This is the year-over-year result from the founder's practice, an independent outpatient imaging practice in Southern California. It ran a patient-billing vendor through December 2025 and has run DueWell since. Both ledgers were available in raw form, so the comparison below uses the same definitions on both sides.
Single site, founder-affiliated, observational, no control group. Read it as one practice's measurement, not as a forecast for yours.
The comparison window
Mar–Aug 2025 on the prior vendor against Mar–Aug 2026 on the platform. Same practice, same calendar months, a year apart. The window skips the prior vendor's wind-down (November–December 2025) and the platform's launch backlog (December 2025–February 2026), when neither system was running normally.
Cash and paying patients
| Monthly average | Mar–Aug 2025 | Mar–Aug 2026 | Change |
|---|---|---|---|
| Patient cash collected | $80,375 | $143,189 | +78% |
| Distinct paying patients | 552 | 849 | +54% |
| Cash events | 738 | 1,148 | +56% |
| Average dollars per cash event | $109 | $125 | +15% |
Over the eight months January through August, the monthly average went from $78,803 to $132,904, up 69 percent.
One caveat belongs next to these numbers. Collections data alone cannot separate the increase into the platform's effect, placement volume, and deductible mix. The safe statement is "same practice, same months," and it stops there.
Recovery by age at first outreach
The number that matters more than total cash is what happens to balances that were already old when the first message went out. These are the dollars a practice would otherwise send to an agency.
| Age when outreach began | Recovered | Window |
|---|---|---|
| 61–90 days | 62.8% | within 90 days |
| 91–120 days | 46.9% | within 90 days |
| 121–180 days | 30.8% | within 180 days |
| More than 180 days | 10.5% | within 180 days |
For context, ACA International's benchmarking data, as reproduced in trade sources, puts agency recovery at roughly 17 to 21 percent of placed medical debt over the life of a placement. The 180-plus bucket is shown because it is below that benchmark. The platform's advantage is in the 60-to-180-day window; balances older than that are hard for everyone.
Against the prior vendor, same rules
The prior vendor's raw account export was re-run under the platform's method: account-level, amount-weighted, right-censored, same age buckets. Its published baseline ($233,471 collected from 1,744 paying accounts over its last 123 days) was reproduced exactly from its own ledger, so both sides of this comparison are primary-data-verified.
The prior vendor's ledger includes non-cash "adjusted" entries that may or may not represent real patient payments. The figures below count them as collections, which is the basis most favorable to the prior vendor.
| Age bucket | Prior vendor, favorable basis | Platform | Window |
|---|---|---|---|
| 61–90 days | 47.6% | 62.8% | 90 days |
| 91–120 days | 26.8% | 46.9% | 90 days |
| 121–180 days | 30.9% | 29.3% | 90 days |
The 121–180 bucket is a mixed result and is shown at full weight. On that basis the prior vendor edges the platform at 90 days (30.9 versus 29.3 percent) and at 180 days (37.0 versus 30.8 percent). On card cash alone the platform leads in that bucket too, but the honest summary is that the platform out-collected the prior vendor in five of six buckets on the favorable basis and six of six on card cash, and that one bucket is a wash.
The two systems' cohort anchors also differ slightly. The prior vendor's clock starts at placement, which triggered a same-day text; the platform's starts at the later of first successful outreach or date of service. Functionally close, not identical.
The clock that got slower
Every time-to-cash measurement from date of service is longer on the platform. Median days from service to first cash went from 30 to 40. The dollar-weighted mean went from 50 to 65 days.
That is not a lapse. It is what recovering old balances looks like in the data. The prior vendor received accounts about 25 days after service, texted the same day, and was paid overwhelmingly by fresh balances. It never recovered the old ones. The platform imports every open patient balance weekly, including balances six to twenty-four months old, and collects some of them: 771 payment events on balances more than 180 days old in the recovery data. Each of those is cash the practice had written off in practice, and each one lengthens the service-date clock.
So: do not read this as faster time to cash, and do not compute a days-sales-outstanding figure from it. A platform whose job is aged dollars will always show a longer service-to-cash median than a vendor that only touched fresh ones.
The clock that isolates the platform's own behavior is days from first message to cash. That median is 15 days, and it has held between 14 and 17 in each of the last seven months.
Method notes
Amount-weighted throughout: dollars collected divided by dollars placed, not account counts. Right-censored: each observation window counts only accounts old enough to be fully observed within it, so the 90-, 120-, and 180-day columns have independent denominators and are not one cohort's cumulative curve. Each bucket is reported in its own window.
A cash event is defined the same way on both ledgers: a successful patient payment dated the day the card was charged, each transaction counted once. The prior vendor's non-cash adjustments are excluded from cash totals and shown separately in the bucket comparison as its favorable basis. The recount of the prior vendor's ledger reconciles exactly to its published lifetime card cash of $1,135,069.
One correction for the record: the first version of the year-over-year analysis reported cash flat and days down, because the script counted each prior-vendor transaction once per line item and included the non-cash entries. After de-duplication the totals reconciled and the results above are the corrected ones.
Sources
- DueWell analysis of the founder's practice ledger, Sept 2026 (platform production ledger, read-only extraction, data as of September 20, 2026; prior vendor's raw account export through December 2025)
- ACA International benchmarking data on agency recovery, as reproduced in industry trade references