Alert immediately on any carrier exception, with no value floor and no delay, because most exceptions require the customer to do something and the window is short. Address problems, failed delivery attempts, customs holds and refused deliveries all resolve if acted on quickly and turn into returns or losses if they sit.
Why this one is worth an interruption
A stuck shipment is silence, which you have to detect. An exception is the carrier actively reporting a problem, which you already have. The failure is not detection, it is routing: these arrive as routine emails in a shared inbox and get read as noise.
The time pressure is real and specific. A parcel held at a depot pending an address correction goes back to sender after a fixed number of days. That clock is running whether or not anybody has read the email.
The threshold
There is nothing to tune. The settings that matter are about speed and coverage.
| Condition | Set it to | Why that number |
|---|---|---|
| Trigger | any carrier exception | No value floor: a cheap parcel returned still costs |
| Delay | none | The customer usually has to act, and the clock is running |
| Priority | customs and address problems first | These have the shortest windows |
| Escalate | unresolved after 24 hours | Most return-to-sender clocks are days, not weeks |
| Coverage | every carrier you use | The one you forgot is where these accumulate |
Who should receive it
A person, not an inbox. This is the single most important routing decision in the cluster.
- Whoever runs the daily shipping check, named, with a named cover.
- They need authority to contact the customer and pay for a redelivery or correction.
- Not a shared inbox. Carrier exception emails landing somewhere unowned is the default state and the reason these get missed.
What to do when it fires
- Read what the carrier is actually asking for. Most exceptions name the required action.
- Contact the customer the same day if the fix needs them, with the specific thing required.
- Fix it directly with the carrier where you can, rather than relaying through the customer.
- If it is returning to sender, tell the customer before the parcel arrives back, and decide then whether to reship or refund.
When not to set this alert
There is no good case for switching this off, but there is a case for consolidating it. If your carrier already delivers exceptions into a system a named person watches hourly, a second alert adds nothing.
If you ship only by a service with no exception reporting, this has nothing to fire on and the stuck-shipment check covers you instead.
Common questions
What is a delivery exception?
A carrier-reported problem preventing delivery: a wrong or incomplete address, a failed attempt, a customs hold, a refusal or damage. Most require somebody to act before a return-to-sender clock expires.
How fast do I need to respond?
Same day. Return-to-sender windows are typically days rather than weeks, and the clock starts when the exception is raised rather than when you read it.
Why do these get missed so often?
Because they arrive as routine emails into a shared inbox nobody owns, and read as noise among delivery confirmations. The fix is routing to a named person, not better detection.