The six numbers this guide walks through, in the order money moves. Four carry a published professional target, one is a repeated rule of thumb, and one has no free small-practice figure at all — but not one of the six rests on a dataset you can open.
So I'm going to do the thing the vendor listicles don't. For each number, I'll tell you exactly who says it, whether there's a dataset behind it or just a professional target — and, in three cases, that the number everyone repeats has no traceable source at all.
The pipeline, in one look
Every one of the six sits at a joint in the same pipe. Money is charged, a claim goes out, it's either clean or it's denied, the balance ages in accounts receivable, and eventually some fraction of it is collected. Each KPI is a pressure gauge on one joint:
- At the claim: clean claim rate and denial rate — did the claim go through, or bounce?
- In A/R: days in A/R and A/R over 90 days — how long is the money stuck, and how much is going stale?
- End to end: net collection rate — of everything you were owed, how much did you actually get? — and cost to collect: what did running the pipe cost you?
You don't fix "revenue." You fix the joint that's leaking, and these six tell you which one.
The six KPIs
A professional target, not a measured dataset — AAFP publishes no survey behind it.
Of the money you were actually allowed to collect — after contractual write-offs — how much did you get? This is the end-of-pipe number, and the most important of the six, because it's the only one that tells you whether the work you did turned into money.
Ignore gross collection rate (payments ÷ total charges); it mostly measures how inflated your fee schedule is. Net is the one that matters.
Below 50 days at minimum. The one measured reference — MGMA’s top performers — sits behind a paid product.
How long, on average, your money sits unpaid — total receivables divided by average daily charges. Collection rate tells you whether you get paid; this tells you when. A practice can collect 97% of what it's owed and still have a cash-flow problem if it takes 60 days to do it.
A widely-repeated rule of thumb, not a published figure. Aim at it as a ceiling, not as a measured norm.
The share of your receivables that has aged past 90 days. Days in A/R is an average, and averages hide tails — a healthy average with a fat 90-plus bucket means a pile of old claims quietly turning into write-offs. The older a balance gets, the less likely it is ever to be collected.
Not the "95%, per HFMA" you will be quoted: HFMA says 98%, citing a trade article, and its MAP Keys publish no value at all.
The percentage of claims that pass every edit with no manual intervention, measured in your billing system before submission. HFMA’s MAP Keys define it that way; payer acceptance on first submission is a different metric, first-pass resolution. It's the cheapest lever in the whole cycle: a claim that goes through clean costs you almost nothing; one that bounces has to be found, fixed and resubmitted.
AAFP calls 5–10% the industry average and below 5% more desirable — a target, not a dataset.
The share of claims a payer rejects on first pass. Two things to keep straight, because the industry blurs them: an initial denial is the first rejection (often fixable and resubmittable), while a final write-off is revenue you gave up after exhausting appeals. HFMA's Claim Integrity Task Force put standard definitions to both — worth using, because a "denial rate" with no definition is uncomparable.
The quoted ~2% is a hospital figure out of HFMA’s paid MAP data; independents almost always run higher.
What it costs you to turn a dollar of care into a dollar in the bank — total billing and collections cost divided by total cash collected. It's the KPI that tells you whether chasing the last few points of collection rate is actually worth it.
The numbers you'll be quoted that aren't real
Here's where this piece earns its keep. Three figures appear on nearly every revenue-cycle page, always as settled fact, and all three fall apart the moment you look for the source.
claim rate,
per HFMA"
rework a
denied claim"
denials are
never reworked"
I'm not saying denials are cheap or rare — they're neither. I'm saying the specific figures used to scare you into a purchase were, as far as I can trace them, invented or borrowed from a vendor and passed hand to hand until they read like fact.
The distinction that runs through all of it
By now you've seen me use two different words for two different kinds of number, and the whole art of reading this space is telling them apart:
- A target is what a body like the AAFP or HFMA says you should aim for. It's a judgement, openly labelled as one — "best-practice tips," in AAFP's words. The AAFP goes further in places, calling 5-10% denials the industry average without publishing a dataset behind it, which is the sharper complaint. Aim at it; be held to it.
- A measurement is a number someone actually observed in the world — a survey, a claims dataset. It has a sample, a date, and a scope.
Almost every "benchmark" in revenue cycle is the first kind wearing the clothes of the second. AAFP's 95% and HFMA's 98% are targets. The real measurements that exist — the ones below — are nearly all hospitals or payers, not practices like yours. When a number can't tell you its sample size and its date, it isn't a benchmark. It's a slogan.
