Multi-Carrier Shipping Platform

Our client moves parcels for people who cannot simply click “deliver to my address” and be done with it high-value items going out for repair, orders from stores that will not ship to a particular location, packages crossing the US and Canadian border.

We built the platform that quotes those shipments, books them, labels them, tracks them and invoices for them. Then we built the piece that turned it from a tool into a business.

Plenty of orders fail at the last field in the checkout. The store does not deliver to that state, that province, or across the border at all. The customer wants the item; the retailer simply will not send it that far.

A virtual warehouse address solves that by splitting one impossible journey into two possible ones. The customer is issued an address at one of the operator's warehouses two in the United States, one in Canada. The store ships there, because that is the destination it already serves. The warehouse then forwards the parcel on.

The customer can check what warehouse-to-door will cost before they order, so the decision is made with the full picture rather than half of it.

Then the parcel arrives, and this is the part that matters. Warehouse staff check it in and record what really turned up its true weight and dimensions. That is the first moment the onward leg can be priced properly rather than estimated from whatever someone typed into a form, and for awkward items it is often a different number. The platform issues a firm delivery estimate to the final address, and the parcel moves once the customer accepts it.

Nothing ships, and nothing is charged, on a guess.

Carrier APIs are not uniform, and they are not always quick. Each has its own authentication, its own address validation rules, its own idea of what a service level is called, and its own occasional slow morning.

Two decisions followed from that. Rate requests go out in parallel and are normalized into one internal shape before they reach the screen, so the customer sees a single comparable list rather than three vendors' vocabularies. And carrier calls run through Laravel's queue system, so one carrier having a bad day degrades that option rather than hanging the page for everybody.

The same applies to failure. If a carrier does not answer, the quote returns with the carriers that did, clearly showing which one is missing rather than showing an error, or worse, silently showing an incomplete comparison as if it were the whole market.

There is a related honesty problem on the warehouse side. Carriers price on dimensional weight as well as actual weight, so a light but bulky item — a surfboard being the obvious example — can cost far more than its weight suggests. Quoting that from customer-entered dimensions would produce a number that is wrong often enough to matter, which is why the firm estimate is issued at check-in, from measurements taken on the real parcel.