Data-Driven UX with Google Analytics

December 15, 2019 • data-driven, Google

We’ve all heard that “data is gold”, and this is no different when it comes to user experience design.

The potential to use Google Analytics is a UX tool is easily overlooked. It’s relatively easy to understand, data rich and other than set up does not often require intervention from developers.

Google Analytics have several features that are relevant to the work we do as UX designers or researchers, and this article will unpack some of the use cases.

Disclaimers: None of the technical aspects of implementation are covered, but I have included some useful links at the end of this article. I will also occasionally refer to Google Analytics as GA for ease of reading.

Google Analytics standard is available for free and can be used on any web-based or native app.

To get started, register a free account on Google Analytics here, and follow the prompts. Once installed, open your website or app and head to the “Real time” section of Google Analytics. You should now see your activity being reported on.

Metrics and dimensions are the various data points within Analytics and are measured for all users. Dimensions can be explained as an attribute associated with a visitor (for example, their age) and a metric is what is used to measure these attributes (such as the amount of times they viewed a page).

Metrics and dimensions provide you not only with numbers, but they can also be compared over time periods or pulled into custom reports which make them very useful to UX designers for research and A/B testing.

The display of metrics and dimensions can also be modified in addition to comparing them. Some examples include changing the time frame (i.e. hourly, daily, weekly or monthly), sorting by metric or displaying the tables as graphs.

Not all of the out of the box data points are necessarily useful to a UX designer, but there are some key ones that can be indicative of user behavior:

Demographics (Audience): Google Analytics gives you access to several demographic data points like age, gender, location, language and interests which are very useful when trying to understand who your users are. These demographics are based on the data collected from users while logged into their Google account which means that it’s not necessarily a full view, so use this information qualitatively.
Browser, Operating System and Devices (Audience): Useful for determining cross-browser compatibility of your product, as well as studying behavior on different device types (like desktop versus mobile).
Time on page/screen: Indicator for whether users are engaging with a single page or screen in your product. Can also be useful for measuring whether or not your layout is too long (i.e. has an early drop-off) or too short (users don’t spend a lot of time on the page).
Session duration — The amount of time a user spent on your entire app or website in a single sitting. This is really useful for measuring the effectiveness of your user journeys. It can also be used in conjunction with bounce rate, exit rate and pages per session.
Landing page, Page depth, Next page path(s): Also useful for measuring user journeys; these dimensions provide a sense of where a user came from and where they are going.

It’s well known that the speed of an application is instrumental to a good user experience, and Google takes this seriously:

The “Speed Update,” as we’re calling it, will only affect pages that deliver the slowest experience to users and will only affect a small percentage of queries. It applies the same standard to all pages, regardless of the technology used to build the page.

Google’s Site Speed feature is useful for analyzing high level performance**. It is paired with browser type by default, but can also be broken down to metrics like page views or bounce rate. The speed metric is based on a sampled average of your website which makes it useful for measuring individual pages.

Example of page views and load time for the top 4 browsers. In this example, Internet Explorer loads 29.53% faster than the site average — great news for your developers!

** Word of warning: Google Analytics will only track page speed after the page loads, which means that it might not necessarily give you the full picture. The data is also sampled, meaning that one bad load can potentially shift the average dramatically. I would therefore recommend using this data qualitatively to test some very high-level objectives as the data might not be completely accurate.

Marketing and UX teams often work in isolation without recognizing that there is a clear symbiosis. UX efforts can greatly inform marketing teams when it comes to marketing funnels for example, and digital marketing can be a great tool for user research and discovery.

Both Google Ads and SEO are very intent driven marketing mediums, which makes them good candidates for understanding your users. Both Google Ads and Google Search Console (Google’s SEO tool) can be connected directly to Google Analytics and gives you access to some new data points.

Search query for example is the query or sentence a user typed into Google to find your ad or website. It’s available for both Google Ads and Search Console (new sections you’ll find under Acquisition) and is perfect for researching user needs or intent. Other sections that might be useful is Display targeting which outlines how your advertisers are targeting users and can be useful for user research.

These users were looking for a free product specifically

The real power of connecting these data sources however, is using them in combination with other data points. Using search query in conjunction with session duration, bounce rate and page depth for example; will give you a clear indication of whether or not you addressed the user’s initial need or intent.

