If you're wondering why ecommerce deliveries fail, the answer is almost never just "the courier". Most orders that arrive late, wrong or not at all go wrong before they even leave the warehouse. The most common causes are a badly captured delivery detail, stock that doesn't add up, a picking error, packaging that doesn't protect the product, or a delivery promise the operation simply can't keep.
The driver only makes visible a fault that was already baked into the parcel. That's why, to reduce incidents, you first need to know at which point in the process each one originates. This article runs through the seven causes we see most often day to day and how to fix them. It includes an example with indicative figures and a checklist you can apply this week.
Why ecommerce deliveries fail: the short answer
Online store deliveries fail mainly due to errors that happen before transport: incomplete addresses, outdated stock, badly prepared orders and inadequate packaging. On top of that come the transport-related causes: the wrong courier chosen for the destination, promised timeframes the operation can't actually sustain, and a lack of follow-up when an incident arises. Almost all of these can be prevented with validated data and checks before the box is sealed.
The key is to stop lumping everything into the same bucket. If every complaint gets logged as a "shipping problem", you'll never know whether the fault lies in the checkout, on the shelf, or on the route.
What counts as a failed delivery (and what is ecommerce parcel delivery)
Ecommerce parcel delivery is the shipping of online store orders to the end customer: small or medium parcels, many different destinations, home deliveries or drop-off points, and a customer expecting real-time tracking. It's nothing like business-to-business shipping: here, every parcel is a purchase and every failure is felt by a person.
A failed delivery isn't just a lost parcel. From your customer's point of view, it's also:
- An order that arrives after the promised deadline.
- A parcel that gets sent back to origin because no one was able to receive it.
- The wrong product, an incomplete order, or the wrong size.
- A broken product or a crushed box.
- An order with no clear tracking, forcing the customer to message you to find out where it is.
For the customer, all of these produce the same thing: uncertainty. For you, it means support tickets, reshipments, returns, and a customer who thinks twice before ordering again.
The 7 most common logistics problems that cause a delivery to fail
1. Outdated stock: selling what you don't have
Selling a unit that isn't actually on the shelf is one of the costliest and most avoidable mistakes. It happens when the stock shown in the store doesn't match what's in the warehouse, when variants are badly configured, or when returns go back into stock without being checked.
It gets worse if you sell across several channels at once. A difference of just two units between Shopify and a marketplace is enough to oversell during a campaign. The customer has paid, is expecting confirmation, and instead gets a cancellation or delay email.
What to do:
- Make sure every goods-in, goods-out, adjustment and return updates stock in a single system that feeds all channels. We cover this step by step in the guide to syncing stock across channels without overselling.
- Separate physical stock from sellable stock: damaged items, those in quarantine or pending review shouldn't show as available.
- Leave a small safety margin on your fastest-moving SKUs during promotions.
2. Addresses and phone numbers captured incorrectly
Before we set up a warehouse, at Furgox we used to deliver parcels for CTT Express and Paack. From behind the wheel, it becomes obvious: a huge proportion of failed deliveries have nothing to do with the parcel itself, but with what the customer typed in at checkout. A missing flat number, a postcode that doesn't match the town, a phone number missing a digit, or a work address given with no company name.
The driver can't call, the parcel is logged as an incident, and if no one acts, it goes back to origin. You end up paying for the outbound trip, the return, and sometimes a second shipment.
What to do:
- Validate the postcode and phone number format at checkout.
- Make the flat/door number field mandatory, or at least very visible.
- Add an optional delivery instructions field.
- Manually check orders with unusual addresses (too short, no house number) before generating the label.
There's a balance to strike: asking for too much data lowers conversion, and asking for too little increases incidents. Look at your own history and adjust wherever it's failing most.
3. Picking without verification
The customer can't tell the difference between a picking error and a transport error. If they receive size M instead of L, as far as they're concerned the delivery has failed. And you pay twice: once for the shipment, once for the return.
Errors cluster around catalogues with similar products, colour variants, bundles and unclear SKU codes. And they multiply when lots of orders come in at once and the team is rushing.
What to do:
- Fixed, labelled locations for every SKU.
- Scan the barcode when picking the product, and again before sealing the box.
- Physically separate variants that get mixed up (don't put the black M t-shirt right next to the black L).
If you want the full method, see our guide on order-picking errors and how to avoid them. Speed without verification doesn't save money: it just shifts the cost onto returns and support.
4. Packaging that doesn't protect the product
A box that's too big lets the product shift around and break. Too little padding won't survive the sorting conveyors. A label stuck over an edge or over tape won't scan at the sorting centres, and the parcel gets stuck.
Packing well doesn't mean using more bubble wrap. It means choosing the right format for each product: made-to-measure boxes for fragile items, envelopes for textiles, seals that won't come open, and the label on a flat, clean surface. Try the drop test: let a packed box fall from waist height. If you hear something shift inside, the packaging isn't right.
5. Choosing the wrong courier for each shipment
Working with a single courier simplifies things, but it's rarely best for every order. One performs very well in major cities and worse in rural areas or islands; another is competitive on light parcels but expensive on bulky ones; another has a better network of drop-off points or a better service into Europe.
Always choosing the cheapest rate usually ends up costing more. The real cost of a shipment includes first-attempt delivery rates, tracking quality, how incidents are handled and how much a return costs with that courier.
Ideally, the choice should be automatic, based on destination, weight, volume and service type. At our warehouse in Murcia we work with Paack, GLS, CTT Express, NACEX, Correos, DHL, UPS, MOVIQ for pallets and Zeleris, and the system decides which courier to use for each order based on these rules. If you want to set this up yourself, read about when multi-courier shipping pays off and how to automate it.
6. Promising timeframes the operation can't sustain
"Delivered in 24 hours" sells well, but failing to deliver on it does more damage than promising 48 hours and sticking to it. The mistake is calculating the timeframe based on the perfect day: order placed early, stock available, on-time collection and a straightforward destination.
A realistic promise takes into account:
- The actual cut-off time (not the one you'd like to have).
- Picking capacity on a peak day.
- Weekends and local public holidays.
- The difference between preparation time and transit time: if you prepare in under 24 hours but transit takes 24-48 hours, the customer needs to see "2-3 days" before they pay.
- The slowest destinations: islands, Ceuta and Melilla, rural areas and shipments to the rest of Europe.
7. Lack of follow-up and communication when there's an incident
Incidents happen in any operation: no one home, delays due to bad weather, tricky access. The difference between a controlled incident and an angry customer is visibility.
If the customer doesn't know where their parcel is, they'll message you. If your team doesn't know either, the search begins across emails, courier portals and spreadsheets. And if no one is monitoring open incidents, the parcel goes back to the warehouse without anyone having called the customer.
What to do:
- Send the tracking number as soon as the label is generated.
- Check orders flagged as incidents every morning and act before they're returned: confirm the address, offer a drop-off point or a new date.
- Flag the delay yourself before the customer has to ask.
An example with indicative figures
Imagine a cosmetics store with 800 orders a month. If 3% have some kind of incident, that's 24 orders. Let's say, as an indicative figure, that each one costs between €10 and €20 once you add up reshipping, returns, unsellable stock and support time. That's roughly €240 to €480 a month, without even counting the customers who never come back.
Breaking down a month's 24 incidents might reveal something like this: 9 due to incomplete addresses, 6 due to picking errors between two shades of the same product, 4 due to broken glass bottles, 3 due to courier delays in a rural area, and 2 due to orders with no real stock. Only 3 out of 24 are actually