Combining Behavioral UX and Usability Testing Data to Make Website Optimization Recommendations
Combining the behavioral UX and usability testing data to analyze results and make recommendations for website optimization is the final step of the 4 Big UX Optimization steps.
Learn how to analyze the data gathered from the prior steps to make recommendations and improve websites. Using this technique can greatly increase the odds that the correct issues will be identified, and the best recommendations made to optimize conversion and increase ROI.
The four Big UX Optimization Steps are:
Step 1 – Define Personas
Step 2 – Conduct Behavioral UX Data Analysis
Step 3 – Conduct UX and Usability Testing
Step 4 – Analyze Results and Make Optimization Recommendations (this article)
In the earlier steps of our optimization process we have; defined personas, conducted a behavioral UX data analysis to find the WHAT is happening information, and performed UX research and usability testing to determine WHY those behaviors are happening. Now that we have a 360 degree view into the who, what and why of our website data we can now analyze the results and make website optimization recommendations.
Our goal is to combine the WHO information from our Persona data with the WHAT is happening information (quantitative data) of behaviors, with the WHY information (qualitative data) coming from our prior analysis.
In combining these sources of data, we:
- Determine WHO our Persona is, and focus on making their experience better
- Find the commonalities in terms of WHAT the consistent patterns of task-flow failure or confusion are for the Persona
- Identify the most common WHY reasons for those failures
- Combine the above data to create UX optimization recommendations which are then A/B tested
Using Persona Information:
As you know, in Step One of the Four Big UX Optimization steps we defined the Persona or Personas who are critical to your website (and business) success.
Personas are fictional representations of our most common website visitors defined by what critical tasks they are trying to accomplish.
What’s a Persona?
A good UX Persona enables you to be able to answer questions in either design or research. By focusing on the user experience of our Persona, we can be centered on the critical tasks your website visitors are trying to accomplish. This provides the framework for analyzing the UX behavior and the usability of the site.
Why Personas Matter:
Personas are very important to the four Big UX Optimization steps process. That’s because we need to understand WHO we are trying to optimize the website for.
Let me repeat that, we need to clearly define WHO we are trying to optimize the website for.
I can’t tell you the number of times I’ve come across firms where there is high activity in conversion optimization and A/B testing, yet Persona data is not involved at all.
This makes no sense to me.
A/B testing without Persona data is like chasing a wild cat. You and the cat have no idea where you’re going. You’re just chasing a random factor without purpose.
In the business world, most firms who are ‘chasing the cat’ are just trying to move the conversion number higher, but without any clear understanding of:
- WHO is using the website
- WHAT their critical tasks are
- WHERE those critical tasks are failing and by how much
- WHY those critical tasks may be failing
Simply chasing a conversion number by making random A/B tests of various elements on the page is not the most efficient or effective way to produce results.
Stop chasing the cat!
By using the Persona you created in Step One, we can be sure we are focusing on optimizing the website experience for WHO most needs to use our site to be successful. The Persona information also tells us what their critical tasks are.
Now that we know WHO we are optimizing the UX for, we can now move to the next step of our Four Big UX Optimization steps, the measurement and analysis of Behavioral UX data.
Using Behavioral UX Data:
Step two in our four step process was conducting the behavioral UX data analysis. This quantitative behavioral data is our sign-post for danger. The data guides us to identify the most problematic pages on our website and enables us to prioritize those problems in terms of deviation from expected behavior.
Identifying where critical tasks are not performing and the severity of their abnormal behavior is essential because otherwise we may be trying to fix things that are just not that important.
Fixing things that may not be important is a common symptom of randomly A/B testing elements on pages without having data to inform us on WHAT to test. It’s just guesswork at that point, something that’s all too common with many companies who conduct A/B testing. And guessing doesn’t sound like a good thing for you or your business, does it?
Using the behavioral UX data you gathered in Step 2 will identify WHAT elements in the critical tasks are not performing to your expectations. You can also prioritize the deviation of expected behavior versus actual behavior you measured. This deviation can be used to prioritize what elements are causing the most problems. Now you have a prioritized list of what to focus on first.
Using the quantitative WHAT data, we can begin evaluating what is happening and from that narrow down our focus to potential usability or heuristic issues that may be causing that behavior.
Types of Behavioral UX Data Analyzed in Step 2:
- Conversion data from Google Analytics, CoreMetrics, or related log file analysis tools
- PPC Keyword Data – PPC Keyword data is a good source of information that helps define what search terms and phrases are attracting visitors to your website.
- Bounce rate – Bounce rate is the rate at which visitors who land on a webpage immediately bounce away to a different site, often never to return. This is useful for evaluating website content, information scent and engagement.
- Visits by Browser – When evaluating the user experience of a website it is important to know what browsers are the most commonly used for viewing site pages
- Session Duration – Time spent per session
- User Flow – Top navigation flow for website visitors
- Screen Resolution – Top screen resolution sizes your website visitors are viewing your site with
- Etc. Etc. Etc. – There are literally hundreds of other data elements that can be analyzed.
