At a glance
Absolute count of Deutsche Post / DHL Paket consignments that arrived after the agreed delivery promise in the last 7 days. The companion to On-Time Delivery Rate. OTD is the percentage you put in front of the board; this is the ticket workload that hits Customer Service. Deutsche Post / DHL Paket is one of the most reliable parcel networks in Europe, so the late count is normally a small number; the value is in the trend and the concentration, not the absolute level.
| What it counts | COUNT(shipments WHERE delivered_at > targetDeliveryDate AND delivered_at IS NOT NULL) over the trailing 7 days, computed at delivered-scan time. |
| API endpoint | DHL Tracking API GET /track/shipments?shipmentNumber={x}, joined to the booking-time targetDeliveryDate from the original POST /parcel/de/shipping/v2/orders response. The card stores both timestamps and computes the late-flag at delivery time. |
| Late definition | Delivered after midnight Berlin time on the DHL-promised date for the service code. There is no grace period. DHL offers service-credit conversations only on Vertragskunden contracts (negotiated SLAs); the card counts every miss, not just credit-eligible ones. |
| Service-tier scope | All German-domestic DHL Paket and Päckchen products with a delivered scan: V01PAK Paket, V62KP Päckchen, V01PAK_S Paket Pro Sunday, V01PRIO Paket priority. Letter products (Standardbrief, Maxibrief) are excluded; international DHL Paket / DHL Express are on separate connectors. |
| Returns / RTO | Outbound only. DHL Retoure consignments are filtered out. RTO volume is on Returned to Sender. |
| In-transit handling | A consignment past the promise date that has not yet been delivered is not counted as late; the card waits for a delivered scan. For forward-looking risk see Exception Rate. |
| Industrial action / weather | Not auto-excluded. Verdi-led postal strikes (rare but periodic during pay-bargaining cycles) and severe weather events (Schneechaos, river-flood disruption) are recorded as actual count without smoothing; merchants should annotate any board-level reads during declared incidents. |
| Currency | The card is a count, not a value. To estimate cost (CS time + remediation refunds + reship), join with Avg Shipping Cost and an internal CS-cost-per-ticket assumption. |
| Time window | 7D (rolling 7-day count). 30D and 90D variants accessible via the time-window control. |
| Alert trigger | >5% of total. For a German-domestic DHL Paket merchant this typically fires only during declared incidents; routine weeks read 1 to 3 percent. A 5 percent reading on Deutsche Post is a meaningful incident signal. |
| Roles | owner, operations |
Calculation
Calculated automatically from your Deutsche 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
The Düsseldorf-based DTC homeware merchant from the OTD card, around 6,500 outbound parcels per week. Reading taken at 09:00 CET on Monday 09 Mar 26 for the trailing 7 days (02 Mar 26 to 08 Mar 26).| Product | Delivered last 7d | Late count | Late % | Avg days over promise |
|---|---|---|---|---|
| V01PAK DHL Paket | 5,250 | 124 | 2.4% | 1.2 |
| V62KP DHL Päckchen | 780 | 41 | 5.3% | 1.8 |
| V01PAK_S DHL Paket Pro Sunday | 390 | 4 | 1.0% | 0.9 |
| All Deutsche Post tracked (this card) | 6,420 | 169 | 2.6% | 1.3 |
>5% is not tripped on the headline. Five things to notice:
- The headline is healthy but Päckchen is over the threshold per-product. 5.3 percent late on Päckchen would normally be concerning; on this product the wider promise window (E+2 to E+4) means the card uses the latest day of the window as the on-time deadline, so a 5.3 percent reading actually reflects parcels arriving day 5 or later. That is a real customer-facing failure even if the contractual interpretation is generous.
- The headline count is the CS staffing number. 169 late parcels = roughly 170 to 240 CS interactions over the next 14 days. At 6 minutes per ticket that is 17 to 24 hours of CS time. Use this when deciding Monday-morning shift coverage.
- Sunday-delivery is the most reliable product on the dial. 1.0 percent late on Paket Pro Sunday confirms the network’s premium-tier handling; if Sunday-delivery rises above 2 percent late, it is a genuine network signal worth chasing.
