I use Amazon obsessively to research and buy all my products — I even considered myself a power user, up until I discovered a feature in the mobile app that I had never noticed before. This made me curious about how other customers were using certain features and inspired me to conduct a UX case study on Amazon’s iOS app.
Quick usability testing revealed several issues with key features in the app, so I designed solutions to solve those problems and validated them with additional usability tests.
Originally published on Medium - featured on UXDesign.cc, Sidebar.io, and other UX design blogs.
Amazon is my go-to place for discovering, researching, and buying the majority of all my products. One thing I love is being able to use my favorite service on the go with the Amazon mobile app. Recently, while shopping for some products at a store, I discovered a feature on the mobile app that I didn’t even know existed — the scan feature. It is more powerful than your basic barcode scanner as it can recognize shapes and objects and surface the exact item in the mobile app, or find something very similar to the item you were looking for.
I was really excited to find this feature because it made product research and comparison an absolute breeze when shopping for products in stores. I did wonder, however, why it had taken me so long to discover the feature in the first place. When had other users figured out the scan feature? Was it just hard to find? Was it possible that there were other features in the app that I wasn’t aware of?
How are other people using the Amazon mobile app, and what issues are they running into?
To get answers to these questions, I decided to take action and do some guerrilla usability testing on the user experience of Amazon’s app.
To identify and solve real UX problems, I followed this design process:
To test the user experience, I asked 7 people to go through a series of pre-defined tasks on the mobile app as I recorded their actions and took notes on anything they said or did during the usability test.
All of the users had previous experience with shopping on Amazon on desktop, but had little to no experience with the mobile app.
Before I assigned the tasks, I set up the scenario below:
“You’re in a bookstore, and you see this book that you want to buy. However, the price looks a little steep and you’re wondering if you could find a better deal online.”
I brought a physical book with me to make the experience more realistic, and also to provide users with several ways to search for the book (search bar, scan, voice, etc.).
Below are a few of the tasks I had the users go through:
I reviewed the user interviews and usability tests to identify any pain points that users were running into. After writing down the insights from the usability tests, I grouped them based on similarity to determine which issues were the most common:
Since this was a personal project, I took it upon myself to determine which would be the top design problems to tackle based on potential impact to business goals and user needs.
As Amazon is an e-commerce company, I assumed that revenue is their top KPI, so I kept that in mind as I narrowed down the problem areas and prioritized them based on potential impact on revenue:
After prioritizing the pain points, it was pretty clear which issues were the most common and would also have the highest impact on revenue:
Reading reviews are a key part of the online shopping process — a quick Google search shows that 77% of customers read online reviews before making a purchase. Yet in the Amazon app, users were confused about how to sort or filter reviews due to the lack of visual cues in the current UI, and were often unable to find these features.
The app has a button to check the shipping address right on the product page. However, some users didn’t even noticed the button and instead used their account settings or went halfway through the checkout process to check their shipping address.
When asked to find the book using a method other than the search bar, none of the users noticed the scan or voice buttons in the top navigation. Instead, they all proceeded to look in different departments and categories.
After identifying and prioritizing pain points, I created a provisional persona to help me create solutions for a specific type of user. This persona was roughly based on users that I talked to during the usability testing, and if I had more time, I would strengthen this persona with additional user interviews and research.
I sketched out some rough ideas on paper to ideate some possible solutions for the top pain points before diving into high-fidelity mockups — using low-fidelity wireframes allowed me to explore several different ideas instead of spending too much time on a single idea.
I created task flows of the top 3 pain points I identified to understand the relationship between the task and the various actions that users took to complete it. Then I created task flows of my proposed solutions to see how it would help improve the user experience — see the comparisons below:
Based on the lo-fi wireframes, I created a high-fidelity mockups in Sketch. The changes I made are annotated below:
Solution: Reducing the friction for sorting and filtering reviews would help users make quicker decisions about the item and increase the purchase conversion rate.
Solution: Providing users with an easy way to check and select their shipping address on the product page would allow them to skip a step in the checkout flow and speed up the checkout process.
Solution: Increasing discoverability for the voice or scan features could potentially increase online sales from users who want to research and compare products when shopping in stores.
Using the high-fidelity mockups, I used Marvel to create an interactive prototype for validation testing.
After creating the prototype, I tested my solutions using the same usability test questions from my initial tests. I also targeted people with a similar background to my initial users — all of them have used Amazon on their computers, but had little to no experience on the mobile app.
Results from validation testing:
Through some quick guerrilla usability testing, I found several issues with filtering and sorting reviews, checking shipping address, and discovering the scan and voice features in the app. Based on the results of the tests, I redesigned these features and validated my design decisions with follow-up usability tests on a high-fidelity, clickable prototype. Though the results from validation testing are qualitative, they show that even small changes to the app can increase discoverability and usability of important features.
This case study was a great opportunity for me to explore ways of improving one of my favorite apps, and through the process, I was able to flex my design skills by creating a solution from beginning to end.