CASE STUDY
API Mapping for a Healthcare Staffing Company
Problem Statement:
A healthcare staffing company needs to have the most up to date and relevant information about available positions from thousands of employers.
The Results
A foundation for the UI and Research team to test with through user interviews and A/B testing
A modernized micro services architecture that saves resources
Defined measurements of success
Team Structure:
Larger group of over 100
Immediate team comprised of 3 groups ( PO’s, Dev, XD) - sole designer on my team and provided design leadership and guidance with designers on other teams
Challenges
System was dated and had a significant amount of tech debt
Current system was home grown and had limitations to growth and scaling
The information available from each employer was not documented
Each employer has different fields for job listings and there is not an industry standard
Key Goals
Maximize cost savings while maintaining functionality - find the redundancies and design for efficiency
Create a repeatable framework that supports the team in the future and is explainable to new team members
Collaborating with the UI and research team determine what the most relevant fields are to job seekers
The Process
Design Workshops (on repeat)
I began the work running workshops with my team of Developers and PO. Candidly the group was skeptical that a designer could support API design but by the end of the first 2 hour workshop they saw the value in putting thoughts to paper. When I’m running workshops my high-level goal is always to sort out ‘what do we know vs. what do we understand’. In this scenario I had an incredibly talented and knowledgeable group of team members who knew a lot about the system but didn’t understand why it worked the way it did.
The first workshop began with defining known pain points, for both internal and external users. Everyone contributed using stickies in an online white boarding tool. From there we worked as a team to group the stickies into categories to see where the most commonalities were and discuss as a group if there were sub-categories to anything. We continued this exercise until we had defined root causes of problems and assigned the business value to solving them. This kept the team aligned on what was worth solving vs. what would provide little return on investment.
API Mapping & Eco-System Maps
By the end of the initial workshop series the team was aligned on the work to be done. With the collective knowledge of the team in one location I took a pass at creating an eco-system map of our product and how it interacted with the business at large. This is incredibly important to be sure we understand both upstream and down stream dependencies of any work to be done. After validating that map with the team I worked with our lead developers to define the granularity of the monolith system we were working with and how to break into microservices that supported our end goal. I ran consistent workshops with them to define the work to be done and get that work onto a backlog.
Information Architecture
With the problem statement defined and micro service foundational work in motion I needed to turn my efforts to the information coming in and determine how to organize it. Each employer posting jobs had wildly different structures and we needed to build a system that not only captured the information but organized it into a consistent and repeatable framework for the front end team to test. Working with developers I stepped out of my comfort zone to learn how to use Postman and reverse engineered all of the data coming in and then reconciled the differences within a spreadsheet. From here I worked with our front-end and research team to support how this information was surfaced and how to test results with users. Understanding the nuances between different professions and the information desired was key to being competitive in this industry.
Ongoing Desk Research
A huge contributor of the success to this project was research. This included utilizing paid tools such as HotJar for website feedback and Dovetail for user interviews but also searching reddit and google for key pieces of information. There is never one perfect location to find all of the information you need. Compiling all of the data together painted a picture that validated our assumptions and gave us confidence in our implementations.
Tools used:
Figma
Miro
Excel
Postman
Dovetail
Hotjar