The check an owner used to make by being present is a reconciliation, and it can be written down. Every dollar should match three ways: the chair against the claim, the claim against the payment, and the payment against the deposit.
Prosperident, the largest specialist investigator of dental‑practice embezzlement in North America, estimates that about 60% of practices are embezzled at some point in their life, that the average case runs two years before discovery, and that the average loss is around $109,000. Dental is over‑targeted for structural reasons: one person often handles the schedule, the ledger and the deposit; cash and card payments arrive at the same desk; adjustments are routine; and the owner is in a treatment room.
At one location the owner's presence was a control — an imperfect one, but a real one. Somewhere between the second and fifth location it stops existing, and nothing replaces it unless it is written down.
Chair against claim. Was the procedure billed the procedure that was done? The evidence is the clinical record: the chart note, the radiograph, the treatment plan. A claim for a composite on a tooth whose most recent radiograph shows no decay, created eleven days after the visit and back‑dated, is the pattern investigators call a phantom procedure. Industry estimates put phantom billing at 1–3% of claim volume.
Claim against payment. Was what the carrier paid what the claim asked for, adjusted per the contract? This is where contract variance lives — but it is also where write‑off concealment lives: a patient pays cash, the payment never reaches the ledger, and the balance is quietly adjusted away. The tell is the same person collecting and adjusting.
Payment against deposit. Did what was posted as received arrive in the bank, on the day it was posted, in the amount posted? This is the oldest check and still the one most often skipped. The scheme it catches is lapping — today's payment from one patient applied to yesterday's shortfall on another, rolling forward until a holiday breaks the chain.
An owner can do this once a month in an hour. Take the day sheet for five random days. For each, tie total posted payments to the bank deposit for that day (not the week — daily). Pull every adjustment over a threshold and ask what it was for. Pick ten claims and open the clinical record behind each. Do it on days chosen by you, not announced, and do it yourself or with your CPA. Done monthly, this catches the large, slow schemes: the write‑off that was really a pocketed payment, the vendor that hasn't existed for two years.
It will not catch the small, fast ones. The average case runs two years for exactly that reason — five random days a month is a 15% sample, and a careful thief is careful on the days that get checked.
The same three matches run on every transaction, every day, are a different instrument. Nothing is sampled. A payment posted with no matching deposit, an adjustment by the same user who took the payment, a procedure with no radiograph behind it, a bank‑detail change on a vendor inside a payroll window — each surfaces the day it happens, not two years later. That is the difference between a check and a control.
Most mismatches are errors. A deposit made a day late; a payment posted to the wrong family member; an adjustment that was a legitimate courtesy. So the language matters as much as the finding. Use graded words — observation, caution, concern, finding — and never a word like theft, fraud or guilty, which you cannot support and which will be quoted back to you. Route anything above observation to the owner and the people they designate, never to a staff account. And keep the chain intact: what was checked, against what, when, by whom, unchanged since. A case file is worth more than a conclusion — to you, to your CPA, to an insurer and, if it ever comes to that, to a court.
You cannot separate everything with four people, so separate the three that matter most. The person who posts payments does not make adjustments. The person who makes the deposit does not reconcile the bank. And nobody changes a vendor's or an employee's bank details in the week before a payroll run without a second person's sign‑off — that one rule closes the single most expensive scheme in the ledger.
This month, pick five days at random. Tie each day's posted payments to that day's deposit. List every adjustment over $50 and ask what it was for. Open the clinical record behind ten claims. Write down what you couldn't tie out.
Fraud Guard runs the three matches on every transaction, every day, at every location, and delivers a graded case file — routed to the owner and the people they designate, never to a staff account. How Fraud Guard works →
Back to the curriculum — pick the seat you're sitting in and read five, in order.
Canopy University →Ask her. She answers from the lessons here, and says so when the honest answer is that it isn't written yet.
Ask about this lesson →
Every lesson here is written for the person doing the work, whether or not they ever buy from us. Inside Vella OS the same questions get answered from your own record.
The scalable operating system that actually improves financial performance on its own.