Positions for Hired
Redesigning one side of a two-sided marketplace
Hired is a two-sided marketplace matching employers with qualified candidates. Late in 2016, we recognized that we had created a problem in our employer experience – we were asking recruiters to search across huge groups of candidates, rather than providing them with a limited, more relevant set based on their specific needs.
I was the design lead for the team responsible for solving this problem. We created the concept of a Position, which is a streamlined job description that the employer can provide to Hired so that we can then return a list of highly relevant candidates. We also added collaborative support to allow hiring managers and recruiters to work together to review these candidates.
A bonus of the Positions solution is that it has also allowed Hired to collect more granular demand data across all our employers, so we can proactively acquire the kinds of candidates our employers are looking for.
I worked closely with Aaron Joseph (Product Management) and Vanessa Van Schyndel (User Research), plus a team of seven engineers to design, test and build the product. Since September 2016, we’ve run four workshops, one client on-site research study (with a total of four employers), eight usability studies (with a total of 24 users), and made 20 prototypes with Invision and Framer.js.
We organized this project into three phases:
I’ll focus on the second phase, the collaborative workflow, to give insights into my design process through a detailed case study.
After Positions launched in 2017, our employers were able to create a position and receive recommendations for those positions. Next we wanted to focus on how employers could identify and connect with the candidates they were most interested in. Our hypothesis was that if we made this easier, more interview requests would be sent, which would in turn increase placements and therefore revenue. We also believed that by improving collaboration features, we would be able to increase retention and sideways growth into our client’s organizational structure.
Vanessa and I ran four on-site research sessions with recruiters at mid-market companies.
We learned how these recruiters process long list of candidates, both through Hired and our competitors (such as LinkedIn and Indeed Prime). They skim through the list, opening promising candidates in new browser tabs. Then they work through the tabs, moving each tab left or right in the browser based on whether that candidate was a match. This process looked cumbersome and inefficient, and team collaboration was near-impossible.
We conducted a synthesis workshop to organize our findings.
Our most significant finding was the detailed understanding of the current workflow. We identified two important “How Might We” opportunities:
Next, we ran a design workshop with designers, engineers, product managers and sales execs. The goal of the workshop was to generate ideas based on the “How Might We” questions.
After the workshop, we grouped the ideas into themes, and chose the ideas that would best serve our business goals and provide the most compelling experience for our employers.
I sketched out various UI options to explore how these ideas could work.
I made an Framer.js prototype after mocking up the UI concept in Sketch. We took this prototype back to our recruiters. We wanted to know if the shortlist could replace their tabbing behavior, and whether the nuances of the collaborative workflow made sense. For example, if a recruiter hid a candidate in the list of recommendations, should that also hide the person from all the other collaborators using the list?
We found that our recruiters initially didn’t see value in the shortlist until they used it. Then they discovered that the candidates added to the shortlist no longer appeared in the full candidate list, so they were able to reach a satisfying “inbox zero” state for the full list – which they found very powerful. This insight, about the relatively late point in the user journey at which the value became clear, showed us that product marketing would be especially important for this feature. We also identified several minor problems with the interface that we were able to fix.