Amazon UX Internship Project

What's the problem?

At Amazon, there is an opportunity to simplify the flow of information between customers and drivers. Particularly, we believe customers can benefit from updating their instructions. Thus, working within a team focused on the post-purchase experience, I was given the task of redesigning delivery instructions.

What are instructions anyway?

Delivery instructions are an opportunity for customers to leave preferences and notes for delivery, such as telling delivery drivers where their package should be delivered (“To my front door”), to how to find and access their address (“This is my security code, I am the first door on the 3rd floor”). Amazon provides many questions on this page that are optional for customers, as these instructions will apply to future orders.

How customers access instructions

In the current experience, customers are typically introduced to delivery instructions through the Checkout page and through the Progress Tracker (PT).

• Checkout: Before placing their order, customers can press a secondary link within the Address card labeled “Add delivery instructions”. This takes them to the main page of instructions.
• Progress Tracker: After the customer makes an order they can press the “Update delivery instructions” button in the main card.
• Main page: Here customers are asked important questions from where specifically their packages should be left (e.g. front door, mailroom), to how to access the property (security code or call box number).

The current CX could use some updating

Having seen the current experience in Checkout and PT, I believe customers can be better encouraged to provide instructions or update delivery information they previously entered.

Little context and visibility is given to the user through these ingress points about what delivery instructions entail and why they might be important.

Moreover, the current main page of instructions consists of expanders, which require a lot of taps from the user and take up more padding than is necessary on the screen.

• The expander UI's unusual expand and collapse behavior—in which opening one question auto-closes another—also creates cognitive load, and can pose an accessibility challenge.

Analyzing research

To better understand the issues around delivery instructions, I dove into current usability studies from my team and engaged in conversations with my partners. With that, I was able to identify different themes, resulting in the creation of three validated customer problems.

Learning from competitors

Keeping in mind delivery instructions exist in other apps, I conducted a competitor benchmarking study to see how other companies are solving this problem.

Three apps I analyzed were Shipt (a delivery service owned by Target), Walmart, and UberEats.

• Though both Walmart and Shipt have simple UI for leaving delivery instructions, they had some unique components that contributed to the customer experience.
• As some customers forget to update their instructions, I appreciated that Shipt gives a reminder that notes they leave behind apply to future orders.

• Likewise, Walmart addresses customer expectations towards drivers by clarifying they will do their best to follow instructions, but they may not always be achieved by drivers. I loved how this component helps drivers, who may not be the target audience but is involved in this area nonetheless.

• As for UberEats, I was inspired by their simple yet well-thought UI.
• Helpful language is interspersed, letting the customer provide delivery preferences with the driver in mind.
• A map experience lets customers drag a pin to where they prefer their dropoff point.

A ride-along impacted the course of my project

In the 2nd week of my internship, I conducted my own research through a ride-along: a day following a delivery driver and observing their process.

Getting to meet someone directly related to instructions helped open my mind to a different side of the problem and impacted my proposed concepts.

• I learned how many environments a delivery person navigates in one day. As drivers are encouraged to deliver packages quickly, mistakes can happen, which are not the fault of the driver.
• Meanwhile, they aren't obligated to follow instructions if they pose a threat to their safety—for example, a dog preventing them from delivering to the front door.

Wireframing my designs

As soon as I had collected research and validated the customer problems, I ideated on possible concepts. Throughout this process, engaging in Crazy Eights, an ideation practice where you create 8 designs in 8 minutes, helped greatly. Moreover, starting with paper helped the idea process feel more organic.

Presenting to my team

In the middle of the internship, I proposed the different ideas I developed to my team. I received a lot of interest in my work, as well as feedback. What I quickly learned about sharing my designs was deciding what advice I wanted to act on, and what thoughts were less feasible to act on given the timeline.

Proposed CX

Seeking to simplify the experience between the customer and the delivery driver, I proposed 5 mobile concepts to my team. Three of these concepts involved rethinking where customers can access instructions.

Concept 1: Button redesign in Checkout

In the current experience, customers can access instructions by pressing “Add delivery instructions” in the address card at Checkout.

• However, the button's design makes it unclear what delivery instructions are or whether customers have saved any.
• In this proposed concept, I used helpful language to add context and increase engagement with delivery instructions.

"Add instructions" screen:

In the first use case, where a customer hasn't inputted instructions, the system displays a description below the button to explain what delivery instructions are and how they help drivers.

"Edit instructions" screen:

In the second use case, when instructions are already saved, customers see a preview of their preferences with a “Show more” option under the button. Changing the button's text to “Edit delivery instructions” in this situtation improves usability with minimal effort.

Concept 2: Progress Tracker Nudges

The PT page, where customers can check the status of their package, is a prime area to increase visibility of instructions.

• Currently, a generic“Update delivery instructions” button allows the user to access instructions. However, it doesn't provide much motivation for users to review or update their instructions.

My solution is to inject nudges throughout that encourage customers in different situations to provide meaningful delivery instructions. This includes:

• A bottom sheet appears when users click a button that shows part of their current instructions, asking, "Is this correct?" This sheet allows customers to quickly edit specific details without leaving the current page.
• Updating the original button to ask for a specific detail (e.g., “Add property type”) and emphasizing how this information helps drivers.
• A new feature to “Add one-time delivery instructions”, summoning a bottom sheet that creates temporary instructions for a specific order—crucial for customers who need to apply instructions to a particular order.

Concept 3: Pin on Map experience

While Amazon offers a map experience that allows customers to update their address with a single pin, it is currently only accessible after leaving negative feedback.

Therefore, I propose introducing this feature more prominently to help customers clearly communicate where to leave their packages.

In the new experience, customers can add up to 2 preferred delivery locations (e.g., front door and side door) and provide pins for each location.

The second location is optional and presented as a “backup” to meet diverse customer needs. For instance, customers can add a neighbor's address as an alternative delivery spot.

Closing out with UserTesting

After I finished my final presentation successfully, I conducted usability testing to determine whether customers are likely to update delivery instructions with 2 of my proposed designs in Checkout.

I met with one of my partners, a senior user researcher to create a research plan, organize questions on UserTesting, and publish the test.

The test had 10 participants and involved assessing whether participants had used delivery instructions, describing each screen shown to them, and indicating which option they saw as most useful and why.

As the results came in, I analyzed customers' feedback through video recordings and the help of a new AI feature summarizing key moments.

I then wrote easy-to-read themes of the raw data, recommended next steps for the project, and included this in a comprehensive final project document.

• While I was compiling this document, and during the internship, my manager and mentor were crucial in providing examples of what typical documents looked like in the team.
• As a result, my efforts received a lot of interest from the team, providing valuable insights for how to improve the ever-complex delivery experience.