Voyager Health Operations Dashboard, a Product Design Case Study

The Program

Mars Veterinary Health is building Voyager Health, an enterprise practice management application to replace disparate legacy systems, including Woofware, the platform I originally designed for VCA in 2012.

Type: Enterprise SaaS
Market: North America • Asia • Europe
Team: 7 Product Managers • 3 Technical Product Managers • 12 Designers • 37 Software Developers
Role On Feature: Sole Product and Visual Design • User Research • Led Design System
Timeframe: 4-month Agile rollout • 2025

The Challenge

Veterinary hospitals run on controlled chaos. Appointment targets, room status, patient flow, client emotion, and follow-up tasks all happening at once. Each tracked a different way.

Having just wrapped the Accounts project, the PM approached me with a new feature idea. Before starting any design work, I returned to the hospitals with a new focus: what did staff need to see at a glance?

We really need a way to see Business Performance KPIs (Appointment Bookings vs. Target Goals) as that is critical for each hospital

- Banfield Staff Participant
It would be amazing to easily see all the arriving patients all in one place.

- Staff Participant
Sometimes it is hard to remember which room is occupied.

- Veterinary Doctor Participant
We need a better way to alert on a euthanasia

- Staff Participant
Wow yeah... I really want to know how many drop-offs we have in the morning. I have to know where to put them.

- Staff Participant
So, my call sheet is important, we need to call and confirm appointments, yeah... and we need to follow up with no-show clients. I usually just use a notebook to remember them.

- Staff Participant

What Success Looked Like

Improve appointment target hit rate across all brand hospitals
Reduce time to identify patient journey status and physical location
Increase room utilization and preparation efficiency
Increase SLA compliance on patient follow-ups and task completion
Ship WCAG 2.0 compliant across all dashboard features
Replace legacy workflows with features staff can learn independently, critical for high-turnover roles like front desk staff

How'd I get there?

A Multi-Brand Veterinary Operation

Voyager Health supports several retail brands within the Mars Veterinary Health Organization. Each brand has its own processes, legal jurisdictions, business goals, and ways it generates revenue. The largest being VCA and Banfield with 1,000+ hospital operations in the USA, Mexico, and Canada.

Diagram showing the organization structure of Mars Veterinary Hospitals and its brands: AniCura, Banfield, Blue Pearl, Linneas, and VCA.

Aligning to a C-suite directive, the product team's Phase 1 focus was on the two largest retail brands: Banfield and VCA, each running a different legacy practice management solution. The product goal was to unify all the retail brands onto a single platform.

The Strategy

Starting With People

My first move was to go onsite rather than start designing. I initiated generative research across five hospitals to understand what the dashboard needed to solve. Several features, including the Maître D' wait time estimator and the address book, weren't in the original brief. I recommended them based on what I observed in the field.

Deep Dive on the research methodology

Lobby and reception desk view of a veterinary hospital, I photographed while at an onsite observation location during discovery research.

In the Field

How Does a Veterinary Hospital Operate?

Veterinary hospitals are busier and more emotionally complex than most enterprise software teams assume. To understand what a dashboard needed to show, I needed to see the environment firsthand and not rely on stakeholder assumptions.

The Visits

VCA hospitals - observing Woofware in daily use
VCA hospitals - observing paper-based analog processes
Banfield hospitals - observing Legacy Software B in daily use

3 panel: 1st Panel two tablets open to the MVH practice mangement app. 2nd Panel a person recording vitals on a vital tracking paper. 3rd Panel a Windows desktop screen with a lot of icons of pdf files saved on the desktop

Observation Methodology

Rather than open-ended observation I focused on specific operational touchpoints:

What I Was Looking For

How are patients greeted, identified, and checked in?
How does staff track patient location and room status throughout a visit?
How does staff manage follow-up tasks and callbacks?
What are the emotional touchpoints for staff, clients, and patients?
Where do analog and digital processes coexist, and why?

UX Observation: Few animal visits are routine

Finding 1 — Technology has limits

Digital tools can streamline check-in, but they can't replace the human-to-human contact that differentiates a clinic when five competitors are within five miles. The personal touch is a business advantage.

Finding 2 — Dignity requires design

Euthanasia appointments demand a level of care and preparation that most software ignores entirely. Advanced notice, buffer zones, and spatial awareness aren't edge cases. This finding directly shaped Feature 4.

Finding 3 — Busyness is predictable

Two different hospital models, same core problem. During Banfield's morning drop-off window, typically 2 hours when 6 to 10 patients arrive at once, staff struggle to track boarding area occupation and check-in status simultaneously. At VCA, a steadier stream of clients accompanying their pets creates a different pressure: exam room status and patient location become hard to track across the day. The chaos isn't random in either model; it happens on a schedule. Designing one feature that solved both operational realities was one of the most satisfying outcomes of this project.

