A practice is told its collection rate is in the fifties against an industry benchmark of 98%. On a $900,000 book that reads as roughly $516,000 walking out the door every year. It is the most common finding in practice consulting, and it is usually an arithmetic error.
The two numbers are not the same measurement. One is a gross collection rate. The other is a net collection rate. Comparing them manufactures a gap that was never collectible by anyone.
The two rates, plainly
| Rate | Formula | What it tells you |
|---|---|---|
| Gross collection rate | Payments ÷ charges | How your posted fee schedule compares to what payers actually allow. Mostly a statement about your pricing. |
| Net collection rate | Payments ÷ (charges − contractual adjustments) | Of the money you were ever entitled to collect, how much you actually got. This is the performance measure. |
The 95–98% benchmark everyone quotes is a net figure. Your gross rate will be far lower — and should be. If you post $200 for a service Blue Cross has agreed to allow $95 for, the $105 difference is not lost revenue. It is not revenue at all. It is the contract you signed.
If a report shows a collection rate under about 70% and compares it to a benchmark in the nineties, ask one question: "Is the denominator net of contractual adjustments?" If the answer is no, or if nobody knows, the finding is not a finding.
How big is the contractual piece? Usually enormous.
In one pilot engagement we followed 2,099 claim lines totalling $144,213 billed to insurance since a system change, and tracked every one to where the dollar actually landed. Not a period-over-period snapshot — a closed cohort, followed forward.
| Where the billed dollar went | Amount | Share |
|---|---|---|
| Contractually written off — never collectible | $48,376.10 | 33.5% |
| Paid by the insurer | $38,873.24 | 27.0% |
| Still open, awaiting payment | $45,428.07 | 31.5% |
| Moved to patient responsibility | $11,481.17 | 8.0% |
A third of everything billed was written off by contract before anyone could collect it. That $48,376 was never available. No amount of follow-up recovers a cent of it. Counting it as lost revenue is the single most common error in practice analysis — and it is the error that makes a healthy practice look like it is haemorrhaging.
Note also that 31.5% was still open, at a median age of 59 days. On a gross-rate snapshot taken today, all of that reads as failure too. Most of it is just a claim that hasn't come back yet.
Not every adjustment is contractual
This is where the analysis gets real. Practice-management systems bundle very different things under "adjustments," and separating them is most of the work.
| Adjustment type | What it actually is | Belongs in net rate? |
|---|---|---|
| CO-45 — charge exceeds fee schedule | The contractual write-down. The gap between your posted fee and the payer's allowed amount. | Yes — remove from the denominator |
| Patient-side write-offs | Could be a time-of-service discount, could be bad debt. These are opposite things wearing the same code. | Depends — see below |
| Migration artifacts | Debris from a system conversion. Not economic events at all. | No — exclude entirely |
In the same engagement, adjustments over three months came to $144,696.89, of which $89,977.25 was CO-45 contractual. The remainder had to be taken apart by hand.
Discount or bad debt? Three tests
One patient-side write-off code carried $41,855. Three independent tests said what it was:
- Timing. Median age at write-off: one day. Forty-five percent same-day. Bad debt does not get written off the day it is incurred.
- Population. 96% of the dollars sat on patients with no insurance adjustment at all — i.e. self-pay.
- Amounts. Eighty percent were exact multiples of $5. Bad debt is not a round number.
Verdict: $32,808 was a self-pay discount — about $10,936 a month of deliberate pricing policy — and only $5,171 was genuine bad debt, roughly $1,724 a month. Those two numbers lead to opposite conversations. One is "review your cash-pay pricing." The other is "tighten collections." Bundled together they are simply a wrong number.
Before accepting any collection-rate finding, get three things: the formula (is the denominator net of contractuals?), the measurement date (a rate measured the day a period closes will be understated — measure a cohort, not a window), and the adjustment breakdown by reason code. If a report cannot produce all three, it cannot support the conclusion it is drawing.
The honest method
Period-based rates divide one month's payments by the same month's charges. Those payments are for claims billed months earlier. It is a ratio of two unrelated quantities, and it will bounce around for reasons that have nothing to do with performance.
The only method that survives scrutiny is the cohort: take a defined set of claims and follow them forward until each one has resolved — paid, adjusted, denied, or moved to the patient. It is slower. It is also the only version you can defend when a doctor asks where the number came from.