Most brands know their shipping and returns cost too much and cannot say where. The invoices are split across carriers, apps and a 3PL, the support queue absorbs the failures, and the returns leak shows up as a refund line nobody owns. An audit is the exercise that puts one number on each leak. This is the checklist we use when we run a shipping and returns audit for a brand: fifteen checks, the number each one produces, and what good looks like. You can run it yourself from your order export and carrier invoices, or have it run for you on your own orders (the last section).
Short answer: Audit five stages in order: checkout (are you promising the right delivery options and dates), carrier (are you paying for the service you get), warehouse handover (how long from order to first carrier scan), tracking (how much of your support queue is "where is my order") and returns (how many come back, why, and how many you keep as exchanges). Each stage has three checks. The three that find the most money in most audits are on-time delivery by carrier and lane, the order-to-scan handover time, and the exchange share of returns.
Before you start: the data you need
One export of the last 90 days of orders with shipping method, carrier, ship date and delivery date. Your carrier invoices for the same period. A dump of support tickets tagged by reason, or a keyword search for "where is my order", "tracking" and "return". Your returns log with reason codes and outcome (refund, exchange, credit). Ninety days is enough to see the pattern and short enough to reflect your current stack.
Stage 1: Checkout
1. Delivery options shown vs options available. Count the delivery methods a customer sees at checkout by destination country and compare with what your carriers can actually do there (home, pickup point, locker, express). Every missing option is conversion left on the table; pickup points in particular convert in markets where lockers are normal. Good: every market shows at least a home and a pickup option with a price and a date.
2. Promised date vs delivered date. For each shipping method, compare the date shown at checkout with the actual delivery date. Compute the share delivered on or before the promise. Good: above 90% per method. A promise that is right 70% of the time is worse than no promise, because the misses become tickets. The mechanics are in what is delivery promise.
3. Shipping price tests. Have you ever tested the shipping price or the free-shipping threshold by traffic split? If not, the number is unknown, which is itself the finding. Good: at least one controlled test in the last year with a measured effect on conversion and average order value.
Stage 2: Carrier
4. On-time rate by carrier and lane. From ship date to delivery date, by carrier and by origin-destination pair. This is the single most useful table in the audit, because it shows which carrier deserves which lane. Good: you can name your best and worst lane and the gap between them. Brands often find one carrier is excellent domestically and poor on one export lane they never looked at separately.
5. Rate paid vs service received. Match invoiced service levels to the on-time table. Paying express rates on a lane where the economy service arrives the same day is common and invisible until the two are put side by side. Good: no lane where a cheaper service would have met the promise more than 90% of the time. The routing logic that fixes this is in what is multi-carrier shipping.
6. Failed, lost and damaged rate. Share of shipments with a failed delivery attempt, a lost-in-transit claim or a damage claim, by carrier. Each failed attempt is a re-delivery cost and a ticket; each loss is a replacement order. Good: under 2% failed attempts domestically, and every loss claim filed within the carrier's window rather than written off.
Stage 3: Warehouse handover
7. Order-to-scan time. Hours from order placed to the first carrier scan, as a distribution, not an average. The tail matters: orders that sit two days before scanning are the ones that miss the promise and generate the first "where is my order" ticket. Good: 90% of orders scanned within one business day of placement, and a visible reason for every order in the tail.
8. Label and dispatch errors. Count re-printed labels, address corrections after dispatch and carrier surcharges for wrong dimensions or weights. Good: address validation before the label, dimensions from the packing station, and surcharges under 1% of shipments. The bench-side view of this is in pick and pack software.
9. Wrong-item shipments. Share of orders where the customer received the wrong product or size. Each one costs an outbound parcel, a return label and a replacement, and often the customer. Good: scan verification at the bench so the number is close to zero, and you can prove it.
Stage 4: Tracking and support
10. WISMO share of tickets. The share of support tickets that are "where is my order" or "when will it arrive". Compute it from tags or a keyword search. Good: under 15% of tickets. Above 30% means the tracking experience is doing no work. The full economics are in what is WISMO.