- The 1.3 average days over promise is the “slipped a day” tail, not the lost-in-transit tail. Lost-in-transit consignments (delivery date 14+ days after booking with no delivered scan) are surfaced via Exception Rate, not this card.
- The week-on-week trend is the actionable read. Last week 152, the week before 158, the week before that 145. The 169 reading is a 11 percent week-on-week jump driven by Päckchen; pair with Cost by Zone and the rural-postcode breakdown to determine whether it is a structural shift or a one-week tail.
Sibling cards merchants should reference together
Late Shipments is a count, not a percentage. Pair with these to read the count in context:| Card | Why pair it with Late Shipments | What the combination tells you |
|---|---|---|
| On-Time Delivery Rate | The percentage form of the same data. | OTD is the board-level number; this card is the operations workload. |
| Exception Rate | Forward-looking signal. | A spike in exceptions with a flat late-count usually means the next 24 to 72 hours will see late-count climb. |
| Failed Deliveries | Subset where the parcel could not be handed over at all. | Late + failed = the full customer-felt failure population. |
| Returned to Sender | Consignments that became RTO after multiple attempts. | A small fraction of late shipments graduate to RTO; chasing late on day +1 prevents RTO on day +14. |
| OTD by Route | Where the late count is concentrated. | Most late counts on Deutsche Post are concentrated in 2 to 5 rural Postleitzahlen routes (East Frisia, Eifel, Bavarian rural). |
Cross-connector: shopify.fulfillment_speed | Upstream cause. Slow warehouse pick-and-pack feeds late deliveries. | If warehouse-to-DHL handover creeps above 24 hours, the late count climbs even with a perfect carrier. |
Cross-connector: gorgias.tickets_open | Downstream impact. Late deliveries drive WISMO tickets at 1.3 to 1.8x ratio in Germany (slightly higher than other markets because German consumers expect Deutsche Post reliability). | A late-count spike on Monday morning typically shows up as Gorgias spike on Tuesday afternoon. |
Cross-connector: klaviyo.transactional_email_delivery | Defensive comms lever. Proactive “your parcel is running late” emails reduce ticket inflow by 40 to 60 percent. | German consumers respond particularly well to proactive comms; “Sorry, your DHL parcel is delayed” is culturally expected. |
Reconciling against the vendor’s own dashboard
Where to look in Deutsche Post’s own portal: DHL Geschäftskundenportal → Berichte → Service Performance → Detailansicht, then filter to Verspätete Sendungen and the trailing 7 days. The detailed view returns row-per-consignment and the headline count should match this card within the lag tolerance noted below. Vertragskunden contracts also see the monthly Quality Report PDF which is the authoritative count for service-credit conversations. Why our number may legitimately differ from the DHL portal:| Reason | Direction | Why |
|---|---|---|
| Timezone (CET / CEST vs UTC) | Off by 1 day at boundary | Portal counts in Berlin local time; the card stores in UTC. A delivery completed at 23:30 CET on Sunday is “Sunday” in the portal and could be “Monday” in the card. |
| In-transit consignments | Portal sometimes higher | The portal occasionally counts past-promise in-transit consignments as late; the card waits for the delivered scan and only counts confirmed-late deliveries. |
| Päckchen promise-window interpretation | Either | Päckchen’s E+2 to E+4 window can be interpreted as “on-time if delivered any day in the window” or “late if delivered on day 5 or later”; the card uses the strict interpretation (latest day of window = on-time deadline). The portal sometimes uses a softer interpretation. |
| Service-credit-eligible vs all late | Portal lower if filtered | The Quality Report PDF for Vertragskunden is sometimes pre-filtered to credit-eligible consignments; the card reads every miss. |
| Returns inclusion | Ours lower | Portal default sometimes includes return-leg late deliveries; the card excludes them. |
| Card | Expected relationship | Causes of legitimate divergence |
|---|---|---|
gorgias.tickets_open | Late deliveries drive WISMO tickets at 1.3 to 1.8x ratio in Germany. | Not every late ticket relates to a DHL shipment; the merchant has other carriers. |
shopify.fulfillment_speed | Slow warehouse-to-DHL handover feeds the late count. | Warehouse-leg lag is upstream; once handed to DHL the carrier-leg performance is what this card measures. |
hermes_germany.her_late_shipments_count | Adjacent German parcel-network late count for the same period. | Different carrier, different shipments; cross-compare to inform carrier mix. |
Known limitations / merchant FAQs
Why does my late-count climb every Monday morning? Two reasons. (1) Weekend tracking lag. DHL’s residential weekend network is reduced; parcels missing Friday’s promise show up as late on Monday’s first scan. (2) The Friday-evening booking surge. German DTC orders concentrate Thursday-Friday; those parcels enter DHL’s induction Monday and any minor sortation backlog shows up on Monday-arriving consignments first. A late parcel from December just appeared in this week’s count. Why? The card filters ondelivered_at (the date the recipient actually got the parcel), not targetDeliveryDate. A parcel booked in December and finally delivered in March (typical for held-at-Packstation undelivered, Nachsendeantrag-redirected, or held-at-Postfiliale) shows up in March’s count when the delivered scan finally lands. Pair with Avg Transit (days) to identify outlier-aged consignments.
