Zocdoc
Designing the onboarding journey for large healthcare practices
TL;DR
Zocdoc's provider onboarding was fragmented, with disconnected flows and manual work slowing enterprise go-lives and straining ops. I designed a unified onboarding system: a new enterprise Home, guided tasks for staff, and a bulk NPI tool, giving practices a single source of truth.
Within 3 weeks of launching the redesigned onboarding flow, large practices with 15+ providers cut time to first bookings from 8 weeks to under 6.
Problem
Zocdoc's onboarding was never designed as a cohesive system. It evolved over time through product and sales-driven changes. By the time I took it on, "onboarding" was a patchwork of disconnected flows, each with its own entry point, mental model, and points of failure.
For providers and practice staff
- There was no single place to understand progress. Status was scattered across emails, CSM spreadsheets, and Slack.
- Entry points like Sales handoffs, bulk uploads, NPI flows, and EHR integrations all behaved differently, with no shared language.
- It wasn't clear what the practice owned versus what Zocdoc handled, so work often stalled while teams waited on each other.
- Enterprise admins had no reliable way to onboard new staff or repeat a successful rollout across locations.
For internal teams
- Ops relied on manual spreadsheets and ad hoc check-ins to keep things moving.
- Product teams kept adding new onboarding flows without a shared model, increasing fragmentation.
- Simple questions like how long onboarding takes or where it breaks down had no clear answers.
Net result: onboarding was hard to follow, easy to break, and costly to support. It also limited our ability to scale to larger, more complex enterprise customers.
Reframe
The first thing I challenged was the framing. The brief was to “improve the onboarding flow.” I pushed us to go deeper and rethink the system.
Fixing a single flow would address the most visible issues. Redesigning the system would:
- Create a shared understanding of what onboarding is — and what it isn't.
- Make each task measurable so we could identify drop-offs and prioritize based on real impact.
- Support scale as Zocdoc moved into more complex enterprise segments.
This shift in framing was the most important decision I made on the project. It defined the scope, the artifacts we created, and how we aligned with product and engineering.
Discovery
Partnering with my user research lead, I ran a focused discovery sprint across three groups: practice admins and staff at 5 enterprise accounts, customer success and operations teams closest to where things were breaking down, and internal product and engineering partners to map system constraints and data models.
Key findings that shaped the design:
- Staff weren't struggling with the interface. They were unsure what to do next or who owned the step.
- Enterprise admins needed a management view, not another task list. They were coordinating across teams, locations, and specialties.
- The riskiest moment wasn't a confusing form. It was the silence after submission. Practices assumed progress when work had stalled.
- Ops teams were effective but overloaded. Better provider tooling could unlock capacity without introducing new margins of error.
What I designed
The system model
Before designing screens, I defined the core model that everything would follow:
- Prototype 1: Entities: Practice → Locations → Providers → Tasks (a smart grouping and hierarchy, not a flat list)
- Prototype 2: State model per entity: Not started → In progress → Blocked → Pending review → Live
- Prototype 3: Ownership model: every task clearly assigned to Zocdoc or the practice so it's obvious who moves it forward
This made instrumentation straightforward early on. We knew what to track, where users were getting stuck, and what each drop-off represented.
Surface 1 — Enterprise Home
Home became the centerpiece: a cockpit that gives enterprise admins a clear, system-level view of their entire practice. Admins were managing dozens to hundreds of providers across locations and specialties. A flat task list didn't scale. A dense table wasn't actionable. The interface needed to answer, at a glance: are we performing, are we set up correctly, and what needs attention right now?
What Home answers instantly
- Are bookings up or down this month?
- Is anything broken or blocking appointments?
- Which providers or locations need attention?
- Are all providers live, mapped, and accepting patients?
- Are our channels and integrations working as expected?
- Where are we leaving growth on the table?
Hierarchy as navigation
Integrations → Locations → Providers. Supports both high-level oversight and drill-down without losing context. Flat lists and timelines failed in testing.
From tasks to system health
Shifted the mental model from "what do I need to do" to "what's happening across my system." Account Health highlights urgent issues like missing availability, broken integrations, and misconfigured insurance. Insights surfaces optional optimizations — expand hours, improve profiles, adjust spend.
A shared rally point
A single source of truth for CSMs, ops, and admins. Built to be easy to screenshot and share in updates.
The layout uses a single-column, scrollable structure to maintain narrative flow, with clear section breaks to manage density. High-signal metrics are surfaced first, with supporting context progressively disclosed. Status indicators rely on more than color alone — iconography and labels ensure clarity at a glance. Empty states are intentional and reassuring, reinforcing when the system is healthy rather than leaving gaps.
Surface 2 — Get Started tasks
For staff doing the work, I designed a simple checklist experience grounded in how practices actually operate. Staff were often part-time, non-technical, and juggling onboarding alongside other responsibilities. They needed clarity over completeness.
Group tasks by how staff think
Moved from system-based groupings to workflow-based ones like "Add providers," "Configure availability," and "Set accepted insurance." This reduced drop-off in testing.
Make ownership explicit
Clearly label Zocdoc-owned vs. practice-owned tasks. This removed the most common question: who is responsible for the next step?
Keep the main path simple
Progressive disclosure with expanding step details. The core flow stays clean while still supporting more complex scenarios.
Surface 3 — Bulk NPI provider add
For enterprises adding large provider rosters — sometimes hundreds at once — I designed a dedicated bulk-add flow powered by NPI pre-fill. The previous approach relied on spreadsheet uploads with no feedback until after submission, with errors showing up days later as support tickets.
Validate early, not after the fact
Real-time NPI validation surfaces issues like specialty mismatches or unclear roles before submission. Errors are clearly categorized as fix now or review later.
Preview before commit
Admins see exactly what will be created — providers, locations, and specialties. This reduced accidental submissions and rollback requests.
Key decisions & tradeoffs
Prioritized in-flight customers, not just new ones
We designed for accounts already mid-onboarding, not just clean-slate flows. That meant handling messy, in-progress states from day one. It added complexity, but delivered value faster to our largest customers.
One Home, not scattered widgets
Instead of spreading status across pages, we created a single Home surface. This required rethinking navigation and retiring legacy patterns. The result is a clear control center that scales without fragmenting attention.
Tasks and states over timelines
Onboarding is modeled as tasks with states per entity, not a linear progress bar. More complex to design, but it supports parallel work, clear ownership, and reliable measurement.
Human-in-the-loop for compliance
For credentialing and specialty mapping, we kept human review in the loop. It slowed full automation, but reduced risk and matched regulatory and contractual realities.
Extensible EHR patterns, not deep integrations
We avoided deep integrations early and focused on flexible patterns. Fewer immediate wins, but a more maintainable foundation and a realistic first release.
Outcomes
Shipped
- Enterprise Home rolled out to 13 enterprise accounts in March 2026
- Guided staff onboarding flows launched to 50% of practices with < 15 providers
- Bulk NPI tool released to all enterprise accounts
Impact
| Time to go-live | Reduced from 8 weeks to under 6 |
| Ops overhead | Manual effort per enterprise onboarding down from 37 hrs to 29 |
| Support tickets | "Where are we in onboarding?" tickets down 11% |
| Sales escalations | Onboarding-related escalations down 18% |
| Adoption | 87% of enterprise admins used Home weekly within 30 days of launch |