Is It Me You Are Looking For? Leveraging and Improving the Search Bar to Improve Students’ Experience on KNET
Applying User-Centric Design to KNET
User-centric design can help design a KNET that students love to use. User-centric design starts with user needs and derives product features from them rather than trying to find users for product features. It starts from a place of empathy, humility, and respect, not pre-supposing what users want or what features will meet their needs. Rather, user-centric design it a continuous process, iterating over product features informed by both quantitative and qualitative feedback.
Students have various customer jobs, pains, and gains. Students’ most important customer jobs are functional and relate to finding information relevant to achieving their educational goals and completing their degrees (e.g., on degree requirements, course registration, etc.). However, students’ customer jobs may also be of social or emotional nature, such as wanting to feel part of the HKS community, or relate to co-creation, such as organizing events or share feedback. Relatedly, students experience various pains, the most significant being missed deadlines and missed opportunities as well as time spent trying to find information. Students’ customer gains include finding helpful information they were not specifically looking for.
KNET’s current value proposition addresses students’ essential jobs and customer pains only to some extent. KNET does serve as a one-stop-shop for authoritative information on degree requirements, suitable courses, etc. However, interviews with focus groups (i.e., breakout rooms in lecture) showed that students currently generally have difficulties to find the relevant information on KNET.
A long and winding customer journey
An exemplary customer journey can illustrate the point. Take a first-year MPA/ID student wanting to know the deadline for course enrolment.
Landing on the KNET homepage, she may feel overwhelmed by the number of links and options to go next presented to her. She may explore quick links (“What are PeopleSoft and Spacebook and why do I have to get there quickly?”), the menu bar, the various links on the website (“Frequently asked questions! But only for Covid-10 resources”), or the search bar, if she notices it.
She may also scroll down the website, only to find more links, such categorized under ‘popular sites’ and ‘training & tools’.
Suppose she starts with the menu bar, as it is a relatively salient object on the landing page. A new student may not know that course registration falls in the domain of the Office of the Registrar (85% of MPA/ID students are international students, and many other higher education systems do not have a separate office under this term). Being an MPA/ID student, she may look for resources for MPA/ID students. However, it is not clear whether she can find the information on course enrollment. In “FAQ” or “General FAQ”? In “Academic Policy”?
She may next try the search bar. The search term “deadline course enrollment”. Many of the 337 results link to out-of-date information (even if they were modified recently). However, hovering over the first search results shows a miniature version of the linked page, which looks relevant.
She lands on the Registration Overview site, which tells her that for the “specific semester’s dates and deadlines“ she should refer to the academic calendar.
She finally finds the information she was looking for. Note, however, that while the academic calendar seems to be housed by the registrar, the link to it was buried in the text, rather than in the side bar.
Recommendation
Make the search bar the most prominent feature on the KNET landing page and the main gateway through which students find information on KNET.
KNET, in its current structure, can be understood as a form of internal Wikipedia. Content creation is decentralized, with different administrators given the right to write the content for the respective domain (i.e., the registrar, dean of students, program directors all manage their own sub-sites).
The problem is that this internal structure is also used for how content is organized for users. The menu bar, for example, reflects this internal structure. This requires users to know the internal structure to know where they can find which information. In the exemplary customer journey, the student had to know that course registration falls into the domain of the registrar rather than, for example, the domain of the dean of students
The first-best solution might be to re-organize the content for students according to particular jobs they want to perform. Want to enroll in courses? Here’s one page with the course requirements for this semester (provided by the student’s program director), course listings (provided by the Dean of Student Affairs), and deadlines for enrollment (provided by the registrar). Currently, this information is spread out across the sites of the respective administrators.
However, such a reorganization of content for students would likely be technically and administratively difficult to achieve. It requires changes to the back-end system where administrators populate different sites, thus further down in the technology stack.
Therefore, a second-best alternative is to leverage and improve the search bar.
The design of the current landing page incentivizes students to use the menu bar instead of the search bar. The search bar is rather small compared to the other elements on the landing page and hidden in the corner.
However, the search bar allows users to access information without having knowledge of which topics fall into the domain of which administrator. In the user journey, the student was able to find the deadline for course registration through the search bar despite not knowing the registrar is responsible for this topic. Think of the analogy to Wikipedia. It would be absurd for users interested in, for example, the history of Harvard to navigate there through USA > Education > Universities > Harvard. Using the search bar is a natural choice.
The search bar should therefore be placed centrally on the website, making it the most salient feature on the landing page. The design of the Apple Support website serves as a good example.
Further, the functionality of the search bar needs to be improved. First, it returns a lot of outdated information. To nudge administrators to update their content and remove old files, all content should by default be taken offline over the summer break unless administrators explicity choose to keep it online for another year. Second, the search engine is not very powerful. For example, searching “deadline course enrolment” instead of “deadline course enrollment” returns only four results and does lead the student from our customer journey to the academic calendar.
The effects of these changes on the user experience should be carefully evaluated. The suggested changes are based on certain hypotheses about users that may or may not not hold. For example, it assumes that students come to KNET to look for specific information rather than just wanting to browse its content. It is therefore essential to evaluate the effectiveness of the proposed changes. For example, each site could have two simple yes/no surveys at the end, asking users whether the information was helpful and easy to find (as, for example, here on Apple Support). If the changes are rolled out first to a random subset of users, comparing the average ratings across groups could help us learn if we improved the user experience (besides, sites rated useful but hard to find could be prioritized in the search).
This memo illustrates how user-centric design can help create, step-by-step, a more student-friendly KNET. It can similarly help improve KNET for other users, such as administrators or faculty.
This blog post was was written in response to an assignment for the course DPI-662 Digital Government: Technology, Policy, and Public Service Innovation at Harvard Kennedy School.
Previously Submitted Pre-Class Assignments
