UX Audit of
Background
Learn Data Insights is a website that prepares users for Microsoft certification exams with practice quizzes and courses on Power BI, Azure, Excel, and Python. Users receive detailed feedback regarding their learning progress as they complete interactive exercises. LDI also offers bespoke solutions to businesses by providing customized training for their employees. The team reached out to me to help them uncover user experience (UX) issues and conduct research that would inform the redesign of the website and product roadmap for the future.
Time Frame
November - December 2022
Platform
Desktop web browser
Tools Used
Notion, Dovetail, Miro, Typeform, Zoom, Fireflies.ai, Loom, Google Analytics
Challenge
Learn Data Insights redesigned their website at the start of the year. According to the team, new visitors and returning users were struggling to navigate the website and understand all that LDI has to offer. The sign-up conversion rate and the number of newly registered users who completed a quiz decreased. To the left is a video that shows the initial state of the website. The new redesign is coming soon!
Goals
Evaluate the user experience of the Learn Data Insights website and identify its points of frustration.
Make the range of services LDI offers quick and easy to understand.
Increase conversion rates for signup and first quiz completion.
The scope of this UX audit was limited to the desktop version as the majority of website traffic came from desktop computers according to Google Analytics data.
Approach
First, I researched Microsoft certification exams and examined blogs and other websites that ranked the best online learning material on the topics that LDI covers. Next, I identified and analyzed 24 direct and indirect competitors, noting their strengths and weaknesses and looking at how they compared on Google Trends and on social media.
In terms of breadth of courses, major EdTech companies like Pluralsight and Udemy are top competitors, but LDI has an edge in terms of depth since it specializes in Microsoft certification exams on data technologies.
(Pictured: "Microsoft Certifications: a competitive advantage you should consider" by Semina Vagikou and "10 Best Microsoft Power BI Courses to Take in 2022" by Archisha Bhar)
I collected as much quantitative information on LDI website usage from the team as available. We looked at data from Google Analytics such as number of new and returning users, website traffic flows, and conversion and bounce rates. After recognizing trends, I asked more questions to the LDI team and we discussed hypotheses on user behavior and information we were still missing that we could begin to track.
A key finding was that the homepage had a high drop-off rate. In addition, the majority of users chose the Data Analyst track and engaged with the Power BI content the most.
Analysis of past survey and design of new one
Upon examining past survey data, I was requested to help design a new survey to gain more insights. We wanted to increase the survey completion rate and collect more demographic user data that would be beneficial for user segmentation.
The original survey on Microsoft Forms received 8 responses over 8 months. I reduced the number of questions and edited what was asked. The new survey we created on Typeform received 32 responses in less than a month, though four are suspected to be duplicate submissions.
The results of the new survey helped validate a few assumptions to complete and revise the primary user persona I developed for LDI.
I conducted a recorded UX walkthrough to identify usability issues. As a part of this process, I performed a heuristic evaluation of the website based on Jakob Nielsen's 10 Heuristics for User Interface Design. After, I completed a cognitive walkthrough where I explored the website step-by-step, aiming to meet the objectives of signing up, onboarding, and completing my first quiz.
Due to time and budget constraints, I was limited to conducting interviews with three power users of the Learn Data Insights website who are experienced data professionals.
One of the selection requirements I met was to interview users located in different countries (UK, USA, and Mexico) to see if there were significant regional differences in their views regarding career development and the role Microsoft certification exams play. The UK and the USA are two of the locations with the highest number of LDI users.
I created a discussion guide and conducted semi-structured interviews remotely using Zoom. Each interview took 30-50 minutes to complete.
I asked participants to describe their professional backgrounds, career goals and challenges, most used learning resources, and their opinions about Microsoft certification exams. Later, participants discussed their experiences using the LDI website and screen-shared how they usually login and what pages they frequent.
Then, I transcribed and tagged the interviews. Finally, I analyzed the qualitative data through affinity mapping.
Findings
Heuristic Evaluation based on Jakob Nielsen's 10 Heuristics for User Interface Design
Using a traffic light rating system, I informed the LDI team which areas required the most attention (red), which could use improvement (yellow), and which performed on target (green).
Visibility of System Status: Overall, users are kept informed of the system status and receive timely and appropriate feedback. Most buttons and hyperlinks change color when the user hovers their mouse over them and/or clicks on them. Still, the "Leaderboard" page takes a considerably longer time to load than other pages. Only the browser loading animation indicates the pending change.
Match between System and the Real World: The website uses specialized vocabulary relevant to the field of data analytics and Microsoft certification exams. However, technical language could be reduced for entry-level users. On the homepage, there are several buttons labeled "LDI Content" without sufficient information presented to elaborate what that includes. It is unclear exactly what is meant by "How To" in the "Blog/How To" of the navigation bar as well.