What the real denial datasets actually say
Three real denial figures — and the one that is missing
One number from the KFF data is worth carrying, with its scope attached: fewer than 1% of those denials were appealed at all — and that's consumer appeals, not the separate channels a practice's billing team uses, so it says nothing about how often you should appeal. When consumers did appeal, insurers upheld about two-thirds of the time. The honest lesson isn't that denials are easy to reverse; it's that the leverage sits upstream. The cheapest denial is the one that never happens — which is why clean claim rate, not appeal-win rate, is the number to chase.
How this fits together
These six are the money-and-billing core of the wider picture. Collection rate and days in A/R each have a full guide of their own; the rest live here, in the hub, because on their own they're a paragraph, not a page. If you want the whole practice on one screen rather than just the billing pipe, that's the five numbers to check every Monday and the broader list of practice KPIs. Dentistry runs on a genuinely different revenue-cycle scorecard — case acceptance, collections as a percent of adjusted production — which is why it gets its own set.
And wherever you build these, the principle from the fake-benchmark section holds: compare each number to your own trend first, to a named professional target second, and to a "measured benchmark" only once you've checked whose hospital it was measured in.
Frequently asked questions
Six cover the pipeline from charge to cash: net (adjusted) collection rate (how much of what you were owed you collected), days in A/R (how long the money takes), A/R over 90 days (how much is going stale), clean claim rate (the share that passes your own edits before it is sent), claim denial rate (the share rejected initially), and cost to collect (what running billing costs you). The first two have deep guides of their own; together the six show where in the cycle money is leaking.
Be careful — most numbers you'll be quoted are professional targets or hospital figures, not measured small-practice data. The AAFP states, as best-practice targets, an adjusted collection rate of at least 95% (top performers 99%+), days in A/R below 50 with 30–40 preferable, and a denial rate of 5–10% (below 5% desirable). A 98% clean-claim target circulates under HFMA's name, but the page carrying it is marked sponsored content, paid for by a revenue-cycle outsourcing vendor, and the figure inside comes from a Becker's trade article about ambulatory surgery centres. These are targets, not datasets, and independents are widely said to run clean claim rates of 75–85% — itself an unsourced figure. Compare to your own trend first.
Gross collection rate is payments ÷ total charges, and it mostly reflects how inflated your fee schedule is above contracted rates — close to meaningless as a performance measure. Net (adjusted) collection rate divides payments by the amount you were actually allowed to collect after contractual write-offs, so it measures how much of your collectible revenue you truly captured. Net is the one that matters; the AAFP's 95%/99% target is the adjusted (net) rate.
No. The widely repeated "95% clean claim rate, per HFMA" doesn't hold up. the 98% circulating under HFMA's name sits on a page marked sponsored content, paid for by a revenue-cycle outsourcing vendor, and cites a Becker's article about ambulatory surgery centres, which is not a stronger source than the 95% it would replace, and its formal MAP Keys framework defines the clean claim rate but publishes no free benchmark value — the comparative data sits behind a paid product. "95%" is a vendor rule of thumb with no dataset — as is the 75–85% independents are said to run.
The figure you'll see everywhere — about $25 per reworked claim ($25.20) — is usually quoted as MGMA's, but in MGMA's own article it appears with no citation at all, and it's dated. A more recent hospital survey by Premier, surveyed between August 2024 and February 2025 about claims from calendar year 2023, put the administrative cost of fighting a single denial at $57.23, up from $43.84 the year before — a related but distinct number. The claim that "65% of denied claims are never resubmitted" has no traceable primary source at all. Treat all three as industry lore, not measured facts.
WRITTEN BY
Olha · clinic data analyst
I build the reporting our managers open every morning at a multi-branch medical clinic — and package it so other practices don't have to start from scratch.
Every benchmark here is labelled by what it actually is: a professional target (AAFP, or a vendor-sponsored page carrying HFMA branding), a measured dataset with its scope stated (KFF, Premier, Change Healthcare, MGMA), or untraceable industry lore — and where a figure everyone repeats has no source, I've said so rather than pass it along. The the AAFP figures are stated targets, not measured distributions; the denial datasets are hospital- or payer-scale, not small-practice censuses. Lucid Vitals is not affiliated with HFMA, AAFP, MGMA or Microsoft.