Your Amazon scorecard just flipped from "Fantastic" to "Fair," and nobody told your bank account — until the settlement lands $12,000 lighter than last week. Multiply that by every driver call-out, every damaged package chargeback, and every van repair bill, and you start to understand why so many Delivery Service Partner (DSP) owners run a $3 million operation and still can't tell you, with confidence, whether they made money last month.
Amazon's DSP program has grown into one of the largest small-business franchise systems in the country — thousands of independent owners dispatching fleets of branded vans, each one a real business with real payroll, real insurance, and a revenue stream that resets every single week. That volume and velocity is exactly what makes DSP bookkeeping so unforgiving. Miss a pattern for one pay period and it's a rounding error. Miss it for a quarter and it's the difference between a profitable business and a slow-motion collapse.
Here's what the accounting actually looks like once you get past the marketing deck.
Why a "Fantastic" Week Isn't Revenue Yet
Every Amazon DSP gets paid through a weekly settlement statement, not a simple invoice. That statement bundles together several very different things:
- Base route rates — a flat amount per route dispatched, typically in the $180–$280 range
- Per-package fees — paid per stop or per package, often $15–$40 per route per day, higher during peak season
- Scorecard incentive bonuses — anywhere from roughly $2,000 to $8,000 a week for DSPs rated "Fantastic," scaling down sharply for lower tiers
- Chargebacks and deductions — typically 1–4% of gross revenue, for damaged packages, missing items, late deliveries, and customer refunds
The single biggest bookkeeping mistake DSP owners make is depositing that net settlement number straight into a "Revenue" account and calling it a day. Doing that buries the chargeback trend inside a number that looks fine on the surface. A DSP whose chargebacks quietly doubled from 1.5% to 3% of revenue over six months won't see it in a lump-sum entry — they'll just notice, one day, that margins feel tighter than they should.
The fix is to book each settlement as multiple journal lines instead of one deposit: base route revenue, per-package revenue, incentive revenue, and a contra-revenue line for chargebacks, each in its own account. Once chargebacks live in their own line, you can watch the percentage every week and catch a spike before it becomes a pattern — and before it becomes the reason you're disputing three months of deductions you can no longer prove were wrong.
The Scorecard Is a Revenue Multiplier, Not a Report Card
It's tempting to treat the weekly scorecard rating (Fantastic Plus, Fantastic, Great, Fair, Poor) as an HR metric — something Ops worries about, not Accounting. That's a mistake. The scorecard tier is effectively a pricing tier, and the gap between the top and the middle is enormous:
| Rating | Incentive impact (monthly, ~25 routes) |
|---|---|
| Fantastic Plus / Fantastic | +$8,000 to +$15,000 |
| Great | +$3,000 to +$8,000 |
| Fair | $0 to +$2,000 |
| Poor | -$5,000 to -$15,000 (route reductions, no incentives) |
In practical terms, the swing between "Fantastic" and "Fair" can run $10,000–$15,000 a month — often the entire difference between a healthy profit and breakeven. That means your bookkeeping should treat scorecard tier as a KPI sitting right next to revenue and labor cost, not as a separate Ops dashboard nobody in Accounting ever opens. If the P&L and the scorecard aren't reviewed side by side every week, you'll find out about a bad month only after it's already happened.
The True Cost of a Driver Is Never the Hourly Wage
Amazon has been raising the wage floor, with target driver pay now landing in the roughly $20–$27 an hour range depending on market. That number is the one everyone budgets from — and it's the one that gets DSP owners in trouble, because a driver's fully loaded cost typically runs 50–60% above the base wage once everything is included:
| Cost component | Annual cost (per driver) | % of base wage |
|---|---|---|
| Base wages (~50 hrs/week) | $41,600–$52,000 | 100% |
| Overtime premium | $6,000–$7,800 | 14–15% |
| Payroll taxes | $4,200–$5,200 | 10% |
| Health insurance | $4,800–$7,200 | 12–14% |
| Workers' comp | $2,800–$4,500 | 7–9% |
| Fully loaded total | $63,000–$82,600 | 152–159% |
A driver earning $20 an hour is really costing the business closer to $30–$32 an hour once taxes, benefits, and workers' comp are added in. Owners who budget only on the sticker wage are budgeting a business that's already 30–50% underfunded on its largest expense line before the first delivery van leaves the lot.
