Delivery exception handling
Also called proactive delivery alerts, shipment exception management, WISMO automation, late delivery alerts
Spotting late or failed deliveries before customers complain: what the evidence says on delivery failures, apologies, the law and the tools.
Late deliveries cost repeat business; evidence that alerts and apologies win it back is thin. After a fulfilment delay, loyal customers of one online retailer ordered less often and placed smaller orders. Each business day cut from delivery time lifted one apparel retailer's online sales by about 1.45%. In a field experiment with about 1.5 million ride-hailing riders after late trips, a US$5 coupon lifted next-week spending by about 1.5%, with or without an apology; an apology alone showed no significant effect in the main estimates.
At a glance
| Status | In pilots |
|---|---|
| What it does | Watches courier tracking for shipments that are late, stuck, returned or failed, alerts your team and messages the customer before they chase you. |
| Typical cost | Free for 50 shipments a month, then from US$11 per month for 100 shipments with one tracking app (vendor list price, checked Oct 2026). |
| Data it needs | Orders with promised delivery dates, tracking numbers and courier names, live tracking events, and customer contact details. |
| Setup effort | Low: tracking apps connect to common shop platforms; the work is writing rules and messages for each exception type. |
| Evidence | Thin. Peer-reviewed studies on the cost of late delivery and one large field experiment on apologies; no independent study of exception-alert tools. |
What is delivery exception handling?
A delivery exception is any shipment that does not go to plan: stuck at a depot, running past the promised date, a failed delivery attempt, a wrong address, damage or a return to sender. Without monitoring, the first sign is often the customer asking where the order is.
Exception handling turns that round. Software watches tracking events from every courier you use, compares them with the date you promised and flags orders that are going wrong. It then alerts a person, messages the customer, or both, before the complaint arrives.
Where vendors use the AI label, it can cover predicted delivery dates (one tracking app offers AI-estimated delivery dates on its higher plan) and drafted customer messages. The underlying job is simple: know about the problem before the customer does, and do something useful about it.
How does it work?
The tool pulls tracking numbers from your order system and checks each courier for status updates. Rules define what counts as an exception: no scan for a set number of days, an estimated date later than the promised one, a failed attempt, or a return-to-sender event. Each exception type gets a route. A failed attempt might trigger a message asking the customer to confirm the address; a long delay might go to a team member to choose between a reship, a refund or a credit.
The value sits in the action, not the alert. A dashboard of red flags that nobody checks changes nothing. Useful set-ups decide in advance what the business will offer for each type of failure and who signs it off, so the customer hears something specific: what happened, the new date and what you will do about it.
What it needs to work
Exception handling depends on joining each order to its live tracking and to the promise you made. Without a promised date per order, the tool cannot tell late from normal.
- Order records with order date, promised delivery date or service level, and delivery address, in your shop or order system
- A tracking number and courier for every shipment, written back to the order automatically rather than pasted in by hand
- Live tracking events from each courier portal or a tracking aggregator, with each courier's status codes mapped to your exception types
- Customer contact details and consent records for the channels you will use, such as email, SMS or WhatsApp
- A written playbook: for each exception type, the message, the remedy (reship, refund or credit) and who approves it
Most use cases in this guide succeed or fail here, on the data underneath them, before any AI is involved.
What does the evidence show?
The evidence that delivery failures hurt is solid. At a mid-sized online retailer, previously loyal customers ordered less often and placed smaller orders after a fulfilment delay. At a US apparel retailer, each business day cut from a seven-business-day delivery time raised online sales by about 1.45%. When Alibaba dropped a high-quality courier for 42 hours, sales fell 14.56%, with bigger drops for expensive products.
The evidence that proactive handling repairs the damage is thinner. The best study is a field experiment with about 1.5 million Uber riders after late trips. Riders hit by the worst delays spent 5% to 10% less afterwards. An apology with a US$5 coupon raised spending over the next week by about 1.5%, as did the coupon alone; an apology alone showed no significant effect in the main estimates, and a third apology backfired. That is ridesharing, not parcels.
We found no independent study of exception-alert software for small online sellers. Claims that it cuts where-is-my-order queries come from vendors.
Proactive exception handling vs waiting for complaints
| Compared on | Proactive exception handling | Waiting for complaints |
|---|---|---|
| When you find out | When tracking shows a problem | When the customer contacts you |
| Customer experience | Hears first from you, with a plan | Has to chase; the delay has already hurt the relationship |
| Workload | A steady, rule-driven queue | Spikes of where-is-my-order messages |
| Evidence of payoff | Thin: one large field experiment outside parcels | The baseline |
| Best for | Steady order volumes across several couriers | Very low volumes with one reliable courier |
Status and evidence
Peer-reviewed studies show late and failed deliveries cost sales and repeat orders. Evidence that proactive alerts and recovery fix this is limited to one large ridesharing experiment, and claims about fewer where-is-my-order queries come from vendors. Tools are cheap and off the shelf, so we rate it pilots with thin evidence.
Where does it go wrong?
- Patchy tracking data. Couriers differ in what they report and how quickly. If a courier scans late or not at all, the tool raises false alarms or misses real failures.
- Alerts with no owner. Exceptions pile up in a dashboard that nobody works through.
- Over-messaging. Repeated apologies can make things worse: in the Uber experiment, a third apology reduced later spending.
- Empty words. In the same experiment, an apology without anything of value had no significant effect on later spending in the main estimates.