Listening Closely

I selected two participants per operational area across five hospitals, varied by tenure where possible. Beyond active staff, I also interviewed two former hospital employees now on the User Advocacy Team, a CSR trainer and a retired veterinary doctor, both with long tenures at VCA and Mars Veterinary Health, having trained staff on both Woofware and Voyager Health.

Field research in a working hospital requires situational awareness. I was a guest in a clinical environment. My presence couldn't disrupt patient care or staff safety, which shaped how I conducted every session.

What I Asked

I structured questions across three facets: technical, aspirational, and emotional. Here are a few examples:

How do you measure success in the hospital against the organization's benchmarks?
What are the common negative experiences during check-in and check-out?
How do you handle clients who are dissatisfied?
Have you had a difficult experience with a patient or owner? What did you do?
How do you create the opportunity for a personal touch with the client?

How I Synthesized

Notes from each session were grouped into an affinity board to surface patterns across roles and hospital types. AI assisted in synthesizing the results into actionable themes.

Meet the Users

During the foundational stages of Woofware, the practice management application I designed for VCA, a former design colleague and I created four archetypes based on hospital roles common across all hospitals.

For the Dashboard, the focus narrowed to Brianna and Asatyra. Their daily hospital duties, operational patient flow and financial performance tracking, mapped most directly to what the dashboard needed to solve. These two roles became the lens through which every dashboard feature was evaluated.

Brianna's day was reactive by nature. Check-ins, check-outs, patient tracking, room management, callbacks. The dashboard was designed around her rhythm.

Answers and makes calls, greets clients and patients, checks in and checks outpatients (CSR, Med Tech, Admin)
Escorts patients, takes initial health metrics, performs some medical exams and procedures (Med Techs)
Prints out discharge papers, collects patient belongings, and gives take home medication (CSR, Med Tech, Admin)
Tracks patient location and safety, stays with drop offs and patients undergoing more involved procedures (Med Tech, Doctor)
Sets and manages care teams (Admin, Med Tech, Doctor)

Asatyra needed situational awareness across the whole hospital at a glance. Her needs drove the KPI and appointment target features on the dashboard.

Tracks number of appointments against target goals (Admin)
Schedules follow ups, confirms appointments, calls no-show clients (CSR, Admin)
Tracks financials and staff schedule to set up care teams (Admin)
Builds client relations, Monitors patient location, staff safety, and client sentiment (CSR, Admin)

Supporting Archetypes

To map the full range of staff workflows and facilitate discussions with product and stakeholders, I created three additional archetypes: Dr. Gloria Monday (veterinary doctor), Crystal Hye-Sook (medical technician), and Dale-k (client).

From Research to Ideas

Feature Board and Ideation

A full brain dump of every possible job-to-be-done across all hospital roles. We sifted through the raw material to define a viable MVP. Every feature on the dashboard traces back to something captured here.

Pre-iteration: I mapped the data points most relevant to the CSR role and sketched a rough feature IA, translating the Miro brain dump into a first hypothesis about what the dashboard might contain. Brought to the PM to align before any formal iteration began.

Early Concepts - Designing the Fix

Iteration 1 - Structured for Unique Staff Roles

Iteration 1: Role-specific views per user type. Discarded when we learned smaller hospitals have staff wearing multiple roles simultaneously and one dashboard had to serve everyone.

Iteration 1: The admin view included comprehensive financial and AR aging features. Stakeholder and user feedback pushed these out as it was too much detail for a landing view.

User feedback changes my direction

Design Strategy

Personalize the dashboard experience with unique features per hospital role.
Why Discarded?

Staff often take on multiple roles in smaller hospitals.

Higher production and maintenance costs. Engineering flagged the significant effort required to maintain a fully personalized experience with unique features per role.
Design Strategy

Add comprehensive financial and AR aging features to dashboard.
Why Discarded?

AR Aging data is not critical for users on the landing view. Internal political sensitivities by other product teams.
Design Strategy

Add medical information for patients to aid in appointment scheduling.
Why Discarded?

Too much information and not useful in this view. Research feedback was that users and stakeholders needed more high-level operational information.
Design Strategy

PM suggested a news and weather widget.
Why Discarded?

Users in testing confirmed this feature wasn't valuable in this context

Rethinking the Approach

Structuring the Level A, B, C model

Iteration 1 told us two things clearly: too many role-specific views created maintenance problems, and too much information on one screen overwhelmed users. In a working session with the PM and VP of Product, we reframed the problem around three levels of information depth.

