Crash Detection UI Layer
OVERVIEW
Arity had been offering customers a crash detection algorithm for several years, and over time the product team identified an opportunity to also provide a crash detection UX.
The team wanted to build a UI layer that was researched and thoughtfully designed, so that customers could quickly and easily integrate it into their app. The target customer of this feature is app publishers that want to provide a crash detection feature for their users, but may not have the time or means to design and build that feature from the ground up.
As the lead designer on the team, I worked to solve both business and user needs with the crash detection experience.
QUICK FACTS
Role: Experience Designer, Researcher
Duration: 4-5 months
Collaborators: product owner, data scientists, developers, design researcher, visual designer
THE DESIGN PROCESS
-
My first step was to scope the work and understand what needs - business, user, or customer - should be solved with the end solution. I collaborated with stakeholders and my team to ask questions and identify key needs.
-
After having a solid grasp of the scope of the work, I started sketching initial ideas for the UI. I included other designers that worked on similar products to bring in additional ideas and perspectives.
-
With low fidelity wireframes of initial UI ideas, I conducted interviews with 10 participants who had been in car accidents to understand how users may want or expect a crash detection experience to work.
-
From those interviews, I identified key recommendations to iterate on the user flows and the UI and developed various options for the UI. I worked with visual designers and other UX designers at Arity to identify the top design to move forward with.
-
After a few weeks, I conducted another round of research - a survey to collect quantitative data to validate our overall concepts for our crash detection experience. The findings from this research were used to prioritize the team's work.
-
I also conducted another round of interviews to broaden the teams understanding of user preferences surrounding the crash detection experience. The findings from this research fed into product strategy and UI iterations.
-
Lastly, I finalized the designs for both Android and iOS platforms, including recommendations from all the rounds of research, and created design documentation to hand off to developers so they could start building the UI layer
Below I've included references and documentation from various steps in the design process.
SCOPING THE WORK
Before diving into the research and design of the crash detection experience, I worked to understand and define the scope of the work. Working with product owners, developers, data scientists, and UX researchers, I identified key business, user, and customer needs, as well as important considerations.
BUSINESS NEEDS
-
A way to deliver a pre-built crash detection UI to sell to customers along with our algorithm services.
-
Collect data from users after we've detected a crash to inform our algorithm and improve its’ accuracy.
-
Account for customers that want to integrate with a call center and those that do not. And account for whether or not customers want to integrate with any third party providers for data storage.
USER AND CUSTOMER NEEDS
-
A straightforward and easy-to-use UI that presents helps options right after a crash is detected.
-
A minimalist and simple design that can have brand colors easily applied to it, to match the branding of whichever app it is integrated in.
-
The crash detection feature and experience should be as accurate as possible, and not trigger a large number of false positives.
DESIGN DECISION 1
-
Create 7 unique user flow options
There were two main needs I aimed to satisfy with this decision
1
2
Customers that we partner with may want to integrate with a call center and/or work with a third party data warehouse to store user information (like emergency contact info) or they may not want any integrations - so I designed experiences to account for every scenario.
Users need accuracy from a crash detection feature, so I worked with data scientists to understand how the algorithm works and how it returns data after detecting a crash. One piece of data that it returns is the ‘confidence level’. This confidence level means how sure we are that a crash occurred and also has a direct correlation to the severity of a crash (i.e. the higher the confidence level, the more severe a crash). Using the 3 existing confidence levels, I designed different experiences for each level, based on what a user may want or expect for a severe crash or a less severe crash. These user expectations were informed by research - both previous research done by others on my team, and research I conducted myself.
Click image to enlarge.
DESIGN DECISION 2
-
Design the UI to be very simple and generic - without excessive branding, or anything else that might distract the user
By doing this, I aimed to satisfy the customer need to easily apply their branding to design components (fonts, colors, etc.).
I also aimed to fulfill user needs for a clean and easy-to-use interface in a moment where they are likely stressed and overwhelmed.
Click image to enlarge.
DESIGN DECISION 3
-
Develop a post crash survey that collects data from a user after we detect a crash on their phone
This choice satisfies the business need to collect accurate data from users and feed that back into the algorithm to improve its predictions. I collaborated with my team's data scientists to understand what information is most valuable to collect from users, and included that in the survey.
I also needed to be mindful to show users the value in sharing this information, since some may not be comfortable answering questions about a crash event.
Use the arrows to scroll through the post-crash survey flow
KEY TAKEAWAYS
-
Once the design work was completed, I transitioned to a different product team and acted as a point person for questions and reviews as the team built the crash detection UI.
-
The business segment found value in the competitive analysis I created and planned to continue research on how they can iterate the crash detection UI layer to continuously improve its value for users and customers.
Some information may be omitted or redacted for privacy concerns.
BROWSE MORE OF MY WORK
