At a glance
Share of Evri parcels that threw at least one tracking exception event before final delivery. Exceptions on Evri include carded (“Attempted, no answer”), refused, address-invalid, courier no-show (“not loaded onto vehicle”), depot misroute, and the Evri-specific failure mode of ParcelShop pickup not occurring. A parcel that threw an exception but eventually delivered still counts here, the card surfaces operational friction.
| What it counts | COUNT(parcels WHERE any tracking_event.code IN exception_codes) / COUNT(parcels WHERE status IN ('DELIVERED','FAILED','RTS')). |
| Exception codes included | CARDED (driver attempted, no answer), REFUSED (recipient declined), ADDRESS_INVALID, NOT_LOADED (parcel did not make courier’s vehicle), MISROUTE (sortation hub error), PARCELSHOP_NOT_PICKED_UP (Evri-specific, sat at ParcelShop but courier did not collect), DAMAGED_IN_TRANSIT, LOST_AT_DEPOT. |
| API endpoint | Evri’s tracking API getTrackingDetails exposes the full event ladder per parcel; Vortex IQ matches event codes against Evri’s published exception taxonomy. |
| Service level scope | All outbound services pooled. Returns leg has separate exception measurement. |
| Multiple exceptions per parcel | One parcel counts once regardless of how many exception events fire. |
| Self-employed-courier model implications | Evri’s exception rate is structurally higher than salaried-driver carriers because courier no-shows, sick days, and turnover produce NOT_LOADED or carded scans. Healthy steady-state is 2 to 4 percent vs Royal Mail’s 1 to 2 percent. |
| Returns / RTO | RTS parcels are included in both numerator and denominator. |
| Time zone | UK local. |
| Time window | 30D vsP (rolling 30 days, period-over-period). |
| Alert trigger | >3% warn. Calibrated to Evri’s structural baseline. |
| Roles | owner, operations |
Calculation
Calculated automatically from your Evri (formerly Hermes UK) 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 DTC home-fragrance brand running Evri Standard for everyday consumer parcels, around 8,400 parcels per month, mixed urban / suburban / rural delivery. Reading taken at 09:00 BST on 12 Mar 26 for the trailing 30 days (10 Feb 26 to 11 Mar 26).| Exception code | Parcels flagged | % of total | Notes |
|---|---|---|---|
| CARDED (recipient absent) | 168 | 2.0% | Driver attempted, no answer, calling card left |
| NOT_LOADED (courier no-show) | 52 | 0.62% | Evri-specific. Self-employed courier did not pick up vehicle. |
| ADDRESS_INVALID | 45 | 0.54% | Customer-side address typos |
| PARCELSHOP_NOT_PICKED_UP | 21 | 0.25% | Evri-specific. ParcelShop drop-off not collected by courier |
| REFUSED | 14 | 0.17% | Recipient declined |
| MISROUTE | 9 | 0.11% | Depot sortation error |
| DAMAGED_IN_TRANSIT | 7 | 0.08% | Driver flagged damage |
| LOST_AT_DEPOT | 4 | 0.05% | Last-scan-at-depot, never re-emerged |
| Total exceptions | 320 | 3.81% | Above alert |
>3% is tripped. Five things to notice:
- CARDED at 2 percent dominates the count. Standard for Evri; the structurally-similar number on Royal Mail Tracked would be ~1.0 to 1.5 percent because Royal Mail’s salaried postman has a fixed local round and tends to know which houses are typically empty during the day. Evri’s rotating self-employed couriers do not build that local knowledge as efficiently. The fix is largely customer-side: add “leave with neighbour” instructions on the parcel via the Evri portal, allow customers to pre-set safe-place preferences.
- NOT_LOADED at 0.62 percent is the Evri-specific signal. It means a self-employed courier was scheduled to pick up the round, didn’t show, and the parcels sat at the depot until a backup courier could be assigned (next day). 0.62 percent is in line with Evri’s published baseline; above 1 percent indicates a depot-staffing crisis.
- PARCELSHOP_NOT_PICKED_UP at 0.25 percent is small but distinctive. This is a parcel sitting at a ParcelShop that the Evri pickup courier did not collect on the daily sweep. The fix is to identify which ParcelShops are involved (a single non-collected ParcelShop drives most of the count). Pair with ParcelShop Drop-off Failure Rate.
- LOST_AT_DEPOT at 0.05 percent (4 parcels) translates to lost-parcel claims. Each is a candidate for Lost-Parcel Rate. Persistent presence of LOST_AT_DEPOT is the trigger to escalate with Evri’s account team.
- The 3.81 percent headline is just above alert. Treat 2 to 4 percent as Evri’s normal range; above 4 percent for two weeks running is escalation territory. Rising trend is more important than absolute level.
Sibling cards merchants should reference together
Exception rate is a leading indicator. Pair with these:| Card | Why pair it | What the combination tells you |
|---|---|---|
| On-Time Delivery Rate | Exceptions predict OTD dips at 24-48 hour lag. | Rising exception rate without a corresponding OTD drop means Evri’s redelivery network is absorbing the friction; both moving means real degradation. |
| Late Shipments | Each carded or refused exception typically converts to late if not resolved same-day. | Exception count - successful redeliveries roughly equals net new lates. |
| ParcelShop Drop-off Failure Rate | The ParcelShop-specific exception subset. | Identifies which ParcelShops are unreliable. |
| Redelivery Attempts > 3 | Carded and refused exceptions accumulate redelivery attempts. | Each row past 3 attempts is on its way to RTS. |
| Lost-Parcel Rate | LOST_AT_DEPOT exceptions feed lost-parcel rate. | Persistent LOST_AT_DEPOT is the structural trigger. |
| OTD by Route | Postcode-level exception concentration. | Identifies courier-event hotspots. |
Cross-connector: shopify.address_validity | Upstream cause for ADDRESS_INVALID. | If Shopify is not running address validation at checkout, the rate of typos passes through to this card. |
| Cross-connector: customer-service ticket volume | Strong lagged correlation. | CARDED parcels generate WISMO tickets at 30-50 percent of carded count. |
Reconciling against the vendor’s own dashboard
Where to look in Evri’s own portal: Evri Business Portal → Reports → Exception Reports. Filter Date Range = Last 30 Days. The portal exposes per-event-code counts; reconcile against that breakdown. Why our number may legitimately differ from Evri’s portal:| Reason | Direction | Why |
|---|---|---|
| Multi-event consolidation | Ours lower | A parcel that throws CARDED then REFUSED counts once on this card; Evri’s portal sometimes counts each event separately. Reconcile against parcel-level filter. |
| NOT_LOADED classification | Either | Evri’s portal sometimes treats NOT_LOADED as an internal operational metric not exposed on the merchant view; this card surfaces it. |
| Driver scan timing | Ours stale | Mobile-app scanning lag in low-signal areas. |
| Returns leg | Either | Outbound only here; portal sometimes pools. |
her_exception_rate >= her_failed_delivery_rate. Every failed delivery threw at least one exception; most exceptions resolve via redelivery.
Cross-connector reconciliation:
| Card | Expected relationship | Causes of legitimate divergence |
|---|---|---|
shopify.address_validity | Upstream cause for ADDRESS_INVALID. | Address-correction services at checkout decouple. |
Documentation cross-reference (UK carrier peers).