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.
Jun 2023 - July 2023
TPM (Sharda)
PM (Anil)
Mentor (Rebekah)
Figma
UserTesting
Miro
Quip
User flows
Thematic analysis
Functional prototypes
User testing
Design system
In September 2022, only 42.3% of new hires had an Amazon Advisor. Multiple internal and external studies have identified ‘getting connected’ as a key input to successful onboarding. Upon research, these are the three main reasons:
When assigning onboarding plans, managers confuse Amazon Advisor with other roles.
Even after assigning, managers don't know what to do, when to connect, and how to connect.
Most importantly, managers lack connections that allow them to know who to assign as an AA.
The design reduced the workload for over 300k managers, enhancing the onboarding experience for more than 400k Amazonians globally.
The design increased the Amazon Advisor adoption rate by 22% and improved the onboarding satisfaction rate from 60% to 85%.
By launching the designs into the Embark portal, the new hires' satisfaction rate increased by 20% in the next quarter.
I began this project by familiarizing myself with the onboarding process. I held meetings with another designer on the OnEx team and the PM to define business and user objectives.
The first goal is to improve the satisfaction rate for new hires building their network for success from 78.8% to 83%. This will be measured through survey after the product has been lauched.
The next goal was to increase the adoption rate. As one of the business goals, we need to increase the adoption rate by 15%. This is the benchmark for our success metric.
Then, I explored Embark independently to conduct a heuristic analysis. These steps helped me map out the existing user flow and frame my research tasks.
To understand how effective the current AA assignment is, I looked at the 2022 research report to see if there were any issues uncovered. Here are some insights.
of historical AA assignments are within an orgdistance of 4 of their new-hire.
of new hires do not had an Amazon Advisor
of new hires had at least 3 Key Amazonians (another onboarding role) to Meet
To understand why managers are not assigning AAs , I recruited 14 managers who had never used Embark before and conducted usability testing through UserTesting. Given the issue, I would ask the participants to speak while they were given certain tasks to complete. Here are some questions I had asked.
Then, I categorized their quotes into a specific theme and ranked the issue's severity based on the frequency of mentions. The result is glaring. 85% of the managers confused AA with another role, and 71% mentioned they did not know who to assign.
These were the most urgent issues that needed to be addressed based on the usability test and follow-up interviews.
Managers often become confused about the differences between Amazon Advisor and the onboarding buddy.
When managers don't clearly understand Amazon Advisors's responsibility, Amazon Advisors will not be fulfilling his/her duty to the fullest.
When making an onboarding plan, managers may skip an Amazon Advisor (AA) because they aren't familiar with suitable individuals outside their team.
After boiling down the main problems, I prioritized brainstorming quantity over quality to ensure I could deliver our designs in 1 month. I then select solutions on the feature-prioritization chart by collaborating with the product and engineering team.
After boiling down the main problems, I prioritized brainstorming quantity over quality to ensure I could deliver our designs in 1 month. I then selected solutions on the feature-prioritization chart by collaborating with the product and engineering team.
Lack of communication with Amazon Advisor
Providing clear in-product guidance
Providing clearer in-product guidance that reminds managers to communicate with the selected AA. Additional banner to reiterate to communicate with AA.
Insert additional reminder banner
In addition to in-product guidance, adding a banner that prompts managers to reach out to their assigned AA.
Confusions between stakeholder roles
Create a comparison chart for the roles
I will be providing a comparison chart, as it clearly and concisely displays the differences between AA and Onboarding Buddy.
Limited connections lead to missing AA
Implement people suggesting pattern
Surface recommendations of qualified people to help managers select AAs at ease. We hypothesize managers will assign AAs more often and lead new hires to establish stronger connections.
Many interviewed managers expressed concern about what to do after completing the plan. The current page obscures key information and does not provide clear next steps for users.
Visual design treatment was limited to only text and did not emphasize key features and information.
Lack of visual treatment to support longer entries and text wrapping.
I have removed redundant information from the page. Then, I simplified the text into concise sentences. Additionally, I changed the banner text to remind users to reach out after completing the plan.
I have replaced lengthy sentences with simple, easily glanceable text. Additionally, I have updated the recommendation to assign an Amazon Advisor to ensure users complete this task.
I have positioned the banner in the task section. Utilizing the Von Restorff Effect, this prominent placement is designed to enhance recall, ensuring that managers can promptly contact their new hires.
Initial research told us the most pressing problem was the confusions between roles. To fixed this design, I looked at the current screen.
The banner does not provide any additional information for managers, failing to serve its intended "notification" functionality.
Welcome message to the new hire is irrelevant to choosing your onboarding connections.
The visual treatment did not emphasize the key points between onboarding roles, making it difficult to grasp them at a glance.
85% of the interviewed managers confused Amazon Advisor with the other role. Hence, my design direction focused on enhancing legibility and glaceability through the addition of structured elements like tables and bullet points.
Organize important information into rows and tables to facilitate easy comprehension, and position roles side-by-side for clear comparison.
I have relocated the welcome message to a different section, thereby reducing the amount of information on the page and lowering the user's cognitive load.
During the design process, my PMs opposed the comparison chart idea, citing concerns about excessive cognitive load for users. As the design lead, I persuaded them to retain the chart by presenting usability testing evidence demonstrating that managers need to quickly understand the roles. Additionally, following UX best practices, we should not obscure important text that aids in making key decisions.
The final issue is the limited network connections, which leads to a shortage of Amazon Advisors. According to the interview, many managers lack the external connections necessary for assigning someone to become an Amazon Advisor.
The current screen asks managers to manually search and assign an Amazon Advisor. The button is functional, yet I need a better design to reduce manager's workload.
I designed an automated people recommendation pattern and held a UX review session. While feedback on the comparison chart and other designs was positive, the search box in the recommendation pattern remains confusing.
The position of the search box within the selection pattern creates confusion regarding its function. It is unclear whether we are searching for 'all' employees or 'suggested' employees.
To fix this issue, I have created multiple iterations and explorations on the position of different elements. I want to make sure it looks like two separate functions while does not impact "load more".
After several iterations, we finalized the second design. It separates the two tasks with a distinct layout and allows room for comparison.
Provide a simple chart on Embark 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."
Designed a reusable pattern on Embark to quickly suggest qualified employees to managers for selection, eliminating the need for manual search and reducing plan creation time.
Prompt managers communicate with AA before their start date with a more personalized start date of the new hire.
When the design was finished, I recruited another 14 managers to test out the new prototype. I wanted to make sure the improved onboarding creation experience worked and made sense. I am so happy with the result. Here are the results.
of the managers claimed the people's recommendation is "useful" to help them find Amazon Advisor.
of the managers have answer the question "what to doe after plan creation process" correctly.
of the new managers rate the comparison table for roles easy to understand.
In November 2023, the redesign was launched officially on Embark. However, future testing and surveys are required to estimate the growth of employee satisfaction. Only in that case, the project is claimed "real" success.
During my internship, I facilitated a meeting with the Stencil team to finalize and demo the people recommender pattern. This highlighted the significance of a design system in ensuring consistent cross-platform design. Embracing uncertainty in patterns and considering user journeys across platforms enables the exploration of diverse variations of the same pattern.
Throughout the project, I collaborated with PMs and the tech team to outline the business scope and assess technical feasibility. While there might be instances where PMs requested design implementation, it's crucial to consider the customer's perspective and respond with customer-centric approach.