Aspen Dental: Patient Financing Work Queue

01Discovery

The moment I knew something was broken

During a visit to an Aspen Dental office, before this project was assigned, I asked the Operations Manager about a dashboard widget called “Schedule Care or Refund.”

He didn’t touch it. Not because he didn’t need to, but because he had no idea what it was asking him to do.

Weeks later, leadership in Patient Financing and Revenue Cycle Management (RCM) brought me the same problem. I already knew why.

FIG 01The original “Schedule Care or Refund” widget, lost among the dashboard’s other widgets.

Operations Managers (OMs) run each Aspen Dental location, so their day rarely slows down. Many are also new to the job.

40%OM turnover, the highest in three years. The workflow had to make sense to someone in their first week.

02The problem

A system built for experts, used by everyone

When a patient takes out financing, a clock starts. There are three phases, each with its own rules and deadlines.

Miss a deadline in any phase and the lender takes the money back, whether or not anyone meant it to.

Trigger-based refund logic

Phase 1Non-starterBook the first appointment within 30 days of the payment date, or provide $100+ of care.

Phase 2Work in progressProvide at least $100 of care by a set date, or the refund triggers.

Phase 3Care completedSchedule the final appointment before the last date. It must fall within 30 days of that date.

FIG 02The three phases I mapped in FigJam. I returned to this map in every stakeholder and engineering session.

The 10-column queue built on top of it had no call to action:

“Refund date” sounded final, but shifted with OM actions.
“Care completed” meant Phase 3 of financing, not finished treatment.
No care totals, so scheduled patients stayed on the list with no explanation.

FIG 03The original 10-column queue. No column told an OM what to do next.

03Key decision

From a data table to a tool for action

I facilitated a workshop with Patient Financing, its executives, and the territory managers who train OMs. We asked of every column: does an OM need this to do their job? We cut 10 columns to 5.

FIG 04The workshop board: every column kept, cut, or questioned.

Cutting alone still didn’t tell an OM what to do, so I proposed a sixth: “Action Needed,” the exact next step to keep each patient’s financing on track, based on their phase and lender.

With RCM, I mapped every phase and lender combination to its instruction, then validated the state table with our PM and engineers so the instruction updates as a patient’s phase changes.

I also moved “Days Until Scheduled Refund” first, so OMs work the list top to bottom by urgency.

FIG 05My first sketch of the “Action Needed” instructions for each phase.

Two ideas didn’t make it:

ExploredOne date instead of twoI asked engineering whether a single date could cover scheduling and completion. The lender logic required both, so the language had to explain two dates clearly.

ExploredA queue per lenderI considered splitting the queue by lender brand, like CareCredit. The actions OMs take were nearly identical across lenders, so one queue kept the page quieter.

FIG 06Before: 10 columns. “Refund type” was industry jargon, and no one used the filters.After: sorted by urgency, with an “Action Needed” instruction per patient.

04Entry point

Redesigning the front door

The dashboard widget leading into the queue had the same problem as the table: it put a red dollar figure up front, but not a next step. This was the widget the OM from my office visit never touched.

I replaced the red "Opportunity at risk" figure with a plain count, "6 accounts left to action," and a "2 refunding soon" badge that appears only in the 0–7 day window. I also renamed it from "Schedule Care or Refund" to "Upcoming Patient Financing Refunds," so it couldn't be mistaken for the dashboard's other scheduling widgets.

Round two would prove me half wrong about this.

Before

FIG 07Before: the workload took a click to see.

After

FIG 08After: the task count, visible at a glance.

05Usability testing

Testing the language, not just the layout

In moderated sessions with 7 Operations and Regional Managers (3 months to 14 years of experience), 5 read the original date copy as one deadline instead of two.

I invited the SVP of Patient Financing and our product team to observe, so they'd see OMs' reactions firsthand.

“It feels like I have to finish everything by this one date. I don't get why there are two dates here.”

Operations Manager · 14 YOE

The simplest rewrite worked for all 7. Jargon mattered too: “charge out” confused people, while “provide over $100 worth of care” landed at every experience level.

BeforeSchedule by Oct 14 to occur/be completed by Nov 28

FIG 09Before: 5 of 7 read this as one deadline.

AfterSchedule by Oct 14, complete appt by Nov 28

FIG 10After: clear to all 7.

06Results

What changed in three months

After a 50-office pilot, we launched to all 1,100+ locations, then compared refund dollars in the three months before and after rollout. No other process, staffing, or policy changes happened in that window.

“This makes it way easier to keep track of patients who will be refunded and when I should schedule them.”

OPERATIONS MANAGER · LAUNCH SURVEY

REFUND REDUCTION$2.98MLess refunded to lenders after rollout.

Refund rate−2.5 pts24.3% → 21.8%, a ~10% relative drop.

07Round two

I disagreed with the SVP, so I tested it

Months later, leadership asked for in-dashboard refund requests and a view of already-scheduled appointments. I disagreed with the second ask: the queue exists to show outstanding work, and a scheduled patient is finished work. Rather than argue from my own assumption, I designed both and tested them with 5 OMs.

FIG 11The interim design I tested: patients to schedule, refunds in progress, and scheduled appointments. Only the first two shipped.

KeptSectioned layoutSplitting the queue into patients to schedule and refunds in progress was a clear win with all 5.

Removed“Treatment timeline” badgeIt read “Not started” next to scheduled patients.

RestoredDollar amount3 of 5 said the widget felt less urgent without it.

ExtendedRefunded-patient visibility7 days instead of 24 hours, because OMs check the queue about twice a week.

“Why does it say ‘Not started’ if I already scheduled them? That’s confusing.”

OPERATIONS MANAGER · 2 YOE

No OM wanted scheduled patients on the list. I took that to the SVP of Patient Financing. He'd sat in on the first round of sessions, so the case was short: five OMs, none of whom wanted the view he'd asked for. He dropped the ask. The same sessions proved me half wrong on the widget: the dollar figure I’d cut came back.

Refund requests

Making irreversible actions deliberate

Refunds used to need a separate ServiceNow ticket. I moved them into the queue's status dropdown alongside outreach tracking, with a confirmation step so nothing irreversible is submitted by accident.

FIG 12The status dropdown: outreach tracking and refund requests in one place.

FIG 13Submitting stays locked until the OM checks the confirmation.

08What’s next

Carrying the work into a new design system

In development · not yet launched

Aspen is adopting a new design system, and I redesigned the full Operations Dashboard for it, across every role tab. Each task widget now names the report card metric it moves, so OMs see why a task matters, not only that it’s due. After launch, I’ll track task completion rates alongside the refund rate.

FIG 14The financing widget in context on the redesigned dashboard.

FIG 15Made to read at a glance in a busy office: one number, one progress bar, and an urgency badge only when refunds are close. The bar is a component I designed for every task widget.

FIG 16More breathing room, larger type, and counted sections. The workflow stays the same on purpose: it’s what moved the refund rate.

09Reflection

Listen earlier, and keep listening

Next timeGo deeper on the current stateI’d talk to more OMs before designing. Round two showed the dollar figure was what made the widget feel urgent, and that OMs check the queue twice a week; two more interviews up front would have caught both.

PrincipleDesign for the first weekWith 40% turnover, I judged every label by whether the newest OM could act on it. The copy that worked for someone 3 months in also worked at 14 years.

A work queue that cut $2.98M in patient financing refunds.