From that session, I concentrated on six features for Iteration 2:

Appointment Target KPIs for Banfield
Appointment Arrivals, Departures, Visit Tracker
Hospital Digital Care Teams Setup
Room Utilization a.k.a. Hoteling Status
Daily Task List - Pending Jobs To Be Done
Address Book (Nice to have)

Iteration 2: Collapsed all roles into a single unified view focused on the six most critical operational features. In an informal user test with hospital staff; the feedback was, the top KPI blocks could be simplified with less data.

Zeroing In On the Solution

Feature 1. Operational Target Appointment KPIs

Through user and stakeholder sessions, the Appointment Target KPI emerged as the highest-priority feature.

Iteration 3: Appointment target KPI one of the most requested features for Banfield.

Feature 2. Appointment Drop-Offs, Outpatient, Inpatient Tracker

Appointment Management is the nucleus of hospital operations, and this feature has many levels of drill down and actions users can take within the phases of the patient journey. For the dashboard we wanted to capture Level A relevant information. Who is coming in and when, where are they in the treatment journey, when are they leaving and with whom, and follow up is needed on discharge or in the future (Finding 2, 3), both surfaced in our hospital observation research.

Iteration 3: Who's arriving, where are they in treatment, when are they leaving? The tracker surfaces the questions staff were answering manually throughout the day.

One of my highest-impact recommendations was surfacing Patient Check-In directly from the dashboard with no navigating away required and reducing the user interaction by three clicks. Banfield stakeholders had explicitly prioritized reducing click depth across the experience.

Iteration No. 3: Surfacing check-in directly from the dashboard removed 3 clicks and saved over 2 minutes per patient. One of the highest-impact UX decisions on the feature.

Another feature I recommended was an End of Day Summary that would show once the last Upcoming Appointment was checked in or cancelled/rescheduled. This feature contributed to a higher user CSAT score and was well received by stakeholders to reinforce tracking the Appointment Target KPIs.

End of Day Summary appears automatically once the last appointment is checked in or resolved which reinforces the appointment target KPI without requiring any extra staff action.

Bonus Feature I recommended à la Maître D'

I designed an estimated wait time feature inspired by my experience in restaurant hospitality. The next intended phase would have been to train our LLM with situational data to make an AI driven estimation feature that would adjust in real time without manual input.

Iteration No. 3: The Maître D' feature: estimated wait time drawn from my restaurant industry experience. The MVP release is manual input; the next iteration will use LLM-driven real-time estimation.

Feature 3. Hospital Digital Care Teams Setup

Every clinic I visited had a literal whiteboard where care teams are listed for the staff, so everyone knows which medical technicians are working with which doctor. The board was updated daily or weekly. In larger hospitals these were up to 8 care teams. The typical structure was 1 or 2 care teams per day.

Collaborating with the feature PM and User Advocacy Team, the digital care team feature went through a few iterations. The first iterative idea can be seen in Iteration No. 2 screen above.

Changes I made to this Iteration No. 3 version were to remove the CSR staff and group the doctors together with the medical technicians as sub elements. Also, we added a new team lead designation feature to the medical technicians.

Iteration No. 3: Every hospital I visited had a physical whiteboard listing care teams. This is its digital replacement, the same mental model, no whiteboard required.

Iteration No. 3: Care team setup directly from the dashboard. I added a team lead designation for medical technicians based on direct feedback from hospital admin staff.

Feature 4. Room Utilization (Hoteling Status)

Based on my Hotel and Hospitality experience and addressing Findings 2 and 3 in Starting With People above, I created a room utilization feature. A recurring theme across all hospital visits was how hard it was to track room statuses. That pain point became an opportunity for delight.

What is the condition of Exam Room 3? Is it still occupied?
Which exam rooms should we not put other patients in during a euthanasia?

There is the need for a quiet, respectful, and dignified buffer zone for the most difficult of times.
Which exam rooms/areas need attention? Did a patient make a mess that needs to be cleaned up?
Which exam room is that abusive/excitable/unpleasant client in?

Through iterative reviews with the PM and User Advocacy Team I created the status display for Phase 1 with a plan for additional features for future phases. Ask me about the plan to make this smart and automated.

Iteration No. 3: Room status during a euthanasia appointment created a buffer zone feature designed around dignity and staff awareness, surfaced from onsite observation and user interviews.

Iteration No. 3: Assistance alert state: room condition and availability tracked in real time, replacing the verbal check-ins and guesswork I observed in every hospital visit.

Feature 5. Daily Task List (Pending Items)

Generative research surfaced a clear wish: a way to manage daily follow-up items without relying on notebooks that were occasionally misplaced.

Iteration No. 3: Staff were tracking callback lists in notebooks. The task list digitized a behavior that already existed, all in one place, following where the staff was, which is why it tested well immediately.

Iteration No. 3: No-show follow-ups consolidated into one list, previously scattered across notebooks, sticky notes, and memory. One participant said 'I misplace my notes all the time. This is cool.'

