Skip to main content
Card class: Hero • Category: Shipping & Courier

​
At a glance

Share of APC Overnight consignments in the period that threw at least one tracking exception event before final delivery. “Exception” means any non-progressing scan: recipient absent, carded for collection, refused, address invalid, access denied, customs held (for Globe export), depot misroute, damage report. A consignment that threw an exception but eventually delivered on time still counts towards this rate, the metric tracks operational friction, not lost deliveries.
What it countsCOUNT(shipments WHERE any tracking_event.code IN exception_codes) / COUNT(shipments WHERE status IN ('DELIVERED','FAILED','RTS')). One exception event flips the consignment regardless of how many subsequent normal scans occur.
API endpointAPC getConsignmentTrackingDetails reads the full event ladder per consignment. Vortex IQ matches event codes against APC’s published exception-code list (CARDED, REFUSED, INVALID-ADDRESS, ACCESS-DENIED, DAMAGE-REPORT, CUSTOMS-HELD, MISROUTE, plus driver-comment exception codes).
Exception codes includedCARDED (calling card left), REFUSED (recipient declined), INVALID-ADDRESS (address could not be located), ACCESS-DENIED (gated property, no access code), DAMAGE-REPORT (driver flagged damage in transit), MISROUTE (parcel sent to wrong depot), CUSTOMS-HELD (Globe export only). Driver discretionary codes (e.g. BAD-WEATHER) are pooled here too.
Delivery success criterionIndependent of exceptions. A consignment can throw an exception, be redelivered the next day, and count as both “exception” (this card) and “on-time” (OTD Rate) if redelivery beat the original cutoff.
Service level scopeAll services pooled (NextDay-9am, NextDay-10:30, NextDay-12, NextDay anytime, Pallet 24, Pallet 48). Pallet exceptions are typically structural (over-height, damaged shrink-wrap); parcel exceptions are usually access / recipient.
Multiple exceptions per consignmentOne consignment counts once regardless of how many exception events fire. A parcel that threw CARDED then REFUSED then RTS still counts as a single exception in the numerator.
Returns / RTOReturned-to-sender consignments are included in the denominator and almost always included in the numerator (the RTS state is itself triggered by an exception sequence).
Time zoneUK local (GMT or BST). All exception events carry depot-local timestamps.
Time window30D vsP (rolling 30 days, period-over-period). Daily readings are noisy below 200 consignments per day.
Alert trigger>3% warn, with the gauge thresholds at 1 percent (good) and 3 percent (warn). UK premium-express benchmark is 0.8 to 1.5 percent steady-state; above 3 percent for two weeks running is operations-team escalation.
Rolesowner, operations

​
Calculation

Calculated automatically from your APC Overnight 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 UK B2B medical-supplies distributor based in Reading, around 2,800 consignments per month split 75 percent NextDay-9am (clinic and surgery deliveries) and 25 percent NextDay-12 (retail pharmacy network). Reading taken at 09:00 GMT on 12 Mar 26 for the trailing 30 days (10 Feb 26 to 11 Mar 26).
Exception codeConsignments flagged% of totalMost common day
CARDED (recipient absent)411.46%Friday
INVALID-ADDRESS180.64%Tuesday
ACCESS-DENIED (gated practices)140.50%Monday
REFUSED40.14%various
DAMAGE-REPORT30.11%various
MISROUTE20.07%various
Total exceptions822.93%
The card reads 2.93 percent on the dial, the alert at >3% is just clear (1 more exception lands amber). Five things to notice:
  1. CARDED dominates the count, and Friday dominates CARDED. Forty-one carded consignments, of which 26 were on Fridays, where surgeries close at lunchtime and the driver finds nobody to sign. The action is upstream: change the booking process to flag Friday clinic deliveries for delivery before 12:00 (move them to NextDay-9am), or hold back to Monday. This is a workflow fix, not a carrier fix.
  2. INVALID-ADDRESS at 0.64 percent points to a data-entry problem, not a postcode-database problem. APC’s depot scanners now use real-time PAF (Postcode Address File) lookup; a parcel reaching INVALID-ADDRESS state has usually had a manual address override at booking. Spot-check the 18 consignments against the order source: if 12 came from a single staff terminal, retrain.
  3. ACCESS-DENIED for medical premises is fixable with delivery instructions. Fourteen exceptions where the practice is behind a gated entry or a coded door. APC’s API supports a deliveryInstructions field per consignment; populating it with gate codes drops the access-denied rate by 70 to 90 percent. Pair this number with the dispatch system’s “delivery notes” usage rate.
  4. The 2.93 percent headline is just under the alert. Premium UK express benchmark is 0.8 to 1.5 percent for an established merchant; this brand is running roughly 2x the peer median. Three months ago the same number was 1.7 percent; the upward trend is the real signal, not the absolute level.
  5. Cross-reference with OTD Rate. This brand’s OTD reads 96.4 percent on the same window. Most of the carded and access-denied consignments did eventually deliver (often next day after the carded scan), so the lateness from this card is partially absorbed by APC’s free-redelivery network. If exception rate doubled but OTD only dropped 1 point, the network is recovering most of the friction; if both move together the network is failing.

