into different properties? Are u utterly mad? That is so far beyond any reason and best practice as possible.
One property, and maybe several views if you dont find the insight by using segments, which is in almost every case a whole lot better.
And just install the script and you are done? Wtf?! Configuration to collect correct data is an absolute necessity.
Please, do not write about stuff you dont know enough about. Please. Stick to UX design. I hope for the love of god, you are way better in that area.

As a marketer I’m happy to see this post in the UX section. It’s been a struggle sometimes to convince product managers who are used to creating classical user personas of the importance of the funnel as GA would demonstrate it. So thanks for this one.