> ## Documentation Index > Fetch the complete documentation index at: https://docs.vortexiq.ai/llms.txt > Use this file to discover all available pages before exploring further. # Late-Delivery Revenue at Risk, DPDLocal > Estimated revenue exposed to refund / chargeback / churn from late deliveries this week. Key Metric - the finance-CFO wake-up call. How to read it, why it matter... **Metrics type:** [Cross-Platform Metrics](/nerve-centre/overview#metrics-types-explained) • **Category:** [Shipping & Courier](/nerve-centre/connectors#connectors-by-type) > Estimated revenue exposed to refund / chargeback / churn from late deliveries this week. Key Metric - the finance-CFO wake-up call. ## At a glance > Estimated revenue exposed to refund, chargeback, and churn from late DPDLocal deliveries this week. Joins the late-shipment list to commerce-sibling order totals and applies a refund-propensity multiplier per merchant baseline. The CFO's wake-up call number, the moment OTD becomes a finance metric. | | | | ------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **What it counts** | `SUM(commerce_sibling.order.total) WHERE order_ref IN (dpdlocal.shipment WHERE actual_delivery > expected_delivery + 24h) × refund_propensity_pct`. The "+24h" buffer excludes mildly-late deliveries that rarely refund. | | **Refund propensity** | Per-merchant baseline computed from the trailing 90-day refund-rate-on-late-cohort vs all-cohort refund rate. Default 12 to 25% for fashion DTC, 5 to 10% for supplements / health, 2 to 5% for B2B. Configurable per workspace. | | **Currency** | Merchant base currency (typically GBP for DPDLocal-native UK merchants). | | **What it excludes** | Mildly-late deliveries (within 24h grace), B2B contract orders with no refund right, and orders already refunded (those are realised, not at-risk). | | **`only_when: has_commerce_sibling`** | Card only renders when a commerce connector (Shopify / BC / Adobe / SFCC) is also live. | | **Time window** | `7D` (the actionable window, the CFO's "this week" view) | | **Alert trigger** | `>£1000`. Default; tune per merchant size. | | **Roles** | owner, finance, marketing | ## Calculation Calculated automatically from your DPDLocal 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 fashion DTC merchant on Shopify, AOV £85, refund propensity baseline 18% on late-cohort orders. Reading 12 Mar 26, trailing 7 days. | Late shipment cohort | Count | Sum order\_total | Refund propensity | Revenue at risk | | ----------------------------- | ------- | ---------------- | ------------------------ | --------------- | | Late by 24-48h | 142 | £12,070 | 18% | £2,173 | | Late by 48-72h | 31 | £2,635 | 28% (raised on severity) | £738 | | Late by 72h+ | 9 | £765 | 45% | £344 | | **Total at risk (this card)** | **182** | **£15,470** | weighted \~21% | **£3,255** | The card reads **£3,255**. The alert at `>£1000` is firing. Three things to notice: 1. **The CFO sees a £3,255 number, not a 96% OTD percentage.** This is the right framing for finance. The OTD card is for operations; this card is for the CFO and the finance partner. 2. **Severity-weighted propensity matters.** Parcels late by 72+ hours refund at roughly 2.5x the rate of 24-48h-late parcels; the simple flat propensity understates the risk on the worst tail. 3. **The pre-emptive intervention budget is £500 to £900.** Even a 30% reduction in refund rate from the late cohort (via "we know your parcel is late, here's a £10 voucher" outreach via Klaviyo) saves more than the voucher cost on this exposure. ## Sibling cards merchants should reference together | Card | Why pair | | ----------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------- | | [Late Shipments](/nerve-centre/kpi-cards/dpdlocal/late-shipments) | The count behind the £ exposure. | | [On-Time Delivery Rate](/nerve-centre/kpi-cards/dpdlocal/on-time-delivery-rate) | The percentage view; aligns operations and finance frames. | | [Carrier OTD by Sales Channel](/nerve-centre/kpi-cards/dpdlocal/carrier-otd-by-sales-channel) | Splits the £ exposure by channel. | | [Claim Value as % of DPDLocal Revenue](/nerve-centre/kpi-cards/dpdlocal/claim-value-as-of-dpdlocal-revenue) | Realised cost on the carrier; this card is the at-risk forecast. | | Cross-connector: [`shopify.refund_rate`](/nerve-centre/kpi-cards/shopify/refund-rate) | Validates the refund-propensity baseline. | ## Reconciling against the vendor's own dashboard **Where to look:** This is a Vortex IQ-derived value. DPDLocal does not produce it (DPDLocal does not see commerce-side order totals); the commerce platform produces refund-rate-on-late-cohort but does not see the DPDLocal late list. The join is the card. **Why retroactive reality may differ from the at-risk forecast:** | Reason | Direction | Why | | ------------------------------------ | -------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Refund propensity baseline drift** | Either | The 18% (or per-merchant) baseline is from trailing 90 days; if customer behaviour has shifted, the forecast under- or over-shoots. Re-baseline quarterly. | | **CS intervention** | Realised lower | Pre-emptive outreach typically cuts realised refund rate by 25 to 45%; the at-risk forecast assumes no intervention. The gap between forecast and realised is the value of the CS programme. | | **Order cohort heterogeneity** | Either | A single high-AOV order in the late list dominates the £ exposure; cohort variance can be high week to week. | **Cross-connector reconciliation:** | Card | Expected relationship | | ------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Realised refund value attributable to late cohort (commerce-sibling, lagged 14 days) | Should land within ±30% of this card's forecast. Persistent under-realisation suggests CS intervention is working; persistent over-realisation suggests baseline propensity needs to be raised. | ## Known limitations / merchant FAQs **Why use a refund-propensity multiplier instead of just summing the late-cohort revenue?** Not every late delivery causes a refund; merchants who refund every late parcel would have a far worse refund rate than they actually do. The propensity adjusts for the fact that customer tolerance, severity, and CS intervention all dampen realisation. Without the multiplier the card would dramatically overstate. **The propensity baseline doesn't match my finance team's number.** Adjust it. The default baseline is computed from your own trailing-90-day data; if you have an internal CFO model with a different rate, configure it in the connector. The card is meant to surface the magnitude, not be the source of truth for finance. **Should I include B2B orders in this calculation?** Default no. B2B trade orders typically have contracted delivery windows with no consumer refund right; including them inflates the at-risk number meaninglessly. The default filter excludes B2B-tagged orders; toggle on if your B2B contracts include service-credit clauses. **How does the card change with a CS intervention programme running?** The forecast number stays the same (it's the unmitigated risk); the realised number drops 25 to 45%. Compare forecast vs realised over 30 days to quantify the programme's value. **The number jumps up and down weekly. Why?** Two reasons. (1) **Cohort heterogeneity.** A single £500 high-AOV late delivery moves the number more than 50 small ones. (2) **Severity distribution.** A week with mostly 24-48h-late vs a week with mostly 72h+-late produces very different propensities even at similar counts. Read the trend over 4 weeks rather than week-to-week. **Can I drill into the late-cohort orders?** Yes, click any segment of the card to see the underlying order list with order\_ref, customer email, and the specific late-by-X-hours measurement. Useful for CS pre-emptive-outreach lists. *** ### Tracked live in Vortex IQ Nerve Centre *Late-Delivery Revenue at Risk* is one of hundreds of KPI pulses Vortex IQ tracks across DPDLocal 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](https://app.vortexiq.ai/login) or [book a demo](https://www.vortexiq.ai/contact-us) to see this metric running on your own data.