2 min readJust now
–
I built a specialized nursing job board with one goal: to replicate Indeed’s filtering experience for healthcare jobs. Multiple filters for shifts, work settings, and license requirements. After months of work and a few weeks live, I scrapped it all for a single text box. That decision increased time on site by 400% in one week.
Press enter or click to view image in full size
Filters for Nursing Jobs (old design) in My RN Jobs
The Original Plan
Nursing job seekers have specific needs. They filter by shift patterns (days, nights, 12-hour rotations, PRN), work settings (hospitals, home health, clinics, travel nursing), license types (RN, LPN, CNA, NP), and specialities ([ICU](https://myrnjobs.com/searc…
2 min readJust now
–
I built a specialized nursing job board with one goal: to replicate Indeed’s filtering experience for healthcare jobs. Multiple filters for shifts, work settings, and license requirements. After months of work and a few weeks live, I scrapped it all for a single text box. That decision increased time on site by 400% in one week.
Press enter or click to view image in full size
Filters for Nursing Jobs (old design) in My RN Jobs
The Original Plan
Nursing job seekers have specific needs. They filter by shift patterns (days, nights, 12-hour rotations, PRN), work settings (hospitals, home health, clinics, travel nursing), license types (RN, LPN, CNA, NP), and specialities (ICU, Emergency, L&D, Paediatrics, Oncology).
I built filters for all of it in My RN Jobs. Dropdowns, checkboxes, multi-select options. If you’re curious about the technical approach, check out “Copy Indeed’s AI Strategy” webinar on YouTube, where Alexander Chukovski** **and I explain the technical details of using AI to create custom taxonomies that power these filters and improve job matching.
The problem? Users didn’t engage. Session times were short. The sophisticated system I’d built went underutilised.
The Pivot to Free Search
With little time on site after searching, I decided to try something different. Just a one-text input with free text to allow users to write whatever they want. No dropdowns. No checkboxes. I didn’t know what to expect.
Press enter or click to view image in full size
The Results
One week after launch, users engaging with search spent 400% more time on the site. But the real value is in what they’re searching for:
Compensation matters. “Bonus”, “weekly pay” appear constantly.
Schedule flexibility. “No weekends,” “day shift,” “PRN”, “flexible.”
Career growth. “New grad”, “residency program”, “tuition reimbursement.”
Every search query is an insight into users’ intent and what matters to them, so we can improve the search and taxonomy for better results.
The Lesson
I spent months asking, “How do I build better filters?” The breakthrough came from a different question: “How do I understand what users actually want?”
Tech Stack
My RN Jobs runs on:
- TypeSense
- Postgres on Neon
- CloudFlare Workers for hosting and cron jobs.
- LLM on OpenRouter