Events is likely one of the most underrated features in Google Analytics. An event, simply put, is any unique tracked action a user takes within your application or website.

Events consist of three dimensions: A category, a label and an action. Categories are useful for grouping events into themes or types and labels can be used to describe something about the event. Lastly, the action can describe the action a user takes. These three dimensions can be combined in several ways to provide detail on how your users interact with your product.

These three UX events track how users interact with some of this website’s eCommerce functionality.

As mentioned, events can consist of any interaction, but these are some examples that are especially useful to a UX practitioner:

Scroll depth: Measure how far a user scrolled down a page or screen; useful for gauging whether or not your pages are too long.
Interactions: Track whether users are clicking or tapping on certain elements like banners, buttons, menus etc. This could be used to test the effectiveness of UI components, or their placement.
Dynamic loading: Events could be useful when components or content in your application is loaded dynamically — for example a “Load More” button.
Form engagement: Form interactions can be tracked on multiple levels: Validation, completions and submissions
Content engagement: It is also possible to measure how users engage with your content by combining some metrics such as time on page and scroll depth
File downloads
Video plays
Third party: Events could also be used to track events on third party applications, such as Internet enabled IoT devices or point of sales systems.

Events are not set up by default and will require some up-front work. They can be implemented in a few ways but using Google Tag Manager will make your events more scale-able. It is also for this reason that events should be based on objectives.

Google Analytics has three tools for analyzing user flow and behavior: Navigation summary, Behavior Flow and User Explorer.

Navigation summary is a straight-forward view — it provides an overview of the page users came from and the page they are going to in relation to a specific page. It also demonstrates how many users viewed this page first, and how many exited from here. Navigation summary can be found under the Behavior section of GA when analyzing individual pages.

In this example, 30.51% of users came from the home page, and 9.57% of them returned to the home page. This could be an indication that they did not find what they were looking for.

Behavior Flow is an excellent feature and provides a lot of in depth detail of how users move through your application or website. This flow can be viewed based on a couple of dimensions (even events) and can be drilled into on multiple levels. The only draw back of this tool is how it groups pages, as it won’t necessary show you a flow chart of every single page on your website. That said though, you can get to this view by adding filters.

Example of how users moved through this website, from the ***Grade 5 Exam Maths*** page.

The Behavior Flow tool is very useful for getting a broad sense of how users move through your website or app, but User Explorer offers a much more granular view.

User Explorer provides an overview of a single anonymous user’s interactions based on their cookie ID.

There is a substantial amount of information available such as session duration, the amount of sessions and how the user found your product. From a UX perspective though, the most valuable data would be the ability to see the detail on each of the user’s interactions — and this includes all the events they triggered. It is also possible to export this data to JSON which can be consumed by third-party applications for visualization.

Cookies do have a few drawbacks: They are device dependent which means you can track the same user cross-device, and they can also be deleted by the user. There are some ways around these scenarios without impacting a user’s privacy but this implementation can be technical.

Example of a user’s interactions on a website

Sometimes you might want to report on data that is not available in GA by default, which is why it is also possible to create custom dimensions. Custom dimensions can be any data point and can come from any interaction. You might want to report on, as an example, how many of your users are married versus single.

In this example you could set up an event on a form that tracks every time a user selects their marital status. This event would then send this data to Google Analytics as a custom dimension, which would then become available for comparison with other metrics.

An example of adding an author name as a custom dimension.

User testing often requires prototyping and focus groups; and finding people can also be difficult or costly. Google solves for this through their free product called Optimize — a tool used for personalization and A/B testing with a direct integration into Google Analytics.

Optimize has what are called Experiences, and there are four types: A/B test, Multivariate test, Redirect test and Personalization.

A/B test is exactly what it sounds like; test two variants of almost anything in your application, whether it’s a button or a content section. A/B tests can have multiple variants and you can decide to what percentage this traffic should be split. Multivariate allows you to test two sections against one another, redirect tests redirects a set amount of traffic to a different page and personalization personalizes a users experience based on parameters.

What makes Optimize incredibly powerful is the ability to create these tests without making any changes to your application code. Optimize uses a visual editor that allows you to edit content, components and styles as if you were making changes to a live website. These changes are then only applied when the conditions are met.

