Cigna (Evernorth) is slowly catching up with the modern world. Their database structures, legacy applications and outdated business models make it increasingly more difficult to productionalize machine learning models and API's for better business insights.
That's where the Analytics Platform comes in: it takes the slow, siloed waterfall process and moves to a cloud-based model where the Data & Analytics team can collaborate, upscale resources and build models in a fraction of the time.
As the sole product designer on the team, it was my responsibility to design the landing page of this product and everything that goes with it. The most of my job ended up being research and discovery, but ultimately the team aligned on several key features for the users.
Internal Marketing for New Users
One of the last but most important features built into the platform, I designed a section for new users to learn about who should join the platform and why, ways to learn more and stay up-to-date as well as seamlessly gain access.
With more and more users reaching out for information about the platform, this page became crucial in the larger strategy for enterprise adoption.
Most of the work I did for the platform was around the strategy for the hub. Users needed a centralized place for all their data science needs and easily access different tools.
Quick access to toolkits
A hub for users to launch any SaaS or cloud-ready data science applications
Non-existent navigation
Easily find cloud data sources and access them in any toolkit
Production requests
Handoff models to create a production model
User resources
Trainings, blog posts, announcements and other necessary collateral
The platform has 4 user types that typically work in conjunction on the same project:
Business Analysts
Business professionals who need a data model or feature for small-scale projects
**Data Scientist (Core Users)**
Develop complex Machine Learning models to satisfy a business need
Machine learning engineers
Scales and deploys ad hoc models from Data Scientists, maintains production model
Measurement and Reporting Specialists
Visualizes findings from model and tracks business impact
I interviewed multiple stakeholders - mostly our core data science users - to understand their process, how they could benefit from the platform and current pain points.
The development team had already started when I joined. They were working on integrating several SaaS applications our users were missing a way to access them.
Working with the product and engineering teams, we concluded the MVP of the product was a simple landing page that guides users (by role) to different cloud applications. From there we could scale to add other primary and secondary features.
Following the wireframes, I collected feedback / tested with our group of beta users on the platform. Several features could improve their workflow but there were several gaps that needed addressing, which brings us to the next section...
Beta users of Domino Data Lab (our first SaaS application) were interviewed regularly for feedback and recommendations. I designed several blueprints to bring transparency to what our users were doing and where there were gaps / painpoints.
What we found was an issue with users finding their data from on-prem sources and moving the resources to cloud storage. We were partnered with another initiative (DataHub) within Cigna to catalog and search metadata but was riddled with usability issues. I was moved to the DataHub team temporarily to test and improve the user experience.
Screenshot of heuristic analysis I presented to our stakeholders
One of the key issues I discovered with users was with the information architecture - users did not understand how the data was structured. The platform also combined a search and cataloging feature, something that made searching the data difficult.
The solution I designed and tested was a more formal data catalog, sorted by the source / application > tables > fields. The users also wanted to features related to the lineage of particular data sets given how muddled the Cigna data was.
After several months of working on different projects, I was consulted to re-strategize the landing page. The problem: users did not know what the platform was / did and had trouble gaining access.
My pitch was to rebrand the platform using a modern color palette, new logo (I did not design this), and dedicated sections for non-users.
The last feature I worked on was a new walkthrough to gain platform access. We landed on designing a quiz to guide the user through which permissions they would need, data access and automating the process for the users.
I then handed off my rebrand specs, strategy documents and research to another designer on my team as I had been reassigned to another project.
.png)