3. User Control and Freedom: Users are easily able to recover from their mistakes. They can use the arrow buttons of their browser or use the navigation bar to go back to certain pages. There are also buttons like the exit button to get out of certain dialog boxes.
4. Consistency and Standards: The navigation does not support users' mental models of what they expect from such a website. After clicking on one of the four buttons to enter a learning path (e.g. Data Analyst, Finance Analyst, etc.), users land unexpectedly on a page with a lot of text explaining the relevant certification exams. The next steps and CTAs are not immediately clear. The presence and appearance of the navigation bar and side bar are inconsistent. In addition, some icons are uncommon and thus hard to understand in terms of what kind of pages they represent.
5. Error Prevention: It is unlikely that LDI users would commit high-risk errors given the nature of the website, aside from those related to quiz completion. Within each quiz, the user can click "Postpone" to skip a question or click "Question List" to navigate to different ones out of the preset order.
6. Recognition Rather Than Recall: At first, it is not intuitive that clicking on the "LDI Content" button will take users to the dashboard where they can edit their profile, complete quizzes, access other resources, and view their performance data. In spite of this, the navigation bar and sidebar lay out all the sections of the website clearly. Users also need to remember to scroll up to access the double arrow button to expand the quiz window. Aside from "My Quiz History" and "Quiz Analytics," which are a little confusing to read, users cannot see information on which quizzes they started or completed.
7. Flexibility and Efficiency of Use: If there are any accelerators (e.g. keyboard shortcuts), they are not easy to detect. The website collects a lot of data in the "Personal Info" and "My Skills" sections of the "Edit Profile" page. However, it is not apparent how this information is utilized to personalize the user experience (e.g. recommendations for next quiz to take or content to consume). Users also do not have customization options (e.g. ability to favorite pages, choose dark mode, etc.).
8. Aesthetic and Minimalist Design: The colors and overall aesthetic of the website are positive and inviting. The brand identity could still be strengthened as the logo, images, colors, and other visual elements of the website are inconsistent with each other. The look and feel of the quiz window especially differ from those of the rest of the website as it seems more old-fashioned and the font is very small. On some pages, there is too much unnecessary text and/or links to resources.
9. Recognize, Diagnose, and Recover from Errors: New users who do not fill out the "Edit Profile" section are not allowed to complete any quizzes. Users who try to skip entering information into these fields are reminded in a visually unappealing way that these steps are required to continue. At one point, an interview participant ran into an error page when she tried to login using Google that only read "ERROR: Invalid Request." It did not explain what went wrong or how she could proceed to resolve the issue.
10. Help and Documentation: The website has "Get Help" buttons on quiz pages and a link to making a ticket on the dialog box that appears when submitting an answer to a question. There is also a "Support" section of the sidebar. There are a couple of embedded YouTube videos as well, though lengthy. Nevertheless, there is a lack of other help options like onboarding screens, walkthroughs, popovers, and a discoverable FAQ page. There is no search engine for content either.
Key Insights from Interviews
For the sake of brevity and the purpose of this case study, the following is a summary of some of the findings. The final report I created for the LDI team went more into depth as I described users' data career challenges, other resources they were using, and their content and feature requests with relevant quotes from the interviews as evidence.
"The main motivation for me is just enjoyment of what I do. So problem solving. And that probably drives it more than certification."
Attitudes and Behaviors
All interview participants emphasized that they wanted to pass official Microsoft certification exams to verify their skills. They are driven by intrinsic motivation and actively seek out challenges as a way to push themselves. Two out of three LDI users interviewed, who work as independent consultants, expressed the desire to acquire knowledge they can apply directly to their work and improve their job performance.
"It's kind of iterative learning… [LDI] states your gaps. It tells you your feedback, your score. 'You don't know this.' And then it gives you a link to say, 'Here's how you can learn more. Here are the topics you need to learn more about to figure out where you need to focus your time to be fully well-rounded in these tools and in these certifications.'"
LDI User Satisfaction
Every interview participant identified the quizzes as being the most valuable feature of the Learn Data Insights website. They love the quizzes the most because they provide feedback on their performance and suggestions for further learning. LDI users also appreciate the convenience and flexibility offered by the self-paced learning format.
"I found myself hitting dead ends in the website. It was inconsistent navigation. I found things that were there that I didn't even realize were there after using it for like a month. And I was on the website quite a bit…'"
LDI Usability Issues
Two out of three interview participants wanted a search functionality on the website to help them find content. The UI design, especially the colors used and the layout of the quiz window, detracted from the user experience. Other issues mentioned included problems with navigation, repeat quiz questions, and unfamiliar UI components.
Recommendations
With these findings, I presented the Learn Data Insights team with recommendations that I categorized as high, medium, or low impact decisions. Considering our project goals, we decided to prioritize solving for pain points around three main issues: 1) website traffic segmentation, 2) website structure, and 3) onboarding. I also predict that by improving these areas, certain supports requested by interview participants, such as the search functionality, would no longer be necessary.
