At a glance
Share of Japan Post consignments throwing at least one tracking exception in the trailing 30 days. Domestic exceptions (held at post office for collection, recipient absent, address invalid) are rare; international EMS exceptions (customs hold, recipient absent abroad, address insufficient) are more common. Rising exceptions today predict a falling On-Time Delivery Rate at 24 to 96 hour lag.
| What it counts | COUNT(shipments WHERE has_exception_event) / COUNT(all_shipments) over 30 days. Multi-exception consignments count once. |
| API endpoint | POST /api/v1/tracking/list reads events[]; any event with status in (HELD_AT_POST_OFFICE, RECIPIENT_ABSENT, ADDRESS_INSUFFICIENT, CUSTOMS_HOLD, WEATHER_DELAY, etc.) flags the consignment. |
| Service-tier scope | All tracked services: Yu-Pack, Yu-Pack Cool, EMS, Letter Pack Plus, Letter Pack Light. |
| Geographic OTD variance | Domestic exceptions cluster around redelivery requests in dense urban Tokyo and Osaka where recipients are out at first attempt. International EMS exceptions concentrate by destination: UK (Brexit customs), US (less frequent but address-format issues), EU (slower customs in Italy/Spain/France). |
| Returns / RTO | Outbound only. |
| Climate handling | Typhoon, snow, earthquake events flagged via WEATHER_DELAY exception code. The card includes them; weather is part of operational reality. |
| Peak-period seasonality | Oseibo and Ochugen peak adds 0.5 to 1.5 points to exception rate from “held at post office” overflow when post office collection capacity stretches. |
| Time window | 30D vsP (rolling 30 days, period-over-period). |
| Alert trigger | >3% warn, >5% critical. Domestic baseline is 0.8 to 1.5 percent; international can run 5 to 12 percent. Cross-border merchants’ aggregate is 2 to 4 percent typically. |
| Sentiment key | {'type': 'gauge', 'thresholds': {'good': 1, 'warn': 3}} |
| Roles | owner, operations |
Calculation
Calculated automatically from your Japan Post 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 Japanese-headquartered DTC apparel brand based in Osaka. ~3,200 outbound parcels per week, mixed mode: 78 percent Yu-Pack domestic, 22 percent EMS international (US, UK, AU, KR, TW). Reading taken at 09:00 JST on 12 Apr 26 for the trailing 30 days (13 Mar 26 to 11 Apr 26).| Service code | Consignments | Exceptions | Exception Rate | Most common exception |
|---|---|---|---|---|
| Yu-Pack (domestic) | 9,840 | 124 | 1.26% | RECIPIENT_ABSENT (52% of cases), HELD_AT_POST_OFFICE (38%) |
| Yu-Pack Cool (refrigerated, domestic) | 880 | 28 | 3.18% | RECIPIENT_ABSENT (60%, cold chain re-attempt is timed) |
| EMS US | 1,420 | 78 | 5.49% | ADDRESS_INSUFFICIENT (34%), RECIPIENT_ABSENT (28%), CUSTOMS_HOLD (22%) |
| EMS UK | 380 | 41 | 10.79% | CUSTOMS_HOLD (54%, post-Brexit), ADDRESS_INSUFFICIENT (24%) |
| EMS EU (DE/FR/IT/ES) | 280 | 32 | 11.43% | CUSTOMS_HOLD (58%, IOSS not configured), ADDRESS_INSUFFICIENT (22%) |
| EMS APAC (KR/TW/AU/SG) | 600 | 28 | 4.67% | ADDRESS_INSUFFICIENT (40%), RECIPIENT_ABSENT (32%) |
| All Japan Post (this card) | 13,400 | 331 | 2.47% | (mixed) |
>3% is not tripped. Five things to notice:
- Domestic exception rate (1.26%) is excellent. Japan Post’s domestic network is among the most reliable globally; an absent-recipient re-attempt is the dominant exception type, and Japan Post automatically schedules redelivery via the Yubin-bin tracking app, most resolve within 24 hours without merchant intervention. Pair with First-Attempt Delivery for the leading indicator.
- EMS UK is the loudest outlier at 10.79%. Post-Brexit customs requirements (full HS code, item-level value declaration, IOSS / VAT registration) trip up roughly half of UK shipments where the merchant hasn’t pre-configured. The 41 UK exceptions in 30 days represent ~JPY 580,000 of revenue tied up in customs holds, plus the customer-experience hit of waiting 7 to 14 extra days. The single highest-leverage fix is IOSS registration, which removes the majority of customs holds for shipments under EUR 150.
