Website Redesign UX Case Study Part II: Usability Testing
#UI #UX
Project Timeline
Sep 2021 - Apr 2022
My role
UI/UX
Tools
Pencils & Papers, Google Docs, Google forms, Google Meet, Miro, Kardsort, Figma, Maze
Deliverables
Empathy map, persona, user journey, user flow, task flow, wireframes, design system, prototype, case study
I chose to do remote unmoderated usability tests with Maze; because it is more cost and time-efficient. With Maze, it can record testers’ screens and their expressions at the same time. Since my Figma file is too big, I have separated the file into three, thus I am conducting three usability tests with each tester during each iteration, as listed below.
Before the test began, I tested my prototype with a friend who is also learning UI/UX for some feedback. Most of the findings are minor, for example, a new path (search) should be added when people are looking for the Lerhamn table on the homepage.
After the internal test, I have then tested the prototype with five testers, of one who was not able to record his screen when doing the test. I have run two iterations of the tests.
According to the comments from the tests, some changes were made based on the Severity Ratings for Usability Problems. For example, instead of a drop-down list to select the year and month for the delivery date, the left and right arrows were added.
And before choosing if the users need assembly or not, the delivery date is greyed (cannot be selected). This improved the misclick rate in the next iteration.
Between this and the last iteration, A/B testing was performed for the shopping cart page. Instead of selecting which products to check out, four testers prefer to delete the products they don't want. When they click delete, a confirmation popup will appear. Click here to try the prototypes.
And according to the comment from a tester, numbering the sections makes the process smoother.
We can see that between the first iteration and second iteration, the misclick rate improved by -3.4%, the direct success rate increased by 20%, and the average duration to complete the task decreased by 36.85 seconds.
Comparing the satisfaction rates from the second iteration with the satisfaction rates for the original IKEA US site and IKEA Taiwan site that were gathered from the survey, we can see that both improved.
Try the three versions of proceeding to the checkout process (on laptop/desktop): checkout Lerhamn table, Stefan chair, and Fado table lamp. Which one do you prefer?
Due to my background in Bachelor of Business, I sometimes focus more on what business wants than what users want. Therefore, I didn’t conduct quantitative surveys and screen interviewees for this project. The interviewees I selected are based on secondary research and my hypothesis, and I realised that they are not dynamic enough. I learned that even though doing primary research takes time, empathising with the users is the most important step when designing services for users. Also, setting out clear hypotheses and goals before starting primary research will give a clear direction and save a lot of extra work.
I have also learned that throughout the design, I should always keep my primary persona, Alice Lee, in mind. What does Alice want? What will she do? This way I can really design a product that empathises with the users.
With the usability tests for my prototype, when a scenario is given, it will be more close to the real-world situation than just giving instructions. Also, I changed the wording on the instructions for the second iteration based on the testers' feedback, this should not have been done in order not to let other factors affect the test results.
With my proposed solution, I have added the "delivered within 24 hours" feature to be competitive with the existing competitors, but now I feel that this is a more strategic part of the business and would need more insights and research.
Lastly, since this is my first UI/UX project, I tried to use as many tools as possible. But looking back, some of them might not be necessary.
I have separated this project into five phases, as listed below, in the beginning, and due to human resources constraints, I was only able to work on the minimum viable product (phase 1 and a bit of phase 2).
The next steps would be to work on the elastic demand, and the different breakpoints, and to develop an app for IKEA Taiwan.
Some elastic demands include adding features like inspiration collection functions (similar to Pinterest) and to cooperate with credit card companies for users to get rebates.