​
Sibling cards merchants should reference together

Exception rate is a leading indicator. Pair it with these to act before the friction becomes a customer complaint:
CardWhy pair it with Exception RateWhat the combination tells you
On-Time Delivery RateExceptions are upstream of OTD misses. A rising exception rate predicts a falling OTD rate at 24 to 48 hours lag.If exception rate climbs but OTD holds, APC’s redelivery network is absorbing the friction. If both fall together, the network is failing.
Late ShipmentsEach CARDED, REFUSED, or ACCESS-DENIED event typically converts to a late shipment if not resolved same-day.Exception count - successful redeliveries roughly equals net new late shipments.
Failed DeliveriesFinal-state failures are a subset of exception events.Exception rate minus failed-delivery rate equals “rescued by redelivery” share, the size of APC’s recovery network for this merchant.
Open ClaimsDAMAGE-REPORT exceptions are direct claim candidates.A spike in DAMAGE-REPORT predicts a 1 to 2 week claim filing surge.
OTD by Origin DepotExceptions are often depot-routing or driver-allocation issues at one or two specific depots.Per-depot exception rate identifies the depot to escalate to APC’s account team.
Cross-connector: shopify.address_validityUpstream cause for INVALID-ADDRESS exceptions.If Shopify checkout is not running address validation, INVALID-ADDRESS exceptions on APC will track the rate of typos at checkout.
Cross-connector: customer-service ticket volume (Zendesk, Gorgias)Downstream impact. CARDED parcels generate “where is my order” tickets at roughly 30 to 50 percent of the carded count.A 50-event CARDED week predicts 15 to 25 WISMO tickets in the following 3 to 5 days.
Cross-connector: shipbob.sb_otd_rateAdjacent 3PL view. ShipBob’s exception rate covers warehouse-floor exceptions; this card is carrier-only.Different stages of the same fulfilment lane; not directly comparable.

​
Reconciling against the vendor’s own dashboard

Where to look in APC’s own portal: APC Overnight Customer Portal → Track & Trace → Exception Reports. Filter Date Range = Last 30 Days, Exception Type = All. The portal exposes per-event-code counts; these should reconcile with the breakdown surfaced in this card. For account-managed customers, APC’s QBR includes a depot-by-depot exception breakdown plus driver-cause attribution. The portal aggregate matches the card; the QBR slices are richer. Why our number may legitimately differ from APC’s portal:
ReasonDirectionWhy
Driver discretionary codesEitherAPC’s drivers occasionally use a generic “delivery delayed” comment code that APC’s portal sometimes counts as exception, sometimes does not, depending on the report tab. Vortex IQ counts it as exception.
Multi-event consolidationOurs lowerA consignment that throws CARDED then REFUSED counts once in this card; APC’s portal sometimes counts each event separately in its events-table view. Reconcile against the consignment-level exception filter, not events-level.
Time-zone of the reportBoundary days offAPC’s portal uses UK local; Vortex IQ also UK local. The 30-day window boundary may differ by a day.
Globe export consignmentsAPC may excludeInternational export via APC’s Globe network often appears in a separate report on APC’s portal. Vortex IQ pools UK domestic and Globe export into the same denominator.
Webhook lagOurs lower for “today”Exception-event webhooks lag depot upload by 30 minutes to 4 hours during peak. The card excludes events not yet ingested; APC’s portal sometimes shows the same lag, sometimes does not.
Definition of “any exception”APC may be narrowerAPC’s headline exception number on the QBR is sometimes restricted to “service-failure” codes (CARDED, REFUSED, INVALID-ADDRESS) and excludes operational codes (MISROUTE, DAMAGE-REPORT). Vortex IQ pools all exception codes.
Internal identity (within APC): apc_exception_rate >= apc_failed_delivery_rate. Every failed delivery threw at least one exception, but most exceptions resolve via redelivery. If the two numbers ever cross, there is an ingestion bug. Cross-connector reconciliation:
CardExpected relationshipCauses of legitimate divergence
shopify.address_validityUpstream input. INVALID-ADDRESS exceptions on APC track Shopify’s address-typo rate at checkout.Address-correction services (Loqate, AfterShip) at checkout substantially decouple the two.
royal_mail.roy_exception_rateCarrier-peer comparison only. Different parcel populations, different exception code taxonomy.Use for documentation navigation; not a like-for-like reconciliation.
Documentation cross-reference (for agencies running multiple UK shippers). Exception-rate metrics exist with conceptually similar definitions across other UK carriers and 3PLs. These are not parallel measurements of the same shipments; the populations are independent.

