At a glance
Count of Amazon Prime Shipping consignments that missed their 1-Day or 2-Day Prime promise in the period. Each missed consignment is one Prime customer who paid the Prime membership for delivery reliability and got something slower. The count drives directly into Amazon’s SFP eligibility floor (99% on-time required); a 7-day late count above 5% of weekly volume threatens the Prime badge.
| What it counts | COUNT(shipments WHERE actual_delivery_date > prime_promise_date AND status IN ('DELIVERED','FAILED')) over rolling 7 days. |
| Delivery success criterion | Carrier POD scan or signature event captured at the recipient. |
| On-time threshold | Prime promise date communicated at customer purchase. End of promise day is the cutoff. |
| Returns / RTO | RTOs excluded. Failed-and-redelivered scores against first successful POD. |
| Service level scope | Prime-eligible consignments only (1-Day, 2-Day, Same-Day where applicable). |
| Money-back-on-late | Amazon does not refund Prime carriage to the seller; the seller already absorbed it. Customer-side, Amazon may issue a goodwill credit to the Prime member; cost flows through Concession metrics, not directly to seller. The card surfaces SFP eligibility risk, not refund recovery. |
| 2-day Prime coverage gaps | Rural ZIPs, AK/HI/PR, and remote zones are excluded from 2-Day eligibility. Card scores only Prime-promised consignments. |
| Buy with Prime cohort | Same definition. Buy with Prime customers are Amazon Prime members buying on the merchant’s direct site; Prime promise applies to those orders. |
| Time window | 7D |
| Alert trigger | >5% of total. On a Prime-heavy seller this corresponds to ~95% OTD; sustained at this level the SFP eligibility threshold (99%) is well-breached. |
| Roles | owner, operations |
Calculation
Calculated automatically from your Amazon Prime Shipping (SFP) data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.Worked example
A US home-goods merchant running SFP from a single Atlanta warehouse: $65 average order value, mostly UPS Ground 2-Day to East Coast, OnTrac 2-Day to West Coast. Reading taken at 09:00 ET on 18 Mar 26 for the trailing 7 days (11 Mar 26 to 17 Mar 26).| Carrier / cohort | Shipments (7D) | Late (POD after Prime promise) | Late % |
|---|---|---|---|
| UPS Ground 2-Day (East/Midwest) | 1,180 | 14 | 1.2% |
| OnTrac 2-Day (West Coast) | 320 | 8 | 2.5% |
| USPS Priority (rural edge ZIPs) | 145 | 11 | 7.6% |
| Amazon Logistics | 80 | 1 | 1.3% |
| Total Prime-promised | 1,725 | 34 | 2.0% |
- The 2.0% headline is comfortable for this card; it is not for SFP eligibility. SFP requires <1% late (i.e. >99% OTD). At 2.0% late, the merchant is at 98% OTD, below the SFP eligibility floor. Pair with
amazon_prime_sfp_otd_slaandamazon_prime_promise_miss_ratefor the badge-retention view. - USPS Priority rural at 7.6% late is the SFP-eligibility-killer. Twelve weeks of USPS Priority for rural ZIPs at 7.6% drags the seller’s account-wide eligibility below threshold. Consider opting these ZIPs out of Prime in Seller Central or shifting to a non-Prime fulfilment path.
- 34 missed Prime promises = 34 Prime-member-disappointed-experiences. Prime member refund / credit-back rate is 4 to 6x non-Prime cohort. Estimated cost: 34 × 10 average concession-cost = 340 of Amazon-mediated goodwill on top of the carrier-side cost.
- OnTrac at 2.5% is borderline. West Coast 2-Day on OnTrac is the cohort to watch second. Consider Amazon Logistics for the OnTrac-served ZIPs where coverage overlaps.
- Compare against last week (10 Mar 26 reading). If late count was 24 last week and 34 this week, the 40% week-on-week rise is the real story. Most likely cause: a new product launch lifted volume past the merchant’s negotiated daily UPS pickup capacity, pushing afternoon orders to next-day-collection cohorts.
