DEFT's Four Demands, Translated to Dollars

In November 2025, a group of Amazon delivery service partners went public as DEFT — DSPs for Equitable and Fair Treatment. The coalition retained legal counsel from Shumaker, Loop & Kendrick and identified four specific program improvements: driver wages and bonuses, vehicle reimbursements, rate cards that keep pace with inflation, and simplified scorecards with operator-controlled metrics. According to coverage in The HR Breakdown and Yahoo Finance, the group formally launched on Black Friday, giving its timing a pointed edge.

What most coverage missed is the dollar math behind each demand. These aren't abstract grievances. Every one of them connects to a cost line or revenue line that operators can measure — and that Amazon's current program structure leaves poorly compensated.

Demand 1: Rate Cards That Keep Pace With Inflation

Amazon raised its per-package rate by 20% beginning in January 2025, part of a $2.1 billion investment in the DSP program announced via About Amazon. The company also committed an additional $660 million specifically toward DSP rate card increases and driver pay, with a target national driver wage of approximately $22–$23 per hour, per program announcements.

The 20% headline figure is real. The problem is the baseline it's adjusting from. Bureau of Labor Statistics CPI data shows fleet insurance up 46.8% since 2019. Vehicle maintenance costs have risen substantially over the same period, though the specific index trajectory differs by component. Used delivery van prices remain approximately 33–35% above 2019 levels even after a partial correction from the 2021 peak, per Manheim Used Vehicle Value Index tracking of commercial van segments. A rate card that receives periodic adjustments against costs that have compounded continuously since 2019 will always be chasing. DEFT's demand isn't for a specific percentage — it's for a structure that adjusts automatically, rather than requiring operators to advocate for corrections every few years.

The rate-compression math shows why the gap matters at scale:

| Rate Compression | Per-Van Annual Loss | 10-Van Fleet | 50-Van Fleet | |---|---|---|---| | 5% | $2,950 | $29,500 | $147,500 | | 10% | $5,900 | $59,000 | $295,000 | | 15% | $8,850 | $88,500 | $442,500 |

A rate card that drifts even 5% below cost inflation annually takes nearly $30,000 per year from a 10-van operation — before the insurance and maintenance lines compound further.

Demand 2: Higher Vehicle Reimbursements

DEFT's specific ask for better vehicle reimbursement is grounded in the lifecycle economics of the program's standard delivery vans. The buy-versus-rent economics are time-limited, and operators are bearing the cost when their fleets age past the crossover.

Using a Mercedes Sprinter as a reference point — the platform most commonly associated with longer-lifecycle DSP operations — Pexara TCO data shows the following all-in monthly cost by mileage band versus a $1,450/month contract rental (Pexara fleet benchmark, derived from operator-reported rental agreements in the 2025 DSP operator survey):

| Mileage Band | Maintenance/mo | Loan Payment | Total/mo | vs. $1,450 Rental | |---|---|---|---|---| | 0–40K | $95 | $1,050 | $1,145 | Save $305 | | 40–60K | $180 | $1,050 | $1,230 | Save $220 | | 60–80K | $380 | $1,050 | $1,430 | Save $20 | | 80–100K | $520 | $1,050 | $1,570 | Over by $120 |

The $305 monthly ownership advantage at low mileage is the economic case Amazon built its DSP model around. By 78,000 miles, that advantage is gone. Past 80,000 miles, ownership costs more than renting — $120 per van per month. A program reimbursement structure calibrated to early-lifecycle economics quietly transfers that cost to the operator as the fleet ages.

The optimal trade window by equity and cost data is 60,000–85,000 miles. Holding to 100,000 miles saves only approximately $168 per month versus new financing payments — not enough to justify turbo and injector exposure in the $4,600–$5,100 and $2,200–$2,400 ranges respectively, per RepairPal data.

Demand 3: Simplified Scorecards With Metrics Operators Can Control

DEFT's scorecard demand asks for "metrics controlled by the DSPs" and compensation for initiatives aimed at "raising the bar." The language is diplomatic. The financial exposure is specific.

Amazon's Fantastic Plus tier unlocks meaningful bonus revenue — the weekly difference between qualifying and not qualifying is real operating income, not a rounding error. The scorecard mechanism that governs that tier includes metrics partly outside operator control: customer-reported issues, delivery photo disputes, access complications at addresses, and any route modification that creates missed-stop risk regardless of driver performance.

The failure cascade looks like this: a single van breakdown mid-route → missed deliveries → Delivery Completion Rate (DCR) degrades → Fantastic Plus threshold missed for the week → bonus revenue lost. The operator controlled step one (van maintenance). They did not control steps two through four. A scorecard structure that penalizes metrics driven by customer behavior, address-level infrastructure, or Amazon routing decisions creates revenue unpredictability that operators cannot budget against.

The simplification demand isn't asking to lower the performance bar. It's asking to remove inputs that aren't operationally within the operator's sphere — which is a reasonable basis for a business relationship.

Demand 4: Sustainable Driver Pay and Bonus Structures

Amazon's investment targets a national average driver wage of $23 per hour, up from approximately $21–$22 per hour, according to About Amazon program announcements. DEFT's specific demand focuses on how bonus structures actually flow — whether program-level bonus criteria are achievable and consistently paid, or whether the criteria shift in ways that prevent reliable budget planning.

The reason driver pay structure matters is the turnover data. Bureau of Labor Statistics JOLTS for NAICS 4921 (courier and messenger services) shows approximately 80% annual driver turnover. SHRM benchmarks put non-CDL driver replacement cost at $3,500 to $5,500 per hire, covering recruiting, onboarding, and productivity loss. At 80% turnover, that math runs continuously:

| Fleet Size | Drivers Replaced/Year | Annual Replacement Cost | |---|---|---| | 10 vans | 8 | $28,000–$44,000 | | 50 vans | 40 | $140,000–$220,000 | | 100 vans | 80 | $280,000–$440,000 | | 300 vans | 240 | $840,000–$1,320,000 |

Higher base pay and clearly structured bonuses reduce turnover. Lower turnover reduces replacement cost. The math is not ambiguous: for most operators, a $1–$2/hour pay increase that cuts turnover by even 20 percentage points returns more than its cost inside 12 months. DEFT's demand on driver pay isn't generosity — it's correcting a cost-of-turnover problem that operators currently absorb without program support.

What DEFT's Existence Signals About Program Economics

A coalition of operators retaining legal counsel and forming a structured advocacy organization is a late-stage signal. Operators don't organize — and spend money on lawyers — when margins are fine. They organize when the program economics have moved far enough from the original business case that informal feedback isn't producing results.

Amazon's response — the $2.1 billion investment, the 20% rate increase, the AI digital assistant rollout — suggests the company is aware the cost picture has shifted. Whether DEFT's specific demands reshape the program structure remains to be seen. What's certain is that operators who already understand their unit economics — cost per stop, per-van maintenance curve, actual driver turnover cost, Fantastic Plus revenue exposure — are positioned to benefit most from any structural improvements DEFT secures. You can't measure the improvement if you don't know the baseline.

Sources: The HR Breakdown (December 2025); Yahoo Finance; About Amazon (program announcements); Chain Store Age; Pexara fleet cost benchmarks and TCO model; Bureau of Labor Statistics JOLTS NAICS 4921; SHRM driver replacement benchmarks; RepairPal repair cost database.