Indeed is the #1 job site in the world with over 300M unique visitors every month. I helped launch a more personalized job search, spearheading the design of core search experiences.
Problem Space
The shear number of jobs on Indeed is a double edged sword. While it gives jobseekers tremendous power to find a job, it also makes it hard to sift through all those jobs to find something relevant. We needed to help job seekers quickly identify relevant opportunities.
Role & Contribution
Several teams where involved in bringing the personalized search strategy to life. I was tasked with figuring out how we would surface matching signals on SERP (search engine results page) and how we could make filters more useful to the job search experience.
Timeline
Personalized search was a multi-GM initiative that spanned the course of a couple of years. My work happened over the course of a quarter and a half.
Outcome
While I can’t specifically state the outcomes I can say that my work directly helped millions of job seekers find better jobs, faster. This increased key down-line KPI’s such as # of qualified candidates (as rated by employers) and the # of positive outcomes for job seekers (e.g. interviews, job-offers, etc).
Job Card
All job meta-data was given a tag treatment to increase scan-ability. Job attributes that could be matched with user a preference were given more visual weight and an icon to accelerate understanding.
Conditional Logic
Job card height affects the number of jobs a jobseeker can see in a given space, which in turn negatively affects KPI’s. To limit job card height and retain the value from current modules, I designed a dynamic display system. The more personalized a module was, the higher priority it received.
Guidelines
In order to spend less time in meetings reviewing the direction with stakeholders across Indeed and more time designing, I put together documentation outlining design decisions. I advised stakeholders review first before scheduling a meeting. This documentation was later rolled into design system documentation.
Job Details
I aligned metadata tags and matching signals with the job card and reworked the visual hierarchy to help accelerate scanning.
Feedback Loop
I felt it was important to leverage matching signals or lack thereof to trigger a behavior of adding or updating job preferences. I successfully pitched and got alignment to design a simple feedback loop to help encourage this behavior.
Multi-select Filters
Multi-select filters had been a long time ask from jobseekers. Teams in the past tried to implement the feature but both attempts ultimately failed. Taking these learnings, I re-worked how to approach the design and functionality of filtering.
Next Up
Building global taxonomies at scale and speed