This feature was met with tears of joy, maybe not literally, but this consolidated list and placement on the dashboard hit the mark.

Wow that will make callbacks so easy!

- Staff Participant 3
I usually have a notebook for remembering who I need to call, and I forget where it is sometimes. Well done team!

- Staff Participant 7

Feature 6. Address Book (The Impactful Value Add)

In a few hospitals we noticed post-it notes all over the staff's work area that contained addresses of outside organizations that the hospital frequently contacts: e.g. surgery hospitals, labs, specialists, other brand hospital locations, local food places for delivery or pickup. We noticed staff looking up local bakeries and restaurants during their shifts. So, I had an aha moment! Let's add the ability to include them in the digital address book as well.

Operations dashboard interface with the right-hand Address Book panel expanded, showing a searchable list of in-network hospital locations and contact cards for local businesses. — Iteration No. 3: The address book started as a post-it note observation, staff had external contacts and local restaurants written on notes all over their workstations. We made it a feature.

Expanded Address Book sidebar panel on a hospital dashboard, with the search input box in a focused, active state for typing. — Iteration No. 3: Search within the address book for internal and external contacts in one place, no more post-its.

Feature 7. Personal Notes (Another Impactful Value Add)

Another curiosity in our observational research was how many notes the staff scribbles down as daily or hourly reminders. We created a simple digital private and personal note taking space in the dashboard. It tested well, but it generated an unexpected, repeated question: were the notes private to the user writing them? Unlike other note-taking features in the platform, which are visible to all users, these are private. To solve this in the MVP release, I agreed with the User Acceptance Team that this question would be answered in the release notes and concierge training program.

Operations dashboard UI view highlighting a fixed 'My Notes' button positioned at the bottom right corner of the main content area. — Iteration No. 3: Notes input dialog: a floating button kept deliberately out of the way since notes are a 'nice to have' rather than a primary workflow. Users found it easy to discover.

Visual Design Solution

The Platform's Visual DNA

While the Design System and Visual Language were a team effort, I drove four main sections of the design system.

Base Visual Design Layout
Illustrated Patient Avatars
Icon Style Design
Accessibility

1. The Base Visual Page Design

The base layout uses a header and columns grid covering 90% of the UI, with a light gray background and off-white cards.

The base grid was established early with engineering to align on page types and reduce layout decisions during development.

2 Column Body Layout (Skinny) - With Header

2 Column Body Layout (Wide) - With Header

2. Illustrated Patient Avatars for Quick Species Identification

The culture and user behavior encourages using photos of patients for identification and species.

Illustration Evolution

Getting the emotional tone right for dogs and cats was the focus as user acceptance depended on it.

Dogs

Early feedback was the dog looked mean or too serious. The "anticipating" dog version won the vote.

Cats

I presented 3 ideas that I reviewed with the design team and then asked users to vote.

3. Icon Set and Style

Standardized the Icon Set and created a Design Ritual and Guidance Documentation for team to create new icons when needed.

4. Accessibility and WCAG Guidance

Established a Design Ritual around Accessibility Best Practices for design and engineering including testing
Worked with a teammate on Accessibility Documentation in Zeroheight. My contribution was the Perceivable, Operable, and Understandable pages.

5. Design System (Contributing)

Design Philosophy and Strategy: As part of the design team I helped develop our visual and experience philosophy as Clear, Concise, Respectful, and Human.

Our 5 principles guiding the Voyager Health User Experience:

Anticipate & Augment User Needs
Foster Trust & Peace of Mind
Support Individual Working Styles
Seamless Continuity & Hand Off
Optimize Task Prep & Information Access

A selection of the Voyager Health component library elements built solely by me as the foundational UI building blocks for the platform. Color and typography were established collaboratively as a team.

Localization and Globalization: With a focus on WCAG and ADA Guidelines and cultural/language support in 40 International markets. We regularly engaged in design rituals with cross-discipline and user feedback reviews to iterate on the Visual Brand and Voice. Our design strategy included adhering to local legal frameworks and agreements; e.g. GDPR, CCPA, CPRA, EFTA, WCAG, and State Boards of Veterinary Medicine.

Design Mechanics: The Design System was based on Material Design, constructed using the Atomic Design Model. Every designer on the team contributed components to the system.

What Shipped

After four months of onsite research, iterative design, and cross-functional collaboration, the Operations Dashboard shipped to production. The results validated the core design decisions: particularly the single unified view over role-specific dashboards and surfacing the patient journey and appointment KPIs directly from the dashboard.

+5%
increase of proactive patient bookings from appointment/actual target rate
5x
faster success rate in identifying patient journey status and physical location
+32%
success rate in optimizing room usage and prepping rooms
+21%
increase in SLA patient follow ups and task completion