There is also a paid for version of Optimize available (called Optimize 360) that would allow you to connect Google Analytics Audiences to an experiment for more precise targeting.

The last port of call is to present your data and findings and there are some options to automate this process. Google Analytics have built in reports that can be customized to show and compare data over time. An alternative option is Data Studio, a free data visualization product from Google. Connecting Google Analytics to Data Studio is relatively easy and allows you to visualize any metric or dimension in a variety of graphs and tables.

Google Analytics offers a tangible way to measure UX initiatives and can be a very useful tool in a UX designer’s arsenal. Remember though, that it only provides us with the data and that it is ultimately up to us to find the insights.

Follow me: Medium / Dev.to / LinkedIn / Twitter

How data-driven email automation gets you closer to your customers

September 7, 2019 • data-driven, Email

AUTOMATED EMAIL marketing

When you’re ready to cozy up to your current customers, email automation is the best way to do it.

You already have your customers’ email addresses. What’s more, email is a channel that you own, unlike social media.

Your customers are (hopefully) already in the habit of opening emails from you for receipts, updates, etc.

For all of those reasons and more, there are a lot of opportunities in email automation, everything from increasing customer LTV to reducing churn to converting customers into affiliates and advocates.

In fact 80% of SMBs rely on email marketing for customer retention.

In this post, we’re exploring how you can get closer to your customers with email automation. And we’re providing you with plenty of examples to inspire your automated campaigns.

What is data-driven email automation?

what is data-driven email automation

Data-driven email automation is the practice of setting up emails to be sent to leads or customers based on triggering criteria, such as when a website visitor takes action on your site or a customer has been active for a certain amount of time.

Email automation is all around us. You download an ebook, and you get five follow up emails. You make a purchase, and you get confirmation emails and then weekly newsletters.

These are examples of pretty simple email automations. As we’ll come to find later in this post, emails can be triggered based on even more specific and more valuable customer behavior.

What can you achieve with data-driven email automation?

achieve with data-driven automation

There are so many ways that data-driven email automation can bring you closer to your customers. Whether you want to increase repeat orders or upsell your customers, we’ve got tips on that plus some email automations you may not have even thought of.

Save abandoned carts

reduce cart abandonment

One of the most essential automated emails for B2C retailers to set up is the abandoned cart email. Why? This type of email receives a click-through rate of 40% on average.

Fortunately, an abandoned cart email is fairly simple to set up. You need to segment out website visitors who have added something to their cart but not completed a transaction.

In GoSquared, you would create a Smart Group that is continually updating with people who had clicked the ‘add to cart’ button but didn’t have any transactional history that day.

In this example from Grove, an online retailer for natural household and personal care products, the reader is served up a warm, friendly message and is reminded of the value of what they left behind.

data driven email automation

Don’t just show a picture of the product and add a CTA to send them back to their cart. Instead, you should use this opportunity to sell the product using short conversion copy that puts your value propositions front and center.

Reduce churn risk

reduce churn risk

Reducing the risk of churn is essential for many different types of companies:

SaaS
Digital publishers that sell subscriptions for premium content
Physical product subscriptions

Detecting churn in physical product subscriptions like product boxes can’t be done based on digital behavior alone, because you’re not able to track engagement. Maybe the customer jumped for joy when they opened the delivery box, or maybe they threw your product at the wall. You’ll need to use email to collect feedback on recent deliveries and then monitor and respond to anyone who has given poor feedback.

But with SaaS companies and digital publishers, you can use website analytics to identify people who are at risk of churning. Maybe they haven’t been on your site (or logged into your product) in 7 days, or 14 days.

Inside of GoSquared, you can create a Smart Group of people who match this criteria, and send them the exact right message via email. Share a product guide, send them a recent article, offer personal help—whatever makes sense for your business.

Upsell to new subscription tiers

upsell to your customers

Email automation can also be used to upsell customers to new subscription tiers. You can use different criteria to help you segment out which customers are the best fit to upsell to. Obviously you wouldn’t want to offer an upsell to brand new customers who haven’t gotten value from your product yet.

