Guides
NEMT Business Plan: Building a Checkable Assumption Register
Build an NEMT business plan around a checkable assumption register for trip demand, payer timing, vehicle capacity, and signals that should trigger a rewrite.
What to take away
- A plan for this trade is an assumption register, not a narrative. Every number in it came from somewhere, and the plan should say where.
- The assumptions that break first are load time, cancellation rate and days to payment. Those three decide whether the rest of the arithmetic means anything.
- Write the falsifier beside each assumptionthe observation that would tell you the number is wrong, and how soon you would see it.
- Payer mix belongs in the plan as a cash timing statement, not as a revenue statement. Who pays and how fast are different questions from how much.
- Nothing here supplies your figures. Rates, coverage terms and payment cycles come from your state Medicaid agency, its transportation broker, your contracts and your own records.
Most plans for a transport operation fail in the same place. They are confident about revenue and vague about time. Revenue is easy to state and impossible to verify on day one. Time is measurable in a week, and every revenue figure in the document depends on it.
So build the plan around what you are assuming, and make each assumption checkable.
The register
One row per assumption. Four columns, and the fourth is the one that does the work.
Assumption Register Falsifiers
- Load timedrivers late from second pickup
- Trips per vehicle hourratio drifting down
- Cancellationsempty slots not backfilled
- Empty milesfuel rising faster than trips
- Days to paymentpayables aging faster
- Driver turnoversecond replacement in quarter
| Assumption | Where the number comes from | What would show it is wrong | When you would see it |
|---|---|---|---|
| Minutes from arrival to secured and moving | Your own timed dry runs, by mode | Drivers running late from the second pickup onward | First week |
| Trips completed per staffed vehicle hour | Daily count of trips against hours clocked | The ratio drifting down while activity feels busier | First month |
| Share of booked trips that cancel or no-show | Your own dispatch log, by account and by day | Empty slots you cannot backfill in time | First month |
| Empty minutes between a drop and the next pickup | Route measurement, not software estimates | Fuel and hours rising faster than trip counts | First month |
| Days from trip completed to money received | Remittances, tracked by payer | Payables aging faster than receivables | First quarter |
| Driver turnover | Your own hiring record | A second replacement hire in the same quarter | First half year |
Fill this in before the financial pages, which are arithmetic on these rows. A plan that has a revenue projection and no register has a conclusion with no premises.
One row filled in, as an example to replace with your own data. Assumption: load time in minutes. Source: four timed dry runs on the dialysis account in March. Falsifier: a median above 12 minutes twice in one week. Window: the first week of April.
Example ranges, to replace with your own data.
- Load time, the minutes from arrival to a secured passenger and a moving vehicle, commonly falls between 8 and 20 minutes in a dense urban market and runs longer in rural service.
- Cancellation and no-show share of booked trips often lands between 5 and 15 percent, clustered by account and weekday.
- Days to payment depends on the payer. Private pay at the curb settles on the day. A facility contract may run about 30 days, and a state broker or Medicaid managed care plan can run 30 to 90 days from a clean claim.
Describe the demand as trip patterns
A lender or a broker reading your plan wants to know what the week looks like, not what the market is worth.
Trip Patterns Behave Differently
Pattern
- Standing orders
- Fixed days
- One-off scheduled
- Booked ahead
- Will-call returns
- Unknown discharge
- Long-distance transfers
- Planned
Scheduling
- Standing orders
- One late pickup costly
- One-off scheduled
- Slot backfill
- Will-call returns
- Vehicle tied up
- Long-distance transfers
- Vehicle hours
Risk
- Standing orders
- One-off scheduled
- Will-call returns
- Long-distance transfers
Say which patterns you will serve and roughly in what proportion: standing orders that repeat on fixed days, one-off scheduled appointments, will-call returns with unknown discharge times, and long-distance transfers. Each behaves differently.
Standing orders make a schedule predictable but make one late pickup expensive, because the clinic chair behind it is booked. Will-call returns tie up a vehicle for an unknown period and can make an otherwise sensible day fall apart.
The public side of the demand is described in general terms by the Centers for Medicare and Medicaid Services transportation benefit page, which is worth citing as context and worthless as a source of local numbers. Local numbers come from local schedulers.
Separate the three parties in the payment section
Plans that treat the customer as one entity are the ones that later discover a receivables problem.
Three Parties, Three Payers
Payer type
- Facility direct
- Facility
- State broker
- Broker
- Family at curb
- Family
Booking party
- Facility direct
- Direct billing
- State broker
- Contract cycle
- Family at curb
- Immediate
Payment cycle
- Facility direct
- State broker
- Family at curb
Name the booking party, the riding party, and the paying party separately for each account. State the payment cycle for each payer.
A facility buying transport directly differs from a broker administering a state contract, which differs from a family paying at the curb. The slowest payer sets your working capital requirement, not the average.
The finance and dispatch systems section of your plan should say how you will track that, because a payer mix is unmanageable from memory.
Suppose 60 percent of monthly billing comes from a broker paying in 45 days, 30 percent from facilities paying in 30 days, and 10 percent private pay at the curb. On 40,000 dollars of monthly billed revenue, roughly 24,000 dollars is still outstanding when the next month's costs arrive.
Those are example figures, to replace with your own. That gap sets the working capital figure the plan has to state.
Where to find the terms. The state Medicaid agency's provider page carries the NEMT fee schedule or rate methodology. The brokerage named in the state contract publishes a provider manual with the authorization and documentation rules. Ask the broker for that manual in writing, and check your own contracts and remittances for the rates that apply to you.
Capacity: vehicles, hours and the single point of failure
State how many vehicles, of which type, staffed for how many hours, and what that yields at your assumed load times. Then state what happens when one vehicle is off the road.
One Vehicle Off the Road
Does one breakdown cancel a morning of dialysis runs?
state contingency and single point of failure
state spare capacity and cover arrangement
This is where a plan either admits a single point of failure or hides it. If one breakdown cancels a morning of dialysis runs, say so and say what the contingency is. This section is where the fleet and premises decisions belong, stated as a capital decision with an operating consequence, not a shopping list.
What to leave out
Leave out market size figures you cannot source. Leave out a five-year revenue curve, which nobody believes and which distracts from the register. Leave out any statement about coverage rules, reimbursement or licensing stated as fact, because your reader may know better than you do and one wrong assertion discredits the document.
Keep the ordinary formation and tax material short and point at the sources. The IRS material on starting a business and the Small Business Administration's planning pages cover it in the standard way, and a plan that reproduces them is padding.
Review triggers, not review dates
Give the plan a short section that says what forces a rewrite rather than when the next version is due.
Reasonable triggers: losing or winning an account that is more than a small share of your trips, or a change to a payer's authorization or documentation process. Also a second vehicle, a change in your service radius, or any of the register's falsifiers actually firing.
The approach to opening the first weeks and the owner's overview of the trade both assume this plan is a working document rather than a filed one. So does the material on growing beyond one service area.
Common questions
How long should the plan be?
Short enough that you reread it. If a section does not change a decision, cut it.
Do I need a plan if I am not borrowing?
The register, yes. The document, only if somebody is going to read it. The value is in writing down what you are assuming while you can still remember that you assumed it.
What if I have no data for the register at all?
Then the first column is filled and the second says "not yet measured", which is honest and useful. Measure the load times first, because everything else scales off them.
Should I include competitor analysis?
Include what you learned by asking schedulers what they currently do. Skip anything you inferred from a website, because a competitor's public material tells you what they sell and nothing about whether they deliver it on time.