- EMS EU rate (11.43%) shares the same root cause. EU customs reform (July 2021 onward) eliminated the EUR 22 threshold for VAT exemption; every shipment now needs IOSS or DDP handling. Without IOSS, parcels held at customs for VAT collection from the recipient produces customer dissatisfaction and in 20 to 30% of cases an outright return-to-sender.
- EMS US (5.49%) is the lowest-friction international destination. US customs is mostly automated for shipments under USD 800 (de minimis); exceptions there are mostly genuine address issues or recipient-absent. The 78 US exceptions in 30 days are usually fixable via email outreach to the customer for an address correction.
- The 2.47% aggregate hides everything that matters. The realm-wide reading is below the alert threshold and looks healthy; the EMS-only sub-rate is 7.4%, which is operationally serious. Always pin per-service per-destination panels for cross-border merchants. The aggregate is a navigation surface; the per-lane breakdown is the actionable read.
Sibling cards merchants should reference together
| Card | Why pair it with Exception Rate | What the combination tells you |
|---|---|---|
| On-Time Delivery Rate | Downstream impact. | Rising exception rate predicts a falling OTD rate at 24 to 96 hour lag. The two should move inversely. |
| Late Shipments | The absolute count companion. | An exception that doesn’t resolve becomes a late shipment. |
| Failed Delivery Count | The unresolved-exception subset. | Exceptions that go more than 7 days without delivery are usually counted here. |
| First-Attempt Delivery | The leading indicator. | A drop in first-attempt rate precedes an exception-rate climb by 24 to 48 hours. |
| Cost by Zone | Operational cost view. | Exceptions are expensive: redelivery costs, customer service tickets, refunds. Pair to triangulate the per-zone economics. |
| Returned to Sender | The terminal state for unresolved exceptions. | EU customs holds that go unresolved typically become RTS at Day 14 to 21. |
Cross-connector: shopify.refund_rate | Downstream impact. | A 2pp exception-rate climb on EMS lanes typically precedes a 0.5 to 1pp refund-rate rise 14 to 21 days later. |
Cross-connector: royal_mail.roy_exception_rate | UK postal peer. | Different network, similar metric. Useful for an agency running both, not a like-for-like reconciliation. |
Reconciling against the vendor’s own dashboard
Where to look in Japan Post’s own portal: Japan Post Business Customer Portal (Yubin Business Customer site, JP-language primary) provides per-shipment tracking and a Tracking Result Inquiry (お荷物追跡サービス) view at the per-tracking-number level. For aggregate exception trends, the dashboard is sparse, most enterprise merchants reconcile by exporting tracking data via the API and computing the rate themselves. For EMS specifically, the International Mail Tracking view shows the full event stream including customs and delivery exceptions; matches this card to within sync-lag timing. For non-Japanese-reading merchants, the third-party tracking aggregator 17Track provides a comparable view with English summaries; useful for sanity-checking but not as authoritative as Japan Post’s own portal. Why our number may legitimately differ from Japan Post’s portal:| Reason | Direction | Why |
|---|---|---|
| Time-zone (JST vs UTC) | Boundary days off | Japan Post’s portal reports in JST; the card stores everything in UTC. JST is UTC+9; midnight JST is 15:00 UTC the previous day. For 30-day windows the gap averages out. |
| Tracking-event ingestion lag | Ours lower for “today” | Japan Post pushes scans to its API in batches; lag varies from 30 minutes (typical) to 4 hours (peak periods, post-Oseibo / Ochugen). |
| Multi-exception consignments | Ours lower than per-event counts | The card counts each consignment once even if multiple exceptions fired. Some Japan Post per-event reports count each exception separately. |
| Service-tier scope | Either | Japan Post’s portal often defaults to a single service code; the card sums every tracked service unless filtered. |
| Strikes / climate events | Same in both | Typhoon, earthquake, snow events flagged via WEATHER_DELAY or operational-notice exceptions; the card includes them. |
| Outbound vs returns | Ours lower for RTS-heavy merchants | The card excludes inbound consignments (the Japan Post Returns API treats them as a separate stream). |
| Card | Expected relationship | What causes legitimate divergence |
|---|---|---|
shopify.unfulfilled_orders | Upstream input | Orders waiting for a Yu-Pack label cannot meet OTD if they sit too long; pre-label-print delays don’t show up as Japan Post exceptions but produce the same customer-experience outcome. |
shopify.refund_rate | Downstream sentiment proxy | Late and exception-flagged deliveries drive WISMO tickets and refund requests; a 2pp exception-rate climb typically precedes a 0.5 to 1pp refund-rate rise at 14 to 21 days. |
royal_mail.roy_exception_rate | UK postal peer; not a reconciliation | Different network, similar definition. Cross-link only. |