Overtime deserves its own line item, not a footnote. Fully loaded labor should run 55–62% of revenue in a healthy DSP; once overtime alone climbs past roughly 22% of total payroll, that's a signal of a staffing or scheduling problem, not just a cost problem — it usually means the business is understaffed and paying a premium to cover the gap.
Van Lease vs. Buy: A Decision Your Books Should Be Driving
Every DSP eventually faces the same fork: keep leasing Amazon-branded vans, or buy and finance a fleet outright. The numbers on paper look deceptively simple:
- Amazon lease program: roughly $1,800–$2,200/month per van, partial maintenance included, no upfront capital, must stay Amazon-branded
- Owned fleet: roughly $600–$900/month in loan payments alone, but $25,000–$45,000 upfront per van, 100% of maintenance risk transferred to the owner, and full tax depreciation benefits (including Section 179 in the year the vehicle is placed in service)
The lease payment looks two to three times more expensive at first glance — until you account for maintenance. Delivery vans routinely rack up 30,000–50,000 miles a year in stop-and-go city driving, and owned vehicles should be budgeted at $250–$400/month per van just for routine upkeep, with real-world repair bills on aging branded vans sometimes spiking into the thousands. Run the comparison as a fully loaded cost per van per month — lease payment vs. (loan payment + maintenance reserve + insurance delta) — not as a bare monthly payment comparison, or the decision will look better than it actually is in year two and three.
Fuel is its own budget line, and it's a bigger one than most new owners expect. Delivery vans typically get 12–15 MPG in stop-and-go conditions; a 30-van fleet running daily routes can easily post $23,000–$35,000 a month in fuel alone. That's a cost that should be tracked per van, per mile, not lumped into a single "vehicle expense" account — a spike in cost-per-mile is often the earliest signal of a maintenance issue or fuel-card misuse before either shows up anywhere else.
Insurance: Book It Monthly, Not All at Once
Amazon's insurance requirements are non-negotiable and substantial: commercial auto with a $1 million combined single limit, general liability at $1 million per occurrence / $2 million aggregate, cargo/bailee coverage, state-mandated workers' comp, and often $5 million or more in umbrella coverage recommended on top. All told, insurance typically consumes 8–12% of gross revenue — one of the largest line items after labor and vehicles.
The bookkeeping mistake here is booking a full annual premium as a single expense the month it's paid. That distorts the P&L for that month and understates every other month, making it look like the business had one terrible month and eleven great ones. The correct treatment is to book the premium to a Prepaid Insurance asset account, then recognize 1/12 of it as an expense each month — the same logic as any other prepaid expense, and it's what keeps month-to-month margin comparisons honest.
Cash Flow: The Timing Mismatch Nobody Warns You About
Even a profitable DSP can run into a cash crunch, because the timing of money in rarely matches the timing of money out:
- Amazon settlements typically deposit Tuesday through Thursday
- Driver payroll runs bi-weekly or semi-monthly, on its own fixed schedule
- Lease payments, insurance, and fuel cards often bill weekly or monthly, on yet another schedule
None of these calendars line up, which means a DSP can be fully profitable on paper and still come up short in the checking account the week payroll, the lease, and an insurance renewal all land at once. The standard cushion recommended for this is two to three weeks of operating expenses held in reserve — for a 25-route operation, that's often in the $80,000–$120,000 range. That reserve isn't idle money; it's what keeps a profitable business from missing payroll during a bad settlement week.
The Real Payoff of Doing This Right
None of this is theoretical. One 25-route DSP generating $3.2 million in annual revenue was taking home just $80,000 — a 2.5% margin, uncomfortably thin for the risk involved. A closer look at the books, using exactly the categorization described above, surfaced $127,000 in recoverable money: $34,000 in chargebacks that were disputable but never disputed, $48,000 in overtime that scheduling changes could eliminate, $29,000 in insurance the owner was overpaying for, and $16,000 in fuel-card abuse nobody had noticed. That's take-home income more than doubling — to over $200,000 — without adding a single route. The routes didn't need to grow. The books needed to.
Keep Your DSP's Books as Precise as Your Delivery Windows
Running a DSP means treating scorecard tiers, chargebacks, fully loaded labor, and fleet costs as data you track weekly, not guess at quarterly — the same discipline Amazon expects of your on-road performance. Beancount.io brings that same rigor to your financial records: plain-text, version-controlled accounting where every settlement, chargeback, and payroll run is a transparent entry you can audit and query, not a black box you have to trust. Get started for free and see why operators who live and die by their numbers are switching to plain-text accounting.
