COURSERA

In 2016, Coursera was working with one school, the University of Illinois, to create an online graduate degree program. This was the pilot for an audacious bet: that Coursera could create the premier online degree platform, one that would allow universities to deliver a degree as valued as their on-campus offerings, but in an online setting.

As one of Illinois’ administrators said at the time, we were building the plane while it was taking off! Everybody knows how it goes when we’re indexing for speed—we end up with technical and UX debt.

Balsamiq wireframes allowed for quick sketches of a future system, including curriculum management.

We decided it would make sense to take time to narrow our uncertainty about the future. We could imagine a UX that would use existing models and conceits while expanding capabilities, such that a university could spin up a degree with operational ease and create programs that offered the rigor of on-campus degrees with the flexibility an online platform could allow. Even at the wireframe stage, in 2016 we had a vision of what the future might look like.

Drop-down madness! On page after page, users had to make sure they had the right combination of content, people, and time.

Still, we knew the tech and UX debt would have to be paid down. The system was originally designed for on-demand courses that learners could work through in their own time, and then updated to provide the structure of automated deadlines. Trying to bolt on limited enrollment, manual deadline-setting, and assignments graded by staff rather than a machine was difficult at best. By 2017, the UX had suffered—page after page in the app sprouted drop-downs to ensure the user was working with the right combination of content, people, and time. Hopefully. Something had to be done. We had to get the basics right before we could achieve the vision.

Post-its with purpose: We conducted multiple workshops to understand how courses are built.

I worked with our PM and put together workshops with internal stakeholders—partnership managers who worked with university admins, in-house instructional designers who worked with course creators, and our partner support team—and conducted user interviews with faculty and staff at partners working on both open courses and degree programs to understand their mental model of how different types of courses were created and run.

This helped us understand the ways the two models differed. Even if they used the same basic elements—content, people, and time—the balance of these at play in any given use case was different, and the system design had to reflect this.

Our design assumed that the difference between large public courses and small private ones was made explicit.

The PM I was working with and I decided I’d dedicate some time to prototyping a higher-fidelity take on what it might look like if we re-architected the system and the UX to distinguish between large, open courses with looping public sessions and private, credit-bearing courses that run when universities need them. We used the opportunity presented by the 2017 Coursera Partners Conference to see if we were on the right track. We got lots of helpful feedback—we were starting to get the mixed model right, which meant we were closer to having our degrees platform.

Team Projects research gave us the chance to show how eliminating system debt could lead to clearer experiences.

Even while we began to address this problem, we added new features to unlock more capabilities for our growing stable of degree programs. One important addition was the late 2017 Team Projects initiative, a system to allow instructors to create flexible teams of learners to whom they could assign projects and discussion topics.

The project offered an opportunity as we conducted our user research. This particular feature was especially difficult to design for, given the existing system’s constraints, and in particular certain parts of the UI’s inability to “know” about what content was associated with a given session and therefore the people in it. I decided to test a version of this system as if we had solved that underlying piece of tech/UX debt, to see if fixing it would provide a simpler solution for our users. We learned that it did, and set the stage for fundamental change.

After iteration and thorough research, we delivered a solution that paid off our debt and opened new doors.

By the middle of 2018 my allies in product, design, and engineering and I were able to get this change onto our roadmap. We took the time to develop multiple medium-fidelity solutions and worked with research to validate our choices. We ended up with a solution that users across our partners loved, and would also support future initiatives driven by partner schools and by Coursera. With the finish work completed by one of my teammates, Vincent Tran, the project ultimately known as Course Types finally shipped in 2019.

Three years in the making: shipping work for a real degrees platform.

While this was in flight, we used it as a springboard to begin working on a true degree management system, a project known as Degree Authoring. With the architecture of the course-level content, people, and time management now able to support it, we could envision how we might allow degree administrators to develop curricula made up of courses and groups of courses, using the same set of models they would understand from developing a course curriculum.

The work I designed here in 2019 was directly inspired by that Balsamiq wireframe I made all the way back in 2016. Spending the intervening 3 years socializing it with product and engineering folks, both leaders and rank-and-file was, I happily learned, skating to where the puck was going.

 

← Back to Work