With the prioritized list of Behavioral UX data gathered in Step 2 you know which pages to work on first, and you have a good sense of WHAT is happening and some initial clues as to where to focus on the WHY.
Your next step then is to analyze that WHY qualitative data. And that brings us to Step 3, Using UX research and usability testing analysis.
Using UX Research and Usability Testing Analysis:
In Step three in our process you gathered UX research and usability testing data to provide you with the qualitative WHY for those behavior-challenged pages identified in Step two.
The qualitative data is our tour-guide, that data will help explain WHY those critical tasks are not performing as expected.
Remember that our goal in this step was to observe and record actual people who match the Persona(s) as they conduct critical tasks on the website. Specifically we want to know:
- What works for them?
- What doesn’t work for them?
- What confuses or causes concerns?
- Are their expectations for the experience met? Why or why not?
You may have used a variety of tools to capture this qualitative data including; moderated usability testing, unmoderated testing, 5 Second tests, click tests, card sorts and more.
Analyzing the results of this qualitative data reveals patterns where usability issues may be occurring. By aligning the What is happening data from Step 2 with this WHY it is happening data in Step 3, you have a much clearer picture of the issues and potential causes for website UX behavior.
This leads us into our final step of the Four Big UX Optimization steps, Combining the Behavioral and Usability data and making recommendations.
Combining Behavioral and Usability Testing Data and Making Recommendations:
Step four in our process is all about using the results gathered earlier and identify patterns of task-flow errors.
We combine the WHAT is happening data with the WHY it’s happening data to obtain a clearer picture of patterns of task flow issues and potential reasons for those issues.
Those patterns are very important. They will be your area of focus for building your main hypothesis for the root causes of measured UX behavioral issues.
The qualitative WHY information will typically point to just one or two rather specific usability or related heuristic issues. The WHY information and results from UX research or usability testing will also support this hypothesis and concurrently will suggest just one or two specific fixes for those issues.
Example of Combining UX and Usability Data:
The following examples demonstrate how combining UX and usability data results in optimization recommendations.
The Step 1 Persona was a person shopping for a luxury watch and going through the buyflow to purchase it.
The Step 2 quantitative behavioral UX data coming from the above example page indicated this page was experiencing poor conversion, was exhibiting pogo-sticking (back and forth movement in pages of the buyflow) and high abandonment rate.
The Step 3 qualitative data from UX research and usability evaluation indicated there were multiple reasons why there were issues. These included:
- Top navigation enabling visitors to go back to other website pages outside the form
- Relatively tight visual spacing and lack of whitespace between labels and form fields
- Centered alignment of labels over fields instead of the normal left justified position
- Lack of information around the number of steps in the process
The analysis and resulting recommendations for the page were:
- Remove the global top navigation
- Create more visual whitespace between each form field for more separation
- Left align form data labels
- Visually clarify where in the process visitors were as they completed each step in the buyflow
Always A/B Test Optimization Recommendations
My advice is that it is important to always A/B test the optimization recommendations. A/B testing is the only way to be sure that the optimizations we are recommending did in fact actually improve the user experience.
Finally, and perhaps just as importantly, how you present your recommendations is important.
Too often I’ve seen UX research practitioners who resort to long, boring, word-packed novels with minimal pictures included.
Look. We humans are visual creatures.
And we in the UX research world are typically researching visual expressions of a user experience, task-flow or process.
Why oh why would we bore people to tears with pages on pages of WORDS when in fact VISUAL PRESENTATIONS OF OUR FINDINGS would work so much better!
Ditch Word. Use PowerPoint.
Create screen shots indicating the Personas, the quantitative information from behavioral UX data, the videos of usability testing or screen shots of UX research data.
Use visuals to demonstrate where the issues are, and how they can be made better.
Telling a story visually is perhaps one of the more important things many UX researchers can improve.
Don’t bore your audience with words, use visuals as an output of your research. You’ll find it’s a far more effective way to communicate the user experience optimization concepts you’re proposing. And you’ll find your audience understands your recommendations quicker and with a deeper level of understanding.
Conclusion: Combining Behavioral UX and Usability Data for Website Optimization
In conclusion, combining the behavioral UX and usability testing data to analyze results and make recommendations for website optimization is the final step of the 4 Big UX Optimization steps.
In this step, we analyze the data gathered from the prior steps to make recommendations and improve websites.
The 4 steps in combining Behavioral UX and Usability testing data to optimize websites are:
Step 1 – Define Personas
Step 2 – Behavioral UX Data Analysis
Step 4 – Analyze results and make recommendations (this article)
By combining the quantitative WHAT data coming from the Behavioral UX analysis with the qualitative WHY data coming from the Usability testing analysis, we now have a much more complete set of information to help us make optimization recommendations.
Looking for patterns between the two sets of data, and identifying potential optimization opportunities is the final step in the process.
Now that you’ve had a chance to review all the steps, and understand how they work, go try it for yourself. I think you’ll see how much more effective and efficient this process is for improving websites and apps.