- Messages that skip the legal basics. In the US, a seller who cannot ship on time must offer the choice of accepting a delay or cancelling for a prompt refund. A friendly update that leaves this out does not meet the rule.
- Assuming the courier carries the risk. Under UK law, goods stay at the seller's risk until the customer physically has them, unless the customer arranged their own courier that the seller did not offer. In South Africa, goods stay at the supplier's risk until the consumer accepts them.
Numbers it should move
Ask the demo business
Northgate Supply is a fictional wholesale distributor of catering and packaging supplies: 28 staff, R46.6m in sales over the last year. Its data is generated for this demo, to 30 September 2026.
94.7% of September orders arrived on the promised day with nothing missing. The worst month was June 2026 at 84.4%: 15.6% of orders were late or short, against an average of 10.1% in the other months.
How this was worked out
Orders delivered on or before the promised date and with every line complete, divided by all orders, per month. The chart shows the rest: orders that were late, short or both.
Where the data usually lives
Where the records this use case needs are usually kept in a small business. It is not a list of recommended tools.
Frequently asked questions
Should I tell customers about a late delivery before they ask?
Yes, but make it useful. Late deliveries reduce repeat orders. In a large field experiment after late rides, an apology with a US$5 coupon lifted spending over the next week by about 1.5%, an apology alone showed no significant effect, and a third apology reduced spending. Give a new date and a concrete remedy, and avoid repeated generic apologies.
How much does delivery tracking and exception software cost?
One tracking app on the Shopify App Store lists a free plan for 50 shipments a month, then US$11 per month for 100 shipments and US$70 per month for 500, with per-shipment charges above that (checked Oct 2026). Prices vary by vendor. Staff time to work through the exception queue is a cost too.
Do I need a developer to set up delivery exception alerts?
Usually not if you sell through a common shop platform and use couriers the tool already supports: setup means connecting the store and writing rules and messages. Smaller or regional couriers may not feed tracking data cleanly, and linking them can need technical help. Test the tool on a month of past orders to check it spots the exceptions you know happened.
What does the law say if a delivery is late or lost?
In the UK, unless agreed otherwise, goods must be delivered within 30 days. Customers can end the contract if a delivery where timing was essential is late, and goods stay at your risk until the customer has them. In South Africa, goods stay at the supplier's risk until accepted, and a consumer may cancel without penalty if delivery comes at a different time than agreed. This is not legal advice.
Does POPIA or UK GDPR apply when I share customer details with a tracking tool?
Yes. Names, addresses and phone numbers are personal information. POPIA requires you to keep it secure and to have a written contract with any operator that processes it for you, such as a tracking or messaging provider. UK GDPR requires a contract with processors that covers your instructions, security and use of sub-processors. Check where the provider stores data. This is not legal advice.
More in stock, supply & fulfilment
Sources
- 1Failure to deliver? Linking online order fulfillment glitches with future purchase behaviorJournal of Operations Management, 2011Study
- 2The Value of Rapid Delivery in Omnichannel RetailingJournal of Marketing Research, 2019Study
- 3Value of High-Quality Logistics: Evidence from a Clash Between SF Express and AlibabaManagement Science (via IDEAS/RePEc), 2020Study
- 4Toward an Understanding of the Economics of Apologies: Evidence from a Large-Scale Natural Field ExperimentNational Bureau of Economic Research, 2019Study
- 5Business Guide to the FTC's Mail, Internet, or Telephone Order Merchandise RuleFederal Trade Commission, 2025Government
- 6Consumer Rights Act 2015, section 28: Delivery of goodslegislation.gov.uk, 2015Government
- 7Consumer Rights Act 2015, section 29: Passing of risklegislation.gov.uk, 2015Government
- 8Consumer Protection Act 68 of 2008Government of South Africa, 2008Government
- 9AfterShip Order TrackingShopify App Store (AfterShip), 2026Vendor
- 10Protection of Personal Information Act 4 of 2013Government of South Africa, 2013Government
- 11What needs to be included in the contract?Information Commissioner's Office (ICO), 2026Government
This entry is educational. It is not legal, financial or professional advice, and it does not recommend any vendor. Figures attributed to vendors are their own claims.
- 1Failure to deliver? Linking online order fulfillment glitches with future purchase behaviorJournal of Operations Management, 2011Study
- 2The Value of Rapid Delivery in Omnichannel RetailingJournal of Marketing Research, 2019Study
- 3Value of High-Quality Logistics: Evidence from a Clash Between SF Express and AlibabaManagement Science (via IDEAS/RePEc), 2020Study
- 4Toward an Understanding of the Economics of Apologies: Evidence from a Large-Scale Natural Field ExperimentNational Bureau of Economic Research, 2019Study
- 5Business Guide to the FTC's Mail, Internet, or Telephone Order Merchandise RuleFederal Trade Commission, 2025Government
- 6Consumer Rights Act 2015, section 28: Delivery of goodslegislation.gov.uk, 2015Government
- 7Consumer Rights Act 2015, section 29: Passing of risklegislation.gov.uk, 2015Government
- 8Consumer Protection Act 68 of 2008Government of South Africa, 2008Government
- 9AfterShip Order TrackingShopify App Store (AfterShip), 2026Vendor
- 10Protection of Personal Information Act 4 of 2013Government of South Africa, 2013Government
- 11What needs to be included in the contract?Information Commissioner's Office (ICO), 2026Government