Insurance decisions for small parcels

The eight ways a small parcel goes wrong

A small parcel has a short list of failure modes, and they behave differently. Some arrive as a parcel that never turns up, some as a parcel that turns up wrong, and some as a parcel that turns up correctly and then becomes a problem because nobody acted inside a window. The last group is the one people leave out of their risk thinking, and it is the group with the most predictable timing.

The list below is the one the desk uses for small consignments. It is deliberately short, because a list of twenty risks is not a decision aid. Each entry is stated as an event rather than as a cause, so that two causes producing the same event are treated the same way when the response is chosen.

1. The parcel never arrives and tracking stops at an intermediate point.
2. The parcel arrives with visible damage to the contents.
3. The parcel arrives with fewer items than the build contained.
4. The parcel is held at the border for a document or a payment.
5. An assessment is raised after delivery and the amount is unexpected.
6. The handling window closes while the item is still in the warehouse.
7. Delivery is attempted at an address where nobody can receive it.
8. The parcel is routed on a line whose rules exclude its contents.

Scoring probability against impact

The scoring below is ordinal, not measured. Probability is judged against the desk own record of how often each event appears in the notes, and impact is judged by what the event costs in money and in days rather than by how alarming it sounds. An ordinal score is honest about what it knows: it ranks the risks relative to each other and does not pretend to be a rate.

Two conventions keep the matrix readable. Probability runs from one, meaning rare in the desk notes, to five, meaning routine. Impact runs from one, meaning a small amount of money and no lost days, to five, meaning the item is unrecoverable or the delay breaks a fixed date.

The product of the two is not used as a single ranking figure on purpose. Two risks can share a product and need completely different responses, and collapsing them into one number is how a matrix stops being useful.

Which risks a claim can actually reach

The reach column is the most useful column in the matrix. Four of the eight risks are things a claim can address: a parcel that never arrives, damage, a missing item, and in some cases a build that does not match its list. Those four share a property, which is that they produce a discrepancy between a record on paper and a parcel in hand.

The other four do not produce that kind of discrepancy. A border hold is a delay, and a delay has no loss attached until a deadline passes. An assessment after delivery is an invoice with its own process. A closed window is a rule, not an event that can be reversed. A refused line is a rule that was knowable before the parcel was built.

This split is what people get wrong when they buy cover for a small parcel. Cover is bought against the first group and then expected to handle the second, and the expectation survives until the first border hold, at which point the cover turns out to be irrelevant rather than insufficient.

1. Reachable by a claim: non-arrival, visible damage, a missing item, and a build that contradicts its own list.
2. Reachable only by a decision: the handling window, the delivery address, and the choice of line.
3. Reachable only by arithmetic: an assessment raised after delivery, which is contested rather than claimed.
4. Not reachable at all: a delay at the border, where the remedy is answering the request quickly.

Responses by tier rather than by risk

Sorting the eight risks into three tiers produces a decision rule that survives contact with a real parcel. The top tier is where probability and impact are both at or above three: those are the events worth spending on, either as cover or as process. In this matrix that tier contains the window, which is a process problem rather than an insurance problem.

The middle tier is where one of the two is high and the other is low. Those get a process fix, because the cost of the fix is lower than the cost of the event. A failed delivery attempt and a border hold both live here, and both are answered by a message rather than by a payment.

The bottom tier is where both are low, and the honest response is to write the risk down and do nothing else. Buying cover against a bottom-tier risk is not wrong, but it is a decision to spend on the least likely event on the list, and it happens most often when the list has never been written down.

The floor that does not depend on a decision

Three things are worth doing on every small parcel regardless of which risks you decide to spend on, because they cost nothing and they are the evidence base for everything else.

1. Keep the build list: what went into the parcel, with descriptions and values, as it was recorded at the warehouse.
2. Photograph the sealed carton and the contents in the same frame before fully unpacking, because a photograph taken after unpacking proves less.
3. Write the handling window end date down on the day the first item is received, because that date, not the delivery date, is the one that expires.

Reading the decision in one pass

The pass is short enough to run on a phone. Name the eight risks, mark which ones a claim can reach, and check that the unreachable ones have a process attached. If a risk has neither a claim nor a process, it has a habit instead, and habits are what the desk notes show going wrong.

Then look at the value of the parcel against the impact column. On a small parcel, the largest impacts are the ones where the item is unrecoverable, and those are exactly the cases where the evidence pack does the work. Cover changes who pays; records change whether anyone pays at all.

The last check is the one people skip. Every risk in the matrix has a moment before which the response is cheap, and the moments are not evenly spread. The window closes in the warehouse, the line rules are fixed at the build, and the delivery address is fixed at booking. By the time the parcel is in the air, three of the eight risks have already been decided.

The data point behind this note

Risk matrix built from the three return paths and the window-closure check in KB-2026-05 (the return model was split into three paths after a single figure was found to hide the case where keeping an item costs less, and the window check was added because an expired window silently invalidated two of the three).

◇KB-2026-05· 2026-09-30 Batch ledger · Challenge a figure

Notes in the same cluster

Related entries and gauges

• What does parcel insurance cover, and what does it exclude?
• How do you file a missing-item claim that gets answered?
• Return path bench

  What each path costs from where you are standing now.