Cityblock
Commons
Context
Cityblock pairs community outreach teams, local clinics, and virtual care staff with a software platform that helps surface medical and social care issues before they spiral out of control and drive up costs. Scheduling is at the core of the Cityblock ecosystem: all patient care happens through scheduled appointments, and the business depends on its software, Commons, to function seamlessly.
Problem
Cityblock’s existing scheduling feature was not meeting the distinct needs of different teams across the Cityblock care system. This led to hours of manual work, which posed a significant risk in care team efficiency and patient care. Below are some numbers to contextualize the amount of work that went into scheduling across teams.
📊
450 appointments scheduled per day
162 appointment types in one dropdown menu
5 FTEs worth of time spent on clean-up
5+ different systems used to manage scheduling
9 types of roles responsible for scheduling
Solution
Build a reimagined end-to-end solution for Cityblock’s scheduling mechanism based on in-depth research insights.
📎
My Role
Lead Product Designer & Researcher
🙌🏼
The Team
1 Designer / Researcher
4 Engineers
1 Strategist
1 Product Manager
1 Head of Product
⏳
Our Timeline
2.5 months
My Work
Design Solutions
I centered my design solutions around 4 major pain points I discovered while conducting user research.
1
Pain point: manual work
50+ draft appointments created daily causing 2+ hours of clean-up every day
The new page now displays a Drafts section so users can quickly identify appointments that need attention. It also features relevant tags, timestamps, and Notification History to help users identify any data that might be causing an appointment to sit in the Drafts queue.
Before closing the scheduling modal, users are also prompted to discard or save the drafted appointment, so they can be intentional about their work.
2
Pain point: inacurate scheduling
All dispatchers reported appointments often get scheduled incorrectly
In order to help people find the correct appointment to schedule, I designed a new scheduling modal with a more transparent way to find what you're looking for. The new category mapping was based on insights from a user card sorting study.
3
Poor UI signaling
Users were spending time scrolling through 162 appointment types to find what they were looking for, not knowing there was a way to type into the Search field.
4
Repetition
All users reported that scrolling felt like a waste of their time because they were typically scheduling the same 3 to 5 appointments every week.
An easy but impactful lift I proposed was to adjust the copy in the input field to help users understand they can type into the field to search. In order to help people schedule the same 3-5 appointments they typically schedule, I included a 'Recents' section at the top of the column.
Centering my design choices around user pain points helped me prioritize features within our Sprint cycles. But what informed all of this? Let's dive into the research I conducted to drive all of my decisions.
My Process
To improve scheduling holistically I proposed we run several iterative efforts centered around a set of research studies. This would help me identify a few low-lift solutions before diving into the long-term strategy.
🔍 ✏️
Phase I
-
15 card sorting exercises with roles across the healthcare organization
-
5 usability testing sessions to determine an MVP solution
-
Low-lift appointments modal design updates
🔍
Phase II
-
5 in-depth user interviews
-
Personas
-
Service blueprint
✏️
Phase III
-
Appointments page redesign
-
Appointment modal redesign
-
Strategy and next steps
Card Sorting and Usability Testing
Methodology
-
15 hybrid card sorting sessions in Maze in order to gather data around how people might categorize different appointments
-
5 usability testing sessions with users in different roles to observe how they schedule appointments
Result
-
During the card sort participants categorized 162 appointments into 9 categories, giving us a better sense of what our new data model might need to look like in the future
-
The usability testing sessions helped highlight how people interact with the current appointment scheduling UI, as well as a testing prototype for a potential UI redesign
 |  |  |  |  |
|---|
User Interviews
After addressing the two most important pain points, we shipped two low-lift, high-impact solutions (addressed later on). The next step was to focus on the user group with the most complex workflow so we could identify their distinct needs and design with edge cases in mind.
I observed and interviewed 5 appointment dispatchers during their daily workflow. We focused on diving deeper into the challenges they experience when dispatching scheduled appointments.
Personas
In order to contextualize my research findings, I started off by outlining outlined the key personas in the scheduling journey. These were based on the qualitative user interviews I conducted the week prior.
It’s basically cleanup all day long. Setting up the Dispatcher Google Sheet takes anywhere from one to two hours in the evening.
– Primary Evening Dispatcher
The Outcome
Measuring Success
As we began shipping the new UI updates, we started tracking some basic metrics to measure the outcomes of our new feature implementation.
📆
Number of inaccurately scheduled appointments
📄
Number of draft appointments created daily
🕐
Time spent on cleaning up draft appointments
The plan was for the new selection modal to be the first step towards rebuilding scheduling from the ground up within our product suite. However, budget cuts forced the team to pivot and postpone this route until 2024. Overall, this was one of my favorite projects because I was given the opportunity to lead from an early problem elaboration stage all the way through shipping a set of new product features. In addition, all of my design decisions were rooted in data, which made the team feel confident in our work.