Sibling cards merchants should reference together
Late count is a count, not a rate. Pair with these:| Card | Why pair it with Late Shipments | What the combination tells you |
|---|---|---|
| On-Time Delivery Rate | The percent companion. | Mathematical mirrors with different windows (this 7D, OTD 30D). |
| SFP On-Time Delivery vs 99% Threshold | The Amazon SFP eligibility-defining card. | Always check; SFP threshold is the Prime-badge contract. |
| Prime 1-Day / 2-Day Promise Miss Rate | Inverse view. | Use both for trend; this card for response capacity-planning. |
| SFP Eligibility Risk by ASIN | Per-ASIN view. | Identifies which products are dragging the headline. |
| Late-Shipment Buy Box Loss | Revenue impact estimate. | The financial cost of a 7-day late-count rise. |
| OTD by Route | Carrier and zone split. | A single underperforming carrier (often USPS Priority for rural). |
Cross-connector: amazon.amazon_otdr | Amazon’s seller-account-wide OTDR. | SFP slice feeds in. |
Cross-connector: amazon.buy_box_win_rate | Direct downstream. | Late count rise predicts Buy Box decay. |
Cross-connector: shopify.unfulfilled_orders | For Buy with Prime cohort. | Upstream of the Prime promise; late dispatch breaks 2-Day. |
Reconciling against the vendor’s own dashboard
Where to look in Amazon Seller Central: Amazon Seller Central → Performance → Shipping Performance → Late Shipment Rate for Prime-eligible orders. For Buy with Prime, the dashboard is at buywithprime.amazon.com/dashboard → Order Metrics. Why our number may legitimately differ:| Reason | Direction | Why |
|---|---|---|
| Time zone | Boundary | Amazon Seller Central defaults to PST. Card uses connected timezone. |
| Promise-clock-start | Either | Both score against customer-promise-at-purchase. |
| Exclusions | Ours higher | Amazon dashboard sometimes excludes seller-cancelled or customer-refunded orders. |
| In-transit | Ours rolling | Consignments without final POD not counted yet. |
| Card | Expected relationship | What causes legitimate divergence |
|---|---|---|
amazon.amazon_otdr | Account-wide OTDR. | Includes FBA. |
amazon.buy_box_win_rate | Direct downstream. | Buy Box also affected by price, stock. |
Known limitations / merchant FAQs
Why is the alert at 5% when SFP requires <1%? This card is the operational alarm; it intentionally runs above the SFP threshold. A 5% weekly miss rate is the “your operations are clearly broken” trigger; a 1.5% rate is the “you are SFP-eligibility-at-risk” trigger surfaced onamazon_prime_sfp_otd_sla. Use both cards in parallel.
Rural ZIPs at high late % keep dragging the count; what is the lever?
Two paths. (1) Opt out of Prime for those ZIPs in Seller Central → Account Info → Prime Settings. (2) Improve the carrier mix to those ZIPs (USPS Priority is often the wrong choice for edge ZIPs; USPS Ground Advantage as non-Prime alternative, or accept the volume drop and protect the badge).
Each Prime miss costs more than non-Prime miss; how do we quantify?
Three layers: (1) carrier-side cost paid (already absorbed). (2) Amazon Concession metric: Prime members refund / credit-back at 4 to 6x non-Prime cohort, costing 10 average concession per miss. (3) Future Buy Box / Prime-badge cost: each miss contributes to the seller-account-wide eligibility floor; sustained at 5%+ weekly the badge is lost and Buy Box win rate drops 30 to 50%. The third cost dwarfs the first two.
Day-2 redelivery on a 2-Day Prime: how is it counted?
Late. The Prime promise is end of promise day; arrival on day 3 (or day 2 of a 1-Day promise) is past cutoff. Re-attempts that successfully deliver after the promise day score as late on this card.
Buy with Prime: same rules?
Yes. Buy with Prime customers are Amazon Prime members who chose Prime-tier shipping on the merchant’s direct site; they receive the same Prime promise. Late on Buy with Prime hurts the merchant’s relationship with Amazon’s Buy-with-Prime program (which can be revoked) and the customer-trust signal.
During Q4 peak, what is realistic?
Q4 typically lifts late count 2 to 4x as carrier networks saturate. Many SFP merchants suspend SFP for 6 weeks (Amazon allows seasonal suspension by request) rather than risk badge loss. Consider this a strategic decision in October, not an emergency in December.
Single-warehouse SFP: realistic 2-Day national coverage?
Hard. From Indianapolis or Atlanta, UPS Ground 2-Day reaches ~70% of US population reliably; OnTrac extends to West Coast metros. Rural and Mountain West typically require either FBA, USPS Priority (less reliable), or accepting Prime-badge-edge risk on those ZIPs. National 2-Day at 99%+ typically requires 2 to 3 warehouses geographically distributed.
Same-Day Prime: how is it scored?
Same definition: delivered on or before the promised date (which is the same calendar day for Same-Day Prime). Same-Day is a small cohort for most merchants and runs higher OTD on Amazon Logistics (which dominates Same-Day delivery).
Compared to ShipBob’s OTD: which is the right comparison?
ShipBob is a 3PL serving direct DTC orders (typically not Prime-promised); ShipBob’s OTD measures against ShipBob’s per-shipment estimated delivery date. Amazon Prime measures against Amazon’s customer-facing 1-Day or 2-Day promise. The two are independent operations; SFP is uniquely Amazon-controlled. Some merchants run ShipBob for direct DTC alongside SFP for Amazon-channel; the cards are not reconciled.
SFP suspension: how does it work?
Amazon allows SFP enrolled sellers to request seasonal suspension via a Seller Central case. Suspension lasts the requested period (typically 4 to 8 weeks). During suspension the seller cannot offer Prime shipping; reverting requires re-qualification (a 30-day trial period at >99% OTD). Many merchants pre-plan a 6-week Q4 suspension.