Winnow Health
The most complex workflows aren't just technically dense - they're socially loaded. Winnow sits at that intersection: a referral tool for physicians who stake their reputation on a recommendation, and a management system for recruiters.
Client Project
Product Design
Lead Designer
Enterprise Web Platform • Webb App
Oct 2025 - Mar 2026
Designing trust into physician hiring.
Physician referrals are the strongest hiring signal in medicine - yet the entire process lived outside the platform. I led the design of Winnow's end-to-end referral experience, turning informal peer trust into a structured, trackable, scalable system.
Problem
Referrals were happening. Just nowhere useful.
Winnow's initial release met beta expectations but lacked enough differentiation to compete against ATS platforms. Customers were benchmarking spend against established tools — and the product needed a genuine value proposition to justify it.
The answer was already in the data. Peer-referred physicians stay 70% longer, perform 33% better, and convert at 3–4× the rate of other candidates. But the entire referral process — sharing connections, following up, tracking outcomes — was happening in static documents, emails, and phone calls. Completely outside Winnow. No structure. No visibility. No data.
Painpoint 01
No platform home
Recruiters had to leave Winnow entirely to initiate, manage, or follow up on any referral.
Painpoint 02
Physicians had no easy path in
No frictionless way for physicians to refer colleagues - any account creation requirement killed participation.
Painpoint 03
Volume had no structure
No filtering, sorting, or urgency signals meant grids filled with noise and strong referrals got lost.
Painpoint 04
No proof of value
Without tracking referral-to-placement conversion, Winnow couldn't demonstrate ROI to customers.
Research
Two users. Two very different relationships with referrals.
We knew from the start that this was a two-sided problem - and that meant two distinct research tracks. We couldn't design for physicians and recruiters from the same set of assumptions. Their contexts, motivations, and friction points were fundamentally different.
On the recruiter side, we ran structured interviews walking through live referral scenarios - before any design decisions were made. What we uncovered went deeper than missing features. Recruiters had lost trust in the process itself. Once a referral was sent, it disappeared. No open tracking, no completion status, no way to know if a strong candidate had gone cold or was simply waiting on a response. Follow-through depended entirely on individual habits, not the system.
The four pain points we identified in the problem phase were validated and sharpened through this research - confirming that visibility, ownership, follow-through, and proof of value weren't just assumptions. They were consistent themes across every conversation.
On the physician side, we focused on understanding what would actually motivate a busy clinician to take time out of their day to refer a colleague. The answer was clear: it had to be effortless. Physicians stake their professional reputation on a referral - the experience needed to honor that weight without adding to their workload. Any friction, including account creation, would kill participation before it started.
Early concept testing with a digital referral form confirmed the core design principle that would drive every physician-facing decision:
No login. No barrier. Just a referral.
Solution
Two connected experiences. One closed loop.
Designed both sides of the referral system simultaneously - a physician-facing form and a recruiter-side management experience - so the loop would be truly closed for the first time.
MVP
Core Loop Foundation
Physician referral form (no account creation), centralized recruiter grid, candidate engagement, basic status tracking. Shipped the concept and validated physician willingness to engage.
Phase 2
Scale & Usability
Full filtering and sorting, archiving, stale indicators, resend capability, sidesheets, and telemetry. Made the MVP usable at real referral volume - completing the core loop.
Phase 3
Intelligence Layer (Planned)
Provider performance dashboards, full conversion tracking, automated reminders, and predictive prioritization based on historical placement data.
Research Synthesis
With a substantial body of existing recruiter research across multiple studies and interview transcripts, I used AI to surface recurring themes, flag contradictions across testimonies, and prioritize pain points by frequency and severity. What would have taken days of manual affinity mapping was condensed into a focused synthesis I could validate and build on - letting me spend more time on interpretation than aggregation.
Early Ideation
In the early stages of designing the referral experience, I used AI to rapidly explore interaction models, challenge assumptions about the physician form flow, and pressure-test structural decisions before committing to them in Figma. It helped me move through a wider solution space faster, so the concepts I brought into critique were more considered and better differentiated.
0 → 1
Centralized Referral System
Before this project, there was no in-platform referral experience - zero infrastructure, zero tracking. We went from nothing to a fully integrated, two-sided system shipped across two epics and 20+ product stories.
2× Faster
Recruiter Triage & Prioritization
Phase 2 filtering, sorting, stale indicators, and sidesheets were designed to cut the time recruiters spent finding and acting on referrals - moving from manual grid scanning to structured, signal-led prioritization.
70%
Longer Tenure — The Design Target
Referred physician hires stay 70% longer than non-referral hires. Every design decision - from zero-auth physician forms to recruiter-side conversion tracking - was made in direct service of this outcome.
20+
Product Stories Shipped
Designed and delivered across two full epics - spanning physician form authentication flows, recruiter grid filtering, candidate sidesheets, email resend logic, and engagement module integration.
3–4×
Referral-to-Hire Conversion Rate
Referred candidates convert to hire at 3–4× the rate of non-referrals. Moving referrals in-platform made this conversion trackable - and improvable - for the first time in Winnow's history.
^ROI
Enterprise Client Value
The referral program was directly tied to Winnow's beta customer value metrics - contributing to the product's ability to demonstrate measurable recruiting ROI and retain referenceable enterprise clients.
Both Phases Shipped
Live
Delivered across two phases and actively in use - a recruiter-side management system and physician-facing referral form, both fully integrated into Winnow
Primary Business Metric
Interviews Generated
The goal was more interviews from referrals. Building the experience in-platform made that measurable - and improvable - for the first time.
Proprietary Data Asset
Physician Graph
Every completed referral validates a physician-to-physician connection, compounding into Winnow's proprietary Physician Graph for future AI-driven matching.
01
Make the MVP cutbacks list a stakeholder communication tool
We tracked descoped items internally - but a more visual, shareable version of that list would have set better expectations and made Phase 2 prioritization conversations significantly easier.
02
Bring physicians into the research earlier
Our research was primarily recruiter-facing. The physician form was designed on low-friction principles and sound assumptions — but earlier co-design sessions with physicians could have surfaced trust and tone considerations we addressed blind.