Curricle
A web platform that leverages interactive data visualization to bring innovation to a student’s course discovery, selection, and planning experience.
MISSION - Lead the Curricle team to identify features for design and implementation, plan and execute a development roadmap, and communicate the platform to key stakeholders and target users through its launches in the 2019-2020 academic year
TIMEFRAME - June 2019 to June 2020
TEAM - metaLAB (at) Harvard and the Berkman Klein Center for Internet and Society, with support from the Office of the Division of Arts and Humanities and the Office of the Provost
ROLE - Product manager and designer
Curricle is a platform for course discovery, planning, and selection across the Harvard curriculum. Unlike traditional course management systems, Curricle reveals the curriculum not as a mere listing of courses but as a dynamic network of connections that, like a living organism, evolves over time. Through leveraging interactive data visualizations, it encourages a novel way of approaching the conventional, keyword-driven course selection experience.
CONTEXT
I joined the Curricle team as its Product Manager in May of 2019. By that point in time, the team has developed and rolled out an early beta version. It was positively received by a select group of testers, and my job was to identify key features to design and implement, plan a roadmap through the summer, and coordinate an even more successful Fall launch.
As part of onboarding, I pored over past research and team communication to get up to speed. One of the first things I did was planning and leading a few all-hands meetings and alignment workshops for everyone on the team to start on common ground and for myself to become better acquainted with the team dynamics. The meetings offered everyone—designer, visualization artist, technologist, developers, and faculty directors—a space to share their vision and goals, and we were able to collaborate on a summer-going-into-fall product roadmap with many elements of agile software development: weekly sprints with planning and review, close and frequent engagement with end-users (Harvard undergraduates), and an MVP to validate the Curricle approach.
Our team aligned on a vision for Curricle: to positively transform the liberal arts course discovery, selection, and planning experience. From there, I identified two key tasks: 1. To validate the value of Curricle’s approach of encouraging non-conventional course discovery, and 2. To deliver a good product by the deadline of the start of Fall semester.
Framework used by team to collaboratively align values against broader short- and long-term goals.
VALIDATING APPROACH
From the work gone into making the first beta, I noticed that much of our former research was centered around usability tests rather than user research with greater emphasis on identifying user behaviors and needs. So as one of the first steps along the summer roadmap, I planned and ran a focus group with target end-users to understand their current course selection journey and to observe how they were using Curricle to perform their tasks. The focus group followed by semi-structured interviews with participants confirmed that while there existed a need for novel—possibility serendipitous—ways of course discovery that gave students confidence in their class picks, there remained a gap in understanding of why and how Curricle would be helpful in facilitating this process and the desired sense of confidence.
ACTIONS TO CLARIFY
To clarify the how, I synthesized the research insights and mapped out a user flow diagram for the platform to identify existing and potential pain points users have while navigating through the site, confirming these barriers with some follow-up interviews. Together with the directors and development team, we drafted up a list of navigational design refinements with level of priority, development cost, and estimated impact. The list included broad changes such as page reorganization to facilitate the course shopping journey, as well as more granular UI additions such as pointers and clearer instructions for the visualizations.
Site flow map showing how a user might navigate through the platform and where friction might arise.
Collaborative effort in using the flow map to identify potential points of further user research and brainstorm methods.
Illustration for design document communicating usability changes to the left-hand navigation UI.
A key part of bridging the gap in understanding what Curricle could do involves building (and positioning) the platform not as a competitor of the existing course selection and registration system, but rather as a companion tool that complements the current process. Close integration with the ecosystem enables us to cut costs and build with features that already exist and seek support from stakeholders. Our team met with key figures at Harvard, including the Dean of College, the Registrar, the Harvard Initiative for Learning and Teaching (HILT), and the Advising Programming Office (APO), to explore ways of integration. Together with our development team, we were able to strategize integration with student evaluation records, syllabus data, and course registration portals, all while ensuring that we stayed on top of our roadmap without being bogged down by possible hurdles, especially in working with access to confidential data.
Clarifying the why was a lot trickier. We lacked clear KPIs and an integrated data collection process from the first beta rollout. In order to validate the value of Curricle, I worked with our directors to draft a list of success metrics, identified the set of user analytics data for each metric, and worked with the team to build the data collection infrastructure. The Berkman Klein Center already employed a site analytic service called Matomo on a few of their own projects, but Curricle was not yet on board. Thus, with a list of success metrics on hand, I pushed for tighter integration with Matomo and worked with development to implement key trackers, all while maintaining constant dialog with the Harvard Law School Cyberlaw Clinics to ensure that we collect and handle data ethically.
