A closer look at 4 simple yet effective uses of analytics data to improve user experience for designers and UX specialists that save so much time.
Whenever I work for a new website to improve UX, I get asked the same question "What are the most important metrics that inform UX?"
This question is not easy to answer because it depends on the overall website's objective and are related to your brand, what you want to achieve and current situation.
Also, another issue with analytics is that it can become a huge black hole of interesting data without any actionable insight. There are lots of free analytics software with massive amounts of data, which means you get distracted looking at those that I define "vanity data". Most of the times, analytics is seen as "let's have a look at how the website is doing" kind of task. This can lead to huge distractions and waste of human resources, so at the end, even if they are free, these software cost you in terms of resources.
This is especially true for analytics beginners, who are easily impressed by the amount and diversity of data. In particular, they get stuck with the 3 following issues:
Number of metrics: so many things can be measured, which one are really tailored to your brand?
Metrics definitions: so many metrics are similar to each other, think of, for example, bounce rate and exit rate, sessions and users. Which do best answer a specific questions?
Dashboards: which dashboard do you need to answer your questions?
This last point in particular is the main problem. Many UX researchers and designers are looking at the dashboard as a tool rather than as an helping hand.
"Here is the complete overview of the website".
Sounds familiar? So familiar and so pointless.
I would recommend to step back for a second and think of what you really need in a dashboard, which analytics data, which metrics are best fit to support your work, to support the users' funnels, web design and the final aim of your website.
Luckily, I have been in so many of these situations and I know exactly which uses you can make of analytics data and in this article I share 4 uses with you.
In summary these are the four uses of analytics for UX specialists, web designers and web developers:
Analytics for spotting issues
Analytics for experimentation
Analytics for investigation
Analytics for usability
Use number 1: spotting issues
Web designers and UX specialists often work with conversion rate optimisation specialists, developers and data analysts when they need to launch a new website, refresh a page or add a new feature. UX teams' performance in many companies is measured by the ability of increasing engagement and they need analytics data to measure it. So the analytics guys will send weekly reports to UX teams to monitor performance.
Rather than sending reports with pre-determined metrics, how about we add a section for spotting issues and drive further investigations?
So the type of weekly dashboard I would recommend to be shared among the teams would look like this one:
Main website goals: a look at how the website is doing on a very high level, for example number of conversions, leads vs the desired number of goals. This last point is very important. It's useless to look at the number without benchmarking it against the goal you want to reach.
Micro conversions: these are the supporting actions, those that support the main goals. For example, funnel reports, percentages of drop-outs, visiting a specific page, clicking a particular button/ link, completing a form; again, as above these goals need to be compared against those you WANT to reach;
Metrics: pure analytics data with metrics to ascertain that the actions are actually happening as intended. These are very helpful to UX, developers and web designers to identify issues.
Descriptions: for those not familiar with the terms, a short description of what the metrics and what it measures.
Here is the template you can use for your weekly reports.
In this table I suggest using scroll depth as opposed to avg. time on page for measuring if users read your content. To know more about scroll depth and how to track it, check my short guide.
Use number 2: Experimentation
UX and web designers can look at the table above and identify issues related to the main goals and supporting conversions. Normally conversion rate experts will do this job and recommend the other teams on how to implement something and fix an issue. However, also UX, web designers and developers can have a look at the table and formulate hypothesis on what needs to be changed.
What role does analytics play in this process? You can use analytics data to prove your hypothesis wrong and right.
Experimentation included various steps and here is what I normally use to test my hypothesis:
The investigation step: in this initial phase, I look at the dashboard and compare the numbers and conversions with what I need to achieve on a weekly basis. For example, if I only have 1,600 signs-ups against a weekly goal of 2,000, I know that this issue needs closer attention.
Hypothesis formulation: I might have an idea of what is going wrong with the sign-ups on the landing page, but my idea is just an opinion not validated by data.
Testing: in this step I set up an A/B test to validate or invalidate my opinion on a small percentage of traffic. I can test the current version of the landing page vs a new version with my changes included and measure it using the same analytics metrics as in the table, namely scroll rate a d page views.
Results: I can let the experiment run for 2 weeks and gather enough amount of data to support or contradict my opinions. If I am right, I can implement the changes to 100% of the traffic. If not, I learn from the data and develop a different test.
Having briefly explained the experimentation, now we need to find out what exactly is causing issues to my website. In other words, what do you need to experiment?
You can try to make quick experiments by looking at your measurement dashboard to identify quickly issues. However, you need to go deeper in your analysis because even if you find that lower signs-ups are due to a decrease in page views, you still don't know what caused the decrease of page views.
So in the next stage you need to identify typical analytics issue categories.
Use number 3: Investigation
This is the third use of analytics data you can make as an UX specialist to avoid wasting time and efforts in trying to find what to look at.
In my experience, I normally look at these issues categories when investigating more on a website:
1. Traffic issues
In this analysis we want to determine if the decrease in page views (for example) is due to a specific traffic source by using custom dimensions. You can do this in Google Analytics for free. Traffic sources can be organic traffic (Google, Bing), referrals, direct traffic, specific campaigns or email traffic.
If you use Google Analytics, select your property, then you can go to Behaviour> Site Content> All Pages to open the report on traffic to pages.
Then click to "Secondary dimension" and type "source" in the bar to open the drop down menu and filter the traffic to those pages by source only to see where the traffic si coming from:
The next step is to insert the page URL in the search bar down on the right.
Now you trigger a report with traffic sources for the page you need, eliminating all distractions and you should do this within 1 minute. It's that simple.
Remember, this investigation into page views was just an example of what you can achieve with Google Analytics. If you want to find different data you can also look at the Landing Page report.
2. Technical issues
In this investigation we want to find out if you have a technical issue on a page by checking the number of events.
The Events report lists all events on your website that have been tracked. If you select one Event you can see all metrics associated to it and find out whether there is an issue.
How do you know if there is an issue? The report can show lower than normal events or unusual metrics data.
Go to Behaviour> Events> Top Events and then click on the event you want to check.
Now compare total events with metrics:
3. Navigation issues
With this report you can check whether buttons, clicks and other events related to navigation have some issues.
Note: to enable this report I would recommend to track buttons click with Google Tag Manager. Luckily, I have written a guide on this.
Alternatively, if you can't track clicks to buttons, you can simply check a different report on Google Analytics called Navigation summary.
Go to Behaviour> Site Content > All Pages.
This report details from which website pages users came before visiting the page of interest and where they went after visiting that page.
4. Content issues
With the following report you can find out whether new wording on landing pages, product pages and other types of pages important for your brand can bring any benefit.
Note: to see these stats you need to enable one of the following advanced functions:
Enhanced Link Attribution on Google Analytics.
Enhanced Link Attribution with Tag Manager.