11. Proactive notification coverage. For the shipments that were delayed (from check 2), how many customers received a message about the delay before they wrote in? Good: every delayed shipment triggers a message, automatically, with the new date. If the answer is "the carrier sends emails", the answer is none, because the carrier's email does not know your promise.
12. Tracking page ownership. Do customers track on your domain, with your products and your support entry points, or on the carrier's page? A carrier page is a dead end for the brand and a confusing one for the customer once two carriers are involved. Good: one branded tracking page across every carrier, with the return flow one click away.
Stage 5: Returns
13. Return rate by reason. Returns as a share of orders, split by reason code and by product. The point is not the headline rate but the concentration: a few products and a few reasons usually carry most of it, and some of those are fixable upstream (sizing information, photos, packaging). Good: reason codes on more than 90% of returns and a monthly review of the top five.
14. Exchange and credit share. Of the returns, how many end as an exchange or store credit rather than a refund? This is where the money is. Switch Nails, a Nordic press-on nail brand running returns on Pango, keeps 19% of returns as an exchange or store credit, up from zero before; 33% of customers who exchange place another order, against 20% of those who take a refund, and one in ten exchangers spent more than they were owed. The figures are in the Switch Nails case study. Good: a rising exchange share and an exchange path that offers any product, not only the same item in another size.
15. Self-serve rate and refund timing. Share of returns completed without a support agent, and days from return scan to refund. Switch Nails runs 99% of returns self-serve. Every manual return is a ticket; every slow refund is a second ticket. Good: above 90% self-serve, refunds triggered by the carrier scan rather than the warehouse count, and cross-border returns handled with the customs paperwork reversed per country. The wider cost model is in the returns leak calculator.
The audit in one table
| # | Check | Number it produces | Good looks like |
|---|---|---|---|
| 1 | Delivery options shown | Options per market vs available | Home plus pickup everywhere, priced and dated |
| 2 | Promise accuracy | On-time vs promise by method | Above 90% |
| 3 | Shipping price tests | Tests run, measured effect | At least one controlled test a year |
| 4 | On-time by carrier and lane | Table by carrier and lane | Best and worst lane named |
| 5 | Rate vs service | Lanes overpaying for service | None |
| 6 | Failed, lost, damaged | Rate by carrier | Under 2% failed attempts, every claim filed |
| 7 | Order-to-scan time | Distribution in hours | 90% within one business day |
| 8 | Label and dispatch errors | Reprints, corrections, surcharges | Surcharges under 1% |
| 9 | Wrong-item shipments | Share of orders | Near zero, scan-verified |
| 10 | WISMO share | Share of tickets | Under 15% |
| 11 | Proactive notifications | Delayed shipments messaged first | All of them |
| 12 | Tracking page ownership | Own domain vs carrier page | One branded page, all carriers |
| 13 | Return rate by reason | Rate, reason, product concentration | Reason codes above 90% |
| 14 | Exchange and credit share | Share of returns kept | Rising, any-product exchanges |
| 15 | Self-serve and refund timing | Share self-serve, days to refund | Above 90%, scan-triggered refunds |
What the audit usually finds
Three findings repeat across brands of every size. First, one carrier lane is quietly failing while the blended on-time number looks fine. Second, the warehouse handover tail, not the carrier, causes most missed promises. Third, the returns flow refunds by default because exchanges are hard to offer, so money that could stay in the business leaves it. None of these show up on a dashboard that reports averages; all of them show up the moment the data is split the way this checklist splits it.
Get the audit run on your own orders
Pango runs this audit for brands free of charge, on their own order and carrier data, as part of a demo. You get the fifteen numbers, the three that matter most for your operation, and a view of the same orders running on one system: delivery promise at checkout, carrier routing by lane, branded tracking with proactive messages, and exchange-first returns, with AI agents doing the routine work. Bring the 90-day export; the rest takes one call. Book the free shipping and returns audit.
The bottom line
A shipping and returns audit turns a vague sense of leakage into fifteen numbers, three of which usually explain most of the cost. Run it yourself from the exports, or have it run for free on your own orders and see the fixes live at the same time. Book the free audit.