Change website appearance to reflect user status: At the moment, new users can access the same dashboard as registered users just by clicking "LDI Content" without signing up. I suggested to only display the dashboard to registered users. The website's appearance should change according to user status to help improve their understanding of the navigation and better personalize the user experience for those who make accounts.
Add a "For Business" page and reorganize the content and pages with a new navigation bar: At this time, the features and services for B2C and B2B users are described on the same "Services" page. The information is mixed up together with different calls-to-action (e.g. sign up versus filling out the form for a consultation), which can confuse both target audiences.
To optimize conversion, I advised to create a separate "For Business" page designed to meet the specific needs and goals of enterprise clients. The navigation bar could be decluttered by removing "Home" and letting the logo serve as a link to the homepage instead. Furthermore, I proposed to use the navigation items "Training" and "Resources" (which would have subpages) instead of "Services," "Blog/How To," and "Contact us." It would be better to clearly define the user flow for new versus registered users by separating "Sign In" and the call-to-action to sign up with a "Get Started for Free" button.
Be mindful of colors and use them consistently: I created the following mid-fidelity wireframes to illustrate these proposed changes, as well as to show how the website would look if more of the LDI logo's colors were used. Though visual design was not the main focus of this project as I am not a designer, I thought this was important to note to the team so they could reflect on how to strengthen their visual identity in the future.
Plus, interview participants mentioned that some of the colors used on the website, especially in the quizzes, and the lack of contrast made the content difficult to see. I recommended to use the same color for particular buttons (e.g. red for call-to-actions like "Sign Up") to maintain consistency as well.
Add onboarding screens with the final one guiding the user to complete a practice quiz: When new users first sign up to Learn Data Insights, they are immediately taken to the "Edit Profile" page and required to fill in their personal information. Asking users upfront for so much of their data in this time-consuming manner can feel intrusive and unpleasant, particularly if it does not seem clear to them that the requested information is relevant to the service LDI is providing. This onboarding process could contribute to a higher drop-off rate.
Because of this, I suggested adding onboarding screens that walk the new user through the process of setting up their account and completing an activity on the website in no more than four steps. The final step would be to complete a practice quiz, the most valuable activity according to interview participants, to see how it works. Ending on this high note follows the UX principles of progressive disclosure and the peak-end rule as the user finishes the onboarding process feeling engaged and accomplished without overwhelm.
Low-fidelity Wireframes of Onboarding Screens
Improve the dashboard for registered users: When a registered user logs back into the LDI website and clicks "My LDI Content," they are taken to the "My Quiz History" subpage under the "Quiz Performance" section. There is no central page for the registered user upon re-entry to feel like they are 'home' on the website. The token balance appears like a notification message in the upper lefthand corner, and the navigation items are jumbled up together in the upper righthand corner.
Below is a mid-fidelity wireframe of my proposed, new dashboard design. I suggested making the sidebar collapsible so that there would be more space for the quiz window. The font size of the quiz text would still need to be enlarged to improve accessibility.
"I’m not overly interested on how I benchmark against others. That doesn’t bother me… the leaderboard wouldn’t bother me either. What I’m more interested in is my own performance as opposed to other people’s." —Interview participant
At the moment, LDI users need to use the menu to search for the activity they would like to start or continue every time. Thus, I suggested to add a feature that would allow users to enroll into certain tracks or favorite/bookmark activities that would show up at the top of the new dashboard. This would allow users to quickly access their preferred content and provide the team with more data they could leverage to make recommendations and/or refine their content strategy.
Instead of the Leaderboard, I advised the team to focus more on highlighting individual learning progress and website usage (e.g. quizzes completed, streaks, etc.) as LDI users care about their personal growth. Competition with others could also feel demotivating for entry-level users.
Outcome
"Asami is a true legend at her art of understanding user motivations and designing experiences to delight! She looks at things from multiple angles, including commercial, heuristics and design, and is able to pick out themes from multiple interviews to bring them into a cohesive framework of recommendations."
The results of this UX audit will be used to inform a review of the site's information architecture and discussions with the development team regarding testing, final design, and implementation. The launch of the redesigned website is set for early 2023. Thank you very much to Rishi Sapra and Naveen Ganpat for allowing me to share this experience as a part of my portfolio.
Reflection
This UX audit went very smoothly thanks to the trust and openness of the Learn Data Insights team and interview participants. Plenty of information was available from the start to better understand LDI users and their experience using the website. The main challenge was digging deeper into user motivations to discover what they feel is LDI's unique value proposition and what more could be done to achieve greater impact. If allotted more time, I would interview non-users who are data professionals and conduct usability testing to further validate the problem hypotheses and design solutions. I would also work on improving the content strategy and exploring the potential opportunity of building an exam prep community. I look forward to seeing what the LDI team implements in the new year!