Insure it or self-insure it: the per-shipment math

The break-even formula: premium vs exposure times loss rate

Shipping insurance is one of the few purchases a seller makes where the fair price is genuinely computable. You are buying protection against a loss whose size you know — the replacement cost of the item, minus whatever the carrier already covers by default — and whose probability you can estimate from your own history. Multiply the two and you have the expected loss on that parcel. The premium is worth paying when it is smaller than that expected loss, and not worth paying when it is larger. Everything else on this page is just a way of getting honest inputs into that one comparison.

The formula is easier to use flipped around: divide the quoted premium by your loss rate, and you get the break-even exposure — the uncovered value at which insurance stops being a fee and starts being a fair trade. A premium divided by a one-in-five-hundred loss rate breaks even at five hundred times the premium in uncovered value. Parcels above that line deserve the toggle; parcels below it are quietly subsidising the insurer. What surprises most sellers who run the division for the first time is how high the break-even line sits for an ordinary domestic parcel that gets scanned at acceptance and delivered to a house. Insurance earns its premium on the tails, not on the typical order.

Worked example bands — compiled 2026-08-15. Illustrative arithmetic only, not a rate card: at a loss rate of one parcel in two hundred, a two-dollar premium breaks even at four hundred dollars of uncovered value; at one in a thousand, the same premium needs two thousand dollars of exposure to be a fair trade. Premiums scale with declared value and change without notice, so never reuse these bands as answers — re-run the division with the actual premium shown on your own quote, because the toggle prices each shipment individually. Re-checked with each pricing review by the named marketing owner recorded with this page.

Estimating your real loss rate from your own shipment history

The loss rate is the input sellers guess worst, because memory is a terrible ledger — one painful loss feels like a pattern, and two hundred uneventful deliveries feel like nothing at all. Count instead. A loss, for this arithmetic, is a parcel that never produced a delivery scan or that arrived damaged beyond use; a parcel the tracking says was delivered but the buyer says never came is a different dispute with different remedies, and it belongs to the delivered-but-not-received guide, not to this count. Divide honest losses by total shipments, and keep the division per lane: domestic ground to residential addresses, international economy, and anything routed through a freight forwarder are three different risks that deserve three different rates.

Small histories need one extra tool. If you have shipped three hundred parcels and lost none, your loss rate is not zero — it is merely unknown and probably small. A useful ceiling: with zero losses in a sample, the true rate is very likely below three divided by the number of shipments, so three hundred clean parcels bound the rate near one percent, not at nothing. Use the ceiling when deciding, and replace it with your measured rate as the history grows. Sellers who skip this step self-insure on the assumption that nothing has ever gone wrong, then meet their first loss with no reserve and no premium either.

Self-insuring in practice: the reserve method

Self-insuring is not the same thing as declining insurance. Declining is simply hoping; self-insuring is running the insurance company yourself, with yourself as the only customer. The mechanics take one spreadsheet column. Each time the arithmetic says skip the premium, credit the amount you would have paid into a reserve line. When a parcel is lost, pay the replacement out of the reserve and record the drawdown. The column now does something subtle and valuable: it turns your insurance decision into a falsifiable bet. A reserve that climbs steadily is proof the premiums were overpriced for your book of parcels. A reserve that keeps going negative is proof your loss-rate estimate was optimistic, and the toggle deserves another look.

The two strategies also mix, and the per-shipment toggle is what makes the mix workable. Because GoatLabels prices insurance per shipment rather than as an account setting, you can self-insure the base of cheap, frequent parcels where the reserve method wins, and buy the premium only on the outliers — the one order this month whose replacement cost would wipe out a quarter of reserve growth in a single loss. One honest boundary belongs here: GoatLabels sells the insurance option at label time and records the premium on the shipment, and the insurer underwrites the policy and decides the claim. No label platform pays claims out of its own pocket, and any page that implies otherwise is selling something it does not control.

The parcels that always clear the bar, and the ones that never do

Run the division across a real order book and the answers cluster. Clearing the bar almost every time: one-off high-value items, because a single loss is a material fraction of the month; international parcels on economy services, where handoffs multiply and the loss rate is genuinely higher; anything shipped to a freight forwarder, where the carrier's obligation ends at the forwarder's dock; and fragile goods whose damage rate dwarfs their disappearance rate. Failing the bar almost every time: inexpensive, replaceable, frequently shipped items — the volume that makes the loss rate measurable is the same volume that makes each individual loss ignorable, and a reserve funded by a hundred skipped premiums absorbs the occasional replacement without complaint.

Two adjacent facts complete the decision, and both live elsewhere by design. Carriers include some declared-value coverage by default on some services, which shrinks the uncovered exposure in your formula, and the per-carrier specifics plus the filing process live on the lost-and-damaged claims guide. And the premium is only one line in the landed cost of a label — the habit that keeps the whole number down is quoting all four carriers on every parcel, which is the cheap-labels argument in one sentence, and it costs nothing to adopt because printing has no monthly minimum here.