Amazon 2023
Redesign internal tool to improve onboarding for new hire
During my Summer 2023 internship at Amazon, I led the design of the Amazon Advisor Inclusion Recommendation project, aiming to increase the integration of Amazon Advisors—a mentor role—into the onboarding process for new employees. My responsibilities included creating high-fidelity prototypes, conducting user testing, and collaborating with cross-functional teams.
In November 2023, my design was launched, impacting Amazon's L3+ workforce of 400,000 employees worldwide. This initiative improved the Amazon Advisor adoption rate from 42.3% to 64.5%, achieving a 1,383 basis point increase compared to the control group.
Time
Jun 2023 - July 2023
Team
TPM (Sharda)
PM (Anil)
Mentor (Rebekah)
Tools
Figma
UserTesting
Miro
Quip
My role
User flows
Thematic analysis
Functional prototypes
User testing
Design system
Background
In Q2 2022, Amazon created a new role for new hire onboarding called Amazon Advisor.
This role serves as an external counselor to help new hires who may feel intimidated asking questions to their managers.
However, after its launch, adoption of the role remained low, with an adoption rate below 50% in Q3 and Q4 of 2022.
In June 2023, I rejoined the GTMC team to conduct research on the root causes of this issue. My efforts helped increase the adoption rate by 23% after implementation.
Problem
How might we increase the adoption rate of Amazon Advisor so that more new hires can receive support when they need it?
The purpose of the project is to conduct research and investigate the underlying reasons for the low adoption rate. Our users are managers who have incoming new hires.
Solution
Redesign the page such that managers can understand Amazon Advisor, communicate their responsibilities, and find suitable candidate easily.
After carefully examining the tool for onboarding, I decided to redesign parts of the page such as managers have clearer understanding of what an Amazon Advisor is. In addition, they will be prompted to communicate with their selected Amazon Advisor.
01/
Provide role comparison chart
Provide a simple chart on the tool to explain the qualifications and purpose of an Amazon Advisor. This will allow managers to quickly understand the role and increase the likelihood of adding it to the onboarding plan."
02/
Recommended Amazon Advisor
candidate
Designed a reusable pattern to quickly suggest qualified employees to managers for selection, eliminating the need for manual search and reducing plan creation time.
03/
In-product guidance and banner
Prompt managers communicate with AA before their start date with a more personalized start date of the new hire.
Success metrics
Success metrics
The final deliverable will be an UX prototype, and the following are the metrics that we used to measure if the project have fully resolved the issue and fulfill its goal.
Dicover and define
Major findings and uncovered insights
I recruited 14 managers who had never used the tool before and conducted usability testing. The result shows glaring issue: 85% of the managers confused Amazon Advisor with another role, and 71% mentioned they did not know who to assign. From our findings, I have identified the following common themes:
Design approach
Since these problems are neither interrelated nor affecting one another, we decided—after discussing the results with PMs and Devs—to approach the redesign in three parts. Based on recommendations, two are low-hanging fruit, easy to address, while one requires high effort but offers high rewards.
Low hanging fruits
Problem 1 of 3 : lacking proper communication
For managers, there is a lack of awareness about how to communicate responsibilities to their selected Amazon Advisor. Even after selection, they often fail to consistently convey the role’s expectations, leading to some new hires not being assigned an Amazon Advisor.
Improvement suggestion
The improved design clearly states the jobs to be done by the manager. Additionally, I provide a status indication with a banner that highlights the next step after selection.
Problem 2 of 3 : confusion between roles
Initial research revealed that one of the most pressing problem was confusion between roles. To fix this design, I analyzed the current screen. Research results highlighted a major issue: 85% of managers confused the Amazon Advisor with another role.
Improvement suggestion
To improve the design, I added a comparison chart to clarify role differences and shortened text into glanceable bullet points. My PM initially opposed this due to cognitive overload concerns, but I defended the approach with positive UX research—though that’s a story for another time.
High efforts high rewards
Problem 3 of 3 : lacking connections
Among the three problems, the most pressing is helping managers establish connections. In the current design, managers must manually search for an Amazon Advisor. Without strong connections beyond their team, they often lack the time to find a suitable candidate for this role.
Improvement suggestion - people recommender
I redesigned this automated people recommendation pattern. This pattern will generate lists of employees for managers for different workflow. This solution effectively addresses managers' challenges in establishing connections.
People recommender iteration
To ensure the pattern meets our user needs, I have created multiple iterations and explorations on the position of different elements. Building on an existing UI pattern, I have landed on a final design along with a star feature - it separates the two selection methods with a horizontal layout and allows room for comparison.
The old people recommender pattern
Implementation
Build component into design system and provide design specs to engineers parners
After finalizing the design, I collaborated with the central design team and front-end engineers to integrate the pattern into the design system. Given Amazon's frugality culture, engineers are hesitant to create new components without proven success. To address this, I proposed leveraging existing components and adapting variants for the new design. I also facilitated meetings, documented the pattern, and assisted in stress testing before implementation.
Final design
Before and after
Finally, after alining and discussing with the cross-functional partners, we have finally landed on the designs. Here are the before and after screens for the design changes.
Problem 1 of 3: lacking proper communication
Problem 2 of 3: confusion between roles
Problem 3 of 3: Lacking connection
.jpg)
Impact
My impact
In November 2023, the redesign was launched officially. Future testing and surveys indicated a huge success to the project. The details are as the following:
Retrospective
My key learnings
01/
Understand Design system
During my internship, I led a Stencil team meeting to finalize and demo the people recommender pattern, emphasizing design system consistency. Embracing uncertainty and cross-platform user journeys fosters diverse pattern exploration.
02/
Pushback when needed
Throughout the project, while there might be instances where PMs rejected and push back design changes, it's crucial to consider the customer's perspective and respond with customer-centric approach.