Here are some criteria you might use:

How long they have been a customer
How many times they login every week
How much they have used the product (minimum number of signatures requested or graphic designs created, for example)

In this example from OptinMonster, the email is sent to anyone who subscribed to the product during a specific period of time.

automated email example

Increase engagement

increase engagement

You can also use data-driven email automation to increase engagement with just about anything.

Maybe you identify that customers who click-through to your YouTube videos are more likely to make a second purchase, or upgrade to a higher plan. In that case, you could use email automation to encourage newer customer (or customers with a low level of product usage) to watch a series of engaging YouTube videos that teach viewers how to use your product.

If you have different types of users, you can also use email automation to segment them based on what part of your product they interact with, and then you can send them content and guidance that’s more relevant.

The opportunities and ideas are endless, and the ultimate goal is to increase stickiness and loyalty.

Grow your affiliate program

grow your affiliate program with email automation

It probably doesn’t make sense to market your affiliate program to a brand new user, does it? A new user hasn’t yet gotten value from your product, so why would they want to recommend it?

However, someone who has been a paying customer of your product for at least two months might be a good fit to become an affiliate.

This email from social media automation tool SocialBee is a great example of how to write in a way that is clear, personable, and engaging. Also, take note of the simplicity of the email subject line.

Automated email data-driven

To segment out the right customers to notify of your affiliate program, you could create a Smart Group based on how long they’ve been a customer, or how engaged they are with your product.

Customise onboarding

customise email onboarding

You can also use data-driven email automation in your onboarding as well, whether for freemium or free trial users in the case of SaaS companies, or for all subscribers in the case of digital publishers.

When we think of onboarding, we think of those critical first few days when a user engages with the product for the first time. If they can’t immediately understand its value, they’ll churn. Chat prompts are smart for onboarding because the communication is more instant, but email can work well too. Here are some examples:

Send an email with a guide or tutorial video the first time someone uses a new feature
Send an email with instructions on an important feature if someone logs into the product after having been away for 10 days or more
Introduce new users to a new feature every day for five days, but remove them from the sequence if they are actively logging into the product

When you trigger your onboarding emails based on user behavior (instead of a one-size-fits-all drip), you’re able to communicate with your customers in a way that matches where they are in their unique journey.

Prompt repeat orders

repeat orders with email automation

Did you know that transactional emails have 8x more opens and clicks than any other type of email and can generate 6x more revenue? That’s why you don’t want to waste the opportunity in your transactional emails.

Check out this example from retailer Sephora, which includes a menu on the top of the order, so that shoppers can click through to look at makeup, skincare, fragrance or sale items. It’s a subtle way, on-brand way of driving traffic back to their website without pushing additional products overtly.

automated email celebrate

Aside from the transactional email of an order confirmation or order arrival, you can also segment your email newsletter and include product recommendations based on customers’ previous purchasing behavior.

Let’s say someone has just bought a shirt from your ecommerce store. You can serve them up with product recommendations for complementary accessories like Coolibar (a line of UPF clothes) does.

recommend products by automated email

Connect with early adopters and power users

early adopters email automation

Early adopters, power users, advocates, influencers, and fanatics are your bread and butter. Through word of mouth, product virality, or social media, these people add more to your bottom line than you could ever measure.

Wouldn’t it be smart to treat these people to something special?

In this example, screen recording tool Loom has created a community just for their power users and is sending out the invites.

email automation engagement

Maybe you want to create a community, host an event, offer a special promotion, provide a sharing incentive, or all of the above.

When you combine customer behavior data with your email automation campaigns, you can give the right message to the right person at the right time.

The 4-step process for getting closer to your customers with data-driven email

four step email automation

Now that you’re armed with plenty of ideas and examples, it’s time to start brainstorming how you can data-driven email automation to deepen your relationship to your customers. We recommend that you follow these steps:

Set a goal – Decide what problem you want to solve or what metric you want to affect
Segment your customers – Use engagement, specific actions taken, time since they’ve been on site and more
Write your messages – Write copy that is personable, on-brand, and most importantly clear
Automate! – Set up the delays and timing for all of your emails—after someone matches the criteria, what happens next

Did you know? Most email automation tools don’t come loaded with customer journey data. From the moment someone lands on your site, GoSquared builds a complete profile of their journey. Learn more.