A screenshot of key metrics compiled and ranked by priority and the infrastructure necessary to capture them.
ACTIONS TO IMPLEMENT
Communicating with development was challenging for me at first: the Curricle platform has been in early development for more than a year, so it took me sometime to understand the existing software architecture and enough developer jargon to even begin writing coherent ticket items on our project management tool. The learning process involved a lot of programming bootcamp videos and one-on-one meetings with our creative technologist and director of technology.
Our backlog seemed to grow endlessly after every meeting. That was exciting for us, as this is evidence of all of our ideas for how our platform could deliver value to a student’s academic experience. But it was a challenge to identify which features to deliver within our tight timeline. The process of prioritization was complex, involving many conversations with our faculty directors, revisits to research insights and our described success metrics, and discussions with our software engineers on cost and timeframe estimates. As I was trying to implement a more agile process, I also had to be mindful of competing projects already on the development team’s plates, which had impeded effective communication and timely feedback. To address the tension, I met up with our Director of Technology separately to bridge the vocabulary gap, adopted more visual communication strategies to minimize back-and-forths, understood and respected the development team’s work patterns and preferences, and moved some of our meetings closer to their primary office. I could not change the priority of projects on their end, but our communication improved, and we were able to get consistent feedback, even if it’s feedback about possible delays.
Illustration for design document showing user interaction when hovering over dot diagram of offered courses.
Graphics accompanying design document used to communicate navigational changes to the development team.
OUTREACH
I worked with my faculty directors and the summer student intern to design and plan the outreach campaign. The outreach package included lobbying university administration to get our product on their start-of-semester official communication to students, flyering on campus bulletins, publicity emails through major student organizations, feature stories on The Crimson and the Harvard Magazine, and a schedule sharing feature that we built to test how much traction we can gain from the organic peer-to-peer interactions that was already happening during the course shopping week. I even leveraged my experience in making animated films and our associate director’s voice-acting background to storyboard, compose, and edit a promo video for dissemination. It felt good to pick up that animation toolset of mine again!
The many version of posters printed and ready to go up at various high-traffic points on campus.
Preliminary storyboards for the Curricle promo video.
LAUNCH AND RESULTS
Our platform launched on August 30, right when the students were starting to return to campus and putting together their shopping schedules. When the user data started pouring in, we were overjoyed! The adoption rate in the undergraduate student body was over 37%, with an average engagement time of 8.1 minutes (returning visits were even higher, at 9.2 minutes). We had almost three times more user interactions (as clicks, schedule adds, annotations, etc.) than our first roll-out, but it’s noteworthy that a large percentage of interactions still happened in our more conventional Search functionality, and the majority of our conversion (defined as when users add courses to the in-app course planning tool) happened through conventional Search as well, rather than from the more innovative visualizations.
The range of qualitative feedback we received indicated a desire for alternate ways of discovering interesting classes, so in that sense, Curricle had found a design opportunity. The approach it has taken in providing new modes of discovery and encouraging students to step out of the convention of keyword-based searches would likely need more work. I synthesized our findings from this Fall launch and collaborated with our directors to write up the Curricle story into a prospectus report for key university administrators, who received the performance of our Fall launch very enthusiastically, and we discussed next-steps for a potential partnership with the university IT department toward full integration into the larger university ecosystem.
In addition to integration within Harvard, we are currently in dialogue with universities around the Greater Boston Area to explore how Curricle can be iterated and implemented with other university catalogs. The mission of Curricle to foster intellectual reflection on the value of cross-disciplinary perspectives in a liberal arts education can positively transform a student’s learning experience, and that is something we want to see happening on other campuses as well.
Refined landing page (mid-summer) with key functionalities sectioned in the left-hand navigational panel and new copywriting.
The Keyword Comparison mode of Curricle, where a student can explore the space of overlap between two departments, fields, or even ideas.
Many thanks to Jeffrey Schnapp, Matthew Battles, Jessica Yurkofsky, Kim Albrecht, Sebastian Diaz, Jon Murley, Patrick Lewis, Sarah Newman, Sydney Lewis, Alex Dolan-Mescal, and the metaLAB (at) Harvard community.