Expanded target user base by 450% by developing a new feature within the consumer-facing mobile app to enable deaf users to contact 988.
Company
AccesSOS
Timeline
Aug 2022 -
Dec 2022
Team
1 UX Researcher
1 Project Lead (ME)
2 UX Designer
1 CEO
Tools
Figma, Card Sorting, Interviews & Storytelling
My Role
✍️ End-to-end iterative process: Conducted research activities to iterate on our design, including:
-
Interviewed 7 emergency call dispatchers
-
Recruited over 15 deaf users to A/B test the user flow
-
Modified and added new elements to our design system
🤝 Cross-function collaboration: designed the logic tree to help the engineering team direct the requests, worked with the legal team to patent our designs
👩💼Stakeholder engagement: Pitched to the CEO and the stakeholders on future iterations to secure fundings
This project highlights my ability to craft accessible design solutions for specific user groups, lead design processes, and manage stakeholders.
Problem
37 Million
70%
Americans can't hear or speak out loud
of 911 call centers can't receive texts
AccesSOS is a nonprofit tech startup on a mission to make emergency help more accessible. Our product aims to bridge this gap: a free app that connects these people with emergency services when they need it.
Scope
Emergency Help Systems
Call for suicide prevention and crisis support
Call for emergencies
Call for information and community resources
Product Requirements:
Add 988 Mental Crisis Helpline into our product
Expand Users by 450%
Estimated 37M impacted
Target Users:
Deaf, domestic violence survivor, language barriers...
Research
Where AccesSOS comes in:
❓Problems:
-
UX: single flow (911)
-
UX: Users are not aware of 988
-
UX: Users don't know what help they need
-
UI: integrate
🎯Design Goals:
-
Turn single flow into multiple flow
-
Educate users about 988
-
Connect users to the helpline they need
-
Design corresponding visuals for 988
Before diving into Figma, I decided to carry out research activities with our real users and SMEs (988 dispatchers and supervisors) to lean on their domain knowledge and better understand how our product fits into their flow and how we could better help our users. Below is the research plan I outlined:
💡Interview
7 Participants
988 dispatchers, deaf community event organizer
1. Reason:
Get rich qualitative insights on the use case on both ends (users and dispatchers), and observe their flows
2. Recruitment:
Leaning into my professional and academic networks, I recruited participants through these channels:
-
Attend deaf community gatherings, 988 Crisis Jam
-
Send the screener survey (see above)
-
Snowball from one contact within our company
3. Research Questions:
-
Figure out how 988 responders take emergency/non-emergency calls;
-
Identify the pain points 988 responders face when helping others;
-
Get more contacts that we can reach out to for more user testing/ interviews.
4. Format
Zoom Call, Interpreter, Text
Because we are talking to deaf users!
5. Findings
-
Taking Call Process:
-
Risk Assessment (Thoughts, Plans, Means, Timeframe): if emergency -> first responder
-
Build Rapport
-
Provide Resource
💡Card Sorting
9 Participants
988 dispatchers, Crisis Helpline Supervisor, deaf community event organizer
In order to lean on our SME's domain knowledge to help users connect with the help they need, I designed this card sorting activity for them to categorize which scenario requires what help. We analyzed the results quantitatively as a team and produced a logic tree (see later section).
Synthesize
✍️Affinity Mapping
In order to synthesize our findings to find the common pain points and tell user stories, together with my team, we used affinity diagrams and whiteboarding to discover patterns from our interviews.
1. The amount of information needed by mental crisis counselors is very few: they avoid asking for personal information to establish trust and build rapport;
2. Most texters who contacted 988 do not want police to show up at their doorstep, so calling 911 on their behalf without a heads-up may be triggering and exacerbate the conversation;
3. While 988 can call emergency services for you, it will be a lot faster if you need help urgently to call 911. If you need emotional support at the same time, call 988 as well
✍️Logic Tree Renovation
After integrating our card-sorting results, we iteratively renovated the original logic tree, which is designed to help the engineering team direct requests in the back end. We introduced more situation elements and which line to contact, simplifying the workflow for our engineering team.
.png)
.png)
Major iteration versions of how we created this logic tree
Prototype
🔍Current User Flow
Let's take a look at the current, single user flow:
❓Problems:
-
UX: single flow (911)
-
UX: Users are not aware of 988
-
UX: Users don't know what help they need
-
UI: integrate
🎯Design Goals:
-
Turn single flow into multiple flow
-
Educate users about 988
-
Connect users to the helpline they need
-
Design corresponding visuals for 988
Recall the problems and design goals, based on research findings, I identified several design opportunites.
💡Design Solution 1: Final screen optimization
Current
Sketch
.png)
Proposed changes:
-
Introducing ETA
-
Give real-time feedback
-
-
Reduce Text
-
Reduce mental load
-
-
(explore) Vertical Layout
-
Icon in sync with text
-
Low-fi
✍️ Iteration I
Before
❌ Too much text
❌ Not enough feedback
❌ Icon is not synced with text copy
After
✅ Reduce and edit text
✅ Add ETA to provide more feedback
🔎 Explore vertical layout
💡Testing and feedback (within company)
I quickly ran some casual user testing sessions with the engineers and PMs at our company to facilitate the iteration process and keep them in the loop. Here is the feedback I got:
-
Text copy edit: Is 988 also ’calling’ the dispatcher?”
-
Not enough distinction between 911 & 988
-
UX Challenge: what if the two lines are in different stage
-
ETA not feasible
✍️ Iteration II
Before
❌ Confusing text
❌ Not enough distinction
After
✅ Separate two lines
Color Distinction
✅ Removed ETA - add to the next release
❌ Vertical Layout
💡Design Solution 2: Consent pop-up
In order to address some pain points discovered during interviews, I explored adding a consent layer in the UX to give user security and option
2. Most texters who contacted 988 do not want police to show up at their doorstep, so calling 911 on their behalf without a heads-up may be triggering and exacerbate the conversation;
Current
Sketch
.png)
Low-fi
✍️ Iteration I
Before
❌ Directly goes to calling 911
💡Usibility Testing
6 Participants
Paid testers, students, company employees
1. Reason:
Test the prototype in a setting that resembles real-life use scenarios, because it's impossible and unethical to test our product in real use cases. I designed the task to be time-sensitive, participants have to finish certain tasks within 30-60 seconds window, to mimic the urgency in emergency situations.
Usability testing feedback document
✍️ Iteration II
Before
❌ What is 988?
After
✅ Added descriptions and icons to represent communication method
💻 Final User Flow:
💻 Deliverables:
After
✅ Added a layer to confirm which line users want to call
Impact & Learnings
✨ Impact
-
988 Integration: shipped, 58% adoption rate
-
Better backend direction: proven to improve efficiency for engineering team
-
Helped CEO to secure funding from investors
✨ Learnings
I grew a lot as a product designer from this project. I learned to deal with sensitive user needs with utmost attention to detail and conduct situational user testing without putting users in the use scenarios (in our case, we cannot actually perform user tests on people who need to call 911 for help). My skills in the following areas were greatly improved from this project:
-
Work Division within teams
-
Project Management
-
UI Design
-
UX research
-
Stakeholder Engagement and Presentation
I reached out to a lot of non-profit organizations and 911/988 respondents, and realized that problems like this lie everywhere in our society, but not a lot of capital and resources were deployed into solving them. Hence I am inspired to develop products with every user group in mind in my future works and always think of and pitch solutions to these problems to relevant parties.