My count keeps oscillating between 130 and 220 weekly with no business change. Why?
Volume noise. At 6,500 weekly parcels and a 2.5 percent late rate, expected weekly count is around 162 with standard deviation of around 35 parcels (poisson approximation). Anything between 90 and 230 is statistical noise. Use 4-week rolling averages for trend, raw weekly counts only for outage detection.
The card says 169 late; my CS team logged 280 WISMO tickets. Why the gap?
WISMO tickets do not map 1:1 to late shipments in Germany. Drivers: (1) German consumers contact CS more proactively than other markets, often before the parcel is technically late. (2) Tickets opened about Packstation collection failures (parcel arrived but recipient cannot collect) feel like “late” but are technically successfully delivered. (3) Tickets about Standardbrief / Maxibrief (excluded from this card) get logged in the same WISMO bucket. Healthy ratio is 1.0 to 1.8x.
My OTD says 97 percent and the count says 169 late. Are these consistent?
Check the math. 6,420 delivered minus 169 late = 6,251 on-time. 6,251 / 6,420 = 97.4 percent. The two numbers should reconcile within rounding. If they differ by more than 1 percentage point, time windows differ (OTD is typically 30D, this card is 7D) or carrier population differs.
How do German postal strikes affect this count?
Verdi-led postal strikes during DHL pay-bargaining cycles (typically March / April every 2 to 3 years) can spike late counts 3 to 8x for affected weeks. The card does not auto-smooth; the count goes up. Three actions during declared strikes: (1) update checkout copy to set later expectations, (2) shift volume to Hermes / DPD / GLS where possible, (3) re-baseline alert thresholds for the strike window only, reset when service resumes.
Will DHL credit me for these late consignments?
Some of them, only on Vertragskunden contracts. Standard Online-Frankierung tariff parcels are aim-only and do not carry SLA-credit. Vertragskunden Paket Pro and Paket Premium have negotiated SLAs and credit-eligible miss thresholds. The card surfaces every miss; the credit-eligible subset depends on your contract terms. File credits within 30 days of the missed promise via the Geschäftskundenportal’s Reklamationsstelle.
Should I exclude Päckchen from the count? The wider promise window inflates the late percentage.
Reasonable choice if your business case requires Paket-only reliability. The card supports per-product filtering at the alert layer; configure the alert to fire only on V01PAK (Paket) for a more conservative reading. Do not delete Päckchen from the underlying data; it is real customer-felt experience.
My Geschäftskundenportal shows 142 late, the card shows 169. Why the gap?
Most likely the portal applies a softer interpretation of the Päckchen promise window (counts as on-time if delivered on the latest day of the window only when “delivery attempted on day 4” event is logged, which excludes some genuine misses). The card uses the strict interpretation. Both are defensible; for service-credit conversations use the portal’s number, for operational triage use the card’s number.
My CS team wants a webhook for new late shipments, not periodic dashboard reads.
Yes. Configure the DHL webhook integration to push tracking events to your CS tool with eventCode = delivered AND delivered_at > targetDeliveryDate as the filter. Pair with Klaviyo Transactional Email for proactive customer-side comms.