australia_post.aus_exception_rate | APAC postal peer | Different network. Cross-link only. |
Known limitations / merchant FAQs
Why is my domestic Yu-Pack exception rate so low (around 1%)? Japan Post’s domestic network is genuinely among the best in the world. Recipient-absent re-attempts are scheduled automatically via the Yubin-bin (Japan Post tracking app), and most consumers either re-direct to a Konbini (24-hour convenience store pickup) or schedule a specific 2-hour redelivery window. The infrastructure absorbs first-attempt failures without producing a customer-visible exception in most cases. Why is my EMS exception rate so high (5 to 12%)? Two structural drivers. (a) Customs friction in the destination country. UK and EU customs reform (Brexit + EU IOSS) requires every shipment to carry HS codes, item-level value declarations, and IOSS / VAT registration; without these, parcels hold at customs for the recipient to pay. (b) Address-format mismatches. Japanese-format addresses (postal code → prefecture → city → ward → street → building) do not always translate cleanly to Western formats; auto-translated addresses sometimes fail destination-country validation. The fix is enforcing structured address capture at checkout. The single biggest lever to cut my international exception rate? Register for IOSS (EU shipments under EUR 150) and DDP-equivalent for UK (under GBP 135). Both pre-pay VAT at the time of sale, eliminating the customs-hold-for-recipient-payment exception class entirely. ROI is typically 2 to 4 weeks of reduced customer-service ticket volume; merchants commonly see EMS UK and EMS EU exception rates drop from 10 to 12% down to 3 to 5% within a month of IOSS configuration. My exception rate spiked during Oseibo (year-end gift season). Is that normal? Yes. Oseibo (December) and Ochugen (July) are Japan’s two major gift-giving seasons, when domestic Yu-Pack volume runs 2 to 3× normal and post offices fill with held parcels. Exception rate (specificallyHELD_AT_POST_OFFICE) climbs 0.5 to 1.5 percentage points during these windows. The pattern recovers within 7 to 10 days after the peak. Don’t tune alert thresholds for the off-season norm; raise them temporarily during these windows or your alerts will fire continuously.
A typhoon shut down service for 3 days. The card flags weather delays as exceptions. Is that right?
Yes, weather-driven exceptions are real exceptions, parcels were delayed, customers were affected, and the operations team needs to know. Some merchants prefer a “controllable exceptions” view that excludes WEATHER_DELAY; that’s a separate filter, not the default. The card surfaces what actually happened to the parcels.
What’s the relationship between exception rate and OTD rate?
Inverse, with a 24 to 96 hour lag. An exception that resolves quickly (recipient-absent → redelivery scheduled within 24 hours) doesn’t always hurt OTD. An exception that doesn’t resolve (customs hold for 7 days, address insufficient with no resolution) becomes a missed OTD. The card pair is the right diagnostic, exception rate climbs first, then OTD drops 1 to 4 days later.
Customers report “no notification, parcel just held at post office for 5 days”. What’s happening?
Two things. (a) Japan Post’s notification SMS / mail goes to the recipient, not the merchant; if the contact details on the label are stale or the recipient is travelling, they don’t see the redelivery prompt. (b) International EMS delivery scans on the destination country’s network can be sparse (especially APAC and EU); the card sees the Japan Post-side exception but not the destination-country delivery action. The fix is adding an end-of-window automated email to the customer (“if you haven’t heard from your local post in X days, contact us”) to bridge the gap.
Should I read this card per-service or aggregated?
Per-service-per-destination, always. The aggregate is a navigation surface; the operationally-actionable read is per lane. A 2.5% aggregate looks healthy until you realise EMS UK is at 11%; the per-lane breakdown is where decisions happen.
My exception rate has been climbing 0.2 percentage points per week for 6 weeks. What’s going on?
Sustained slow climbs usually indicate either (a) a creeping address-validation issue (e.g. a checkout deploy stripped a building-line input), (b) a destination-country customs change (Japan Post publishes Service Updates when customs rules shift in major destinations), or (c) a sender-side label-data quality drift (e.g. an OMS upgrade changed how addresses are formatted on labels). Investigate in that order: checkout first (cheap to fix), Japan Post service updates second, OMS configuration third.
The card threshold is >3% warn but my international rate is always above that. How do I tune?
Raise the alert to a per-service threshold. Run a domestic-only alert at >2% and a separate international alert at >8% (or >5% if you’ve configured IOSS / DDP). The aggregate alert is not informative for cross-border merchants; the per-lane alerts are.