Objective

Create a new Job Match experience utilizing new personalization and AI capabilities in a new codebase.

Role

Generative Research Lead, Research Synthesis, Wireframing, High Fidelity Design, Design QA

Goals

• Create new web product to smartly match users with available jobs

• Increase job applies + job apply rate

• Reduce “dead time” spent browsing lists of jobs

• Increase monthly active users

―

Searching for jobs takes a lot of time. Some job seekers like to use boolean search and filters to only see jobs that are an exact fit for them, but they could miss seeing jobs on the fringe of their requirements. Some job seekers use as few filters as possible and read through every single listing, but they could miss a good-fitting job for them by browsing too fast, or not going far enough in the results to see their perfect job.

Our task was to serve both of these opposite ends of the job seeker spectrum with the introduction of Job Matching to Built In. Users who like to tailor their job search will only see exactly the jobs they want to see. Users who want to see more variance can get started with Job Matching with just a job title and location.

In both cases, Built In began leveraging new personalization and AI capabilities to learn from job seekers’ behavior to expand the matching criteria for super filter users, and narrow the criteria for those with minimum search criteria.

Initial Research

Job Matching had a tight timeline to get to market. Job seeking was trending upward in general, and a competitor had just exited the US market. Built In saw an opportunity to fill a gap with a brand new job match product.

Because of this accelerated timeline, there was a focus on building the minimum number of filters that users would find important. I worked closely with Product, Engineering, and Data to determine all of the facets we had available to us, and then set out to prioritize which facets were most important to users.

Without time or budget for more traditional research opportunities, I worked with Product to conduct 8 quick card sorting sessions with individuals we knew. This allowed us to understand what attributes were most important in evaluating whether or not a job is a “match,” and prioritizing those attributes for the Job Match MVP.

Ideation

After determining which job filter options were feasible to display, we prioritized based on the card sorting data. This led to the final step: create an efficient and responsive design that kept the desired information hierarchy regardless of what device a job seeker was using.

Design + Impact

Job Matching launched to a subset of users in October 2022. Even though I left Built In before attaining significant measurable results, there was high confidence in the org with the new user experience for job seeking on the platform. So much so that when I was tasked to redesign Built In’s traditional (non-match) job posts for the new backend technology, we opted to re-use the same design as Job Matches with minor modifications to displayed information.

Preliminary Results

2x

increase in conversion to job apply compared to Built In’s job board experience in the first two weeks