modernizing the sub
onboarding experience
status: shipped
role: design lead and product manager
challenges
conversion
As of Nov 2023, conversion rate was 16% across regions. In California, 4,742 users entered the flow and only 215 activated.
support volume
52% of support time was spent on onboarding. From Aug to Apr, we had 8,114 tickets and spent 714 hours resolving them.
duration
On average, it would take California users 31 days to complete all onboarding tasks.
scalability
The legacy experience was built using Clojure and due to significant technical debt, it was extremely difficult to ship.
approach
The team was relying on bandaid fixes that didn’t tackle the root of the problem.
team
Throughout the course of the project, we had 5 different product managers which would caused fatigue and misalignment.
outcomes
conversion
Onboard rate increased by 7%. We grew sub activation by 16% while keeping sub ad spend flat, leading to a 10% decrease in customer acquisition cost.
support volume
Onboarding is still the highest ticket volume category, but we’ve reduced ticket volume by 12%.
duration
Now on average, California users take 33 days to complete all tasks. We made a tradeoff that reduced support ticket volume but extended overall time to onboard.
scalability
We modernized the backend and created a design system that is enabling us to modernize onboarding in other regions and make copy edits without engineering.
approach
We transformed the way we tackled onboarding to truly follow the design thinking process.
team
I proactively stepped up as product manager during an extended vacancy to thoroughly test, triage bugs, and release the experience.
discovery
The background check step had significant drop-off, and I wanted to understand what made this task so notoriously difficult. To better understand the user journey, I took the initiative to become a Swing sub — including fingerprinting and TB testing - gaining firsthand insight into how to improve the flow.
medium fidelity explorations
designing for scalability
Legacy was built without a design system which introduced an incredible number of variants for how information was presented to users. I stress tested my designs to consolidate the types of components we would need to support the onboarding requirements. I collaborated with engineering to ensure the design system aligned with the data model and prioritized atomic design principles.
Now, for the onboarding flow in other states, engineering is able to create “skeleton files” and I’m able to configure the front-end using the Step, Question, and Answer Options structure. When customer support shares that users are confused at a given step, I’m able to create merge requests with minor UI or copy updates to address the issue with minimal engineering support.
When the team wanted to create “Skill Builders” for substitute teachers, we were able to reuse the components to ship a brand new feature very quickly.