​
Known limitations / merchant FAQs

Exception rate is rising but OTD is steady, why? APC’s redelivery network is absorbing the friction. Roughly 60 to 80 percent of CARDED and ACCESS-DENIED exceptions deliver successfully on the next attempt, often the same week. As long as the second attempt lands by the original promised date, the consignment counts as on-time on OTD Rate but exception on this card. The leading-indicator value of this card is precisely that it moves before OTD does. Treat a 1-point exception-rate rise as a 0.3-point OTD rise pencilled in for the following week unless action is taken. What is the difference between a CARDED parcel and a FAILED parcel? CARDED means the driver attempted delivery and left a calling card; the parcel returns to the depot for redelivery (free, normally next working day). FAILED means the consignment exhausted attempts (typically three) without successful delivery and is now waiting for instructions (return-to-sender, customer collection, or further redelivery via a paid request). CARDED is a friction signal; FAILED is a near-loss. Should I aim for zero exception rate? No. A meaningful share of exceptions are recipient-side (CARDED, REFUSED, ACCESS-DENIED) and structurally outside the merchant’s control. The healthy steady-state for a UK B2C book on APC NextDay is 0.8 to 1.5 percent; for B2B clinic-and-surgery deliveries it is 1.5 to 2.5 percent (more access-control complexity). Aim to keep the trend flat and the cause-mix sensible, not the headline at zero. My exception rate spiked yesterday, what do I check first? In order: (1) APC’s service alerts in the customer portal, depot-level disruption produces uniform exception spikes regardless of cause-code. (2) The dominant cause-code in the spike, if CARDED, the day-of-week is the answer (Friday is the typical culprit on B2B routes). If INVALID-ADDRESS, check whether a recent CSV-import of orders introduced bad data. If ACCESS-DENIED, check whether a B2B customer changed gates or codes. (3) Geography, if the spike concentrates in two postcodes, the issue is local; if uniform, it is network-wide. Why is my INVALID-ADDRESS rate higher on B2B than B2C? Counter-intuitive but consistent across the network: B2C addresses pass through Shopify / WooCommerce checkout with PAF lookup and rarely produce invalid addresses (modern checkout libraries catch typos). B2B addresses come from spreadsheet imports, EDI orders, and account-team-keyed bookings, where PAF validation is often skipped or run loosely. Tighten the upstream booking-system validation; do not blame APC. Can exceptions be removed from the count if they were customer fault? The card surfaces every exception event regardless of cause; that is by design (the customer experiences the friction whether it was their fault or not). For QBR conversations with APC’s account team, filter the cause-code breakdown to exclude REFUSED and ACCESS-DENIED on B2C addresses; APC will accept those as out-of-network. CARDED on a B2B medical clinic where the surgery published wrong opening hours is a process fix on the merchant side, not an APC service-credit candidate. Why does APC’s portal show a different exception rate? Two usual reasons. (1) Different denominator APC’s portal often divides exceptions by attempted deliveries, Vortex IQ divides by all consignments in the period (delivered + failed + RTS). The portal number is therefore systematically higher. (2) Different code list APC’s portal sometimes uses a narrower “service-failure exceptions” subset that excludes MISROUTE and DAMAGE-REPORT. Reconcile by toggling the portal’s “all event types” filter. How does this interact with insurance and damage claims? DAMAGE-REPORT exceptions on this card flow through to Open Claims and Claim Value (open). Each DAMAGE-REPORT is a candidate for an APC carriage refund or insurance claim, depending on whether the merchant carries APC’s enhanced liability cover. The card does not auto-flag claim-eligible exceptions; that is a manual workflow.

​
Tracked live in Vortex IQ Nerve Centre

Exception Rate is one of hundreds of KPI pulses Vortex IQ tracks across APC Overnight and 70+ other ecommerce connectors. Nerve Centre runs the detection layer; Vortex Mind investigates the cause when something moves; Ask Viq lets you interrogate any number in plain English. Start for free or book a demo to see this metric running on your own data.