Amazon reported third quarter results on Thursday, October 29. Worldwide shipping costs came in at $28.3 billion, up 11% from a year ago. Fulfillment expense was $30.0 billion, against $27.7 billion in the same quarter last year. (Amazon investor relations, October 29, 2026)
Paid units grew 12%.
That one-point gap is the only number in the release that matters to a brand shipping 1,000 to 25,000 orders a month. Amazon moved more boxes than its shipping bill grew. If you ran the same comparison on your own invoices this quarter, most brands would find the gap pointing the other way.
Strip out the fuel and the gap widens
On the earnings call, Amazon's CFO said the quarter carried higher transportation costs from fuel inflation and higher line haul rates. Excluding those two pressures, shipping costs grew 8%, which is roughly 400 basis points slower than unit growth. (Amazon Q3 2026 earnings call transcript, October 30, 2026)
Read that backwards. Fuel and line haul added about three points to Amazon's shipping cost growth. Amazon buys more freight capacity than any retailer in the country and it still ate three points of fuel and linehaul inflation it could not negotiate away.
Your invoices are seeing the same three points. You are just seeing them without the offset.
The offset came from two kinds of decision
Amazon credited the improvement to two things, and neither one is exotic.
The first is placement. Where inventory sits before an order exists, and how far the package travels after it does. Amazon said it lowered its overall cost to serve through better inventory position and shorter shipping distances.
The second is packing. Consolidating a multi-item order into fewer boxes.
Notice what is missing from that list. Amazon now runs over a million robots in its network and doubled its fleet of robotic arms during 2026. Those investments showed up in productivity and in record-low safety numbers. They are not what the company pointed to when explaining the shipping line. (Amazon Q3 2026 earnings call transcript, October 30, 2026)
Distance moved the shipping cost. Automation moved other lines.
That distinction is the practical takeaway, because distance is the one lever on Amazon's list a mid-sized brand can actually buy. You cannot buy a million robots. You can buy a zip code.
Run the number on your own invoices
This is a 20-minute exercise and most brands have never done it.
Pull your parcel spend for the last 90 days. Divide by units shipped. That is your real cost per order, carrier-blended, surcharges included. Now do the same for the same 90 days last year.
If cost per order grew faster than your unit count, you lost the race Amazon just won. If it grew slower, find out why before you assume it was skill, because a product mix shift toward lighter items will do it too.
Then pull the zone distribution on those same shipments. Count the share landing in zone 5 and beyond. That number is the one you can change without renegotiating anything. A rate discount applies to a rate. A shorter zone applies to a smaller rate.
Four things that move it, in rough order of how fast they work:
- Audit your packaging against dimensional weight before peak, not after. Oversize and additional handling charges do not care about your discount.
- Check whether your fulfillment location actually covers your customer map, or whether it covered it three years ago.
- Consolidate multi-item orders at the pack bench instead of letting a WMS split them by pick path.
- Push back on line haul. It is the fastest-rising part of inbound freight and the least audited.
What this means for your brand
Amazon's answer to distance is more buildings. That is the right answer when you have Amazon's volume and it is an impossible answer at 5,000 orders a month, where splitting inventory across coasts usually costs more in safety stock and transfers than it saves in postage. We covered the version of this argument that showed up in Prologis research earlier this week. (Next-day delivery and why speed beats price)
The single-node version of Amazon's lever is warehouse placement. We ship from Canal Winchester, Ohio, in the Columbus metro. That puts 60% of the US and Canada population within 600 miles and reaches 90% of customers in two days by ground, from one pool of inventory. Short zones are how a brand with one warehouse borrows the lever Amazon pulled with hundreds. (Ecommerce fulfillment)
The packing lever is less glamorous and often bigger. Orders here are processed within 24 hours, and consolidation happens at the bench, so a four-item order does not leave as four parcels because that was easier to pick.
One honest caveat. Amazon's 11% was achieved while it was also hiring more than 300,000 seasonal workers and expanding its fastest delivery tier to over 450 cities. It spent heavily to grow cheaply. The lesson is not that cost control is free. It is that the cheapest unit of cost control available to you is the mile you never drive.
If you want to know what your current zone mix is actually costing you, send us 90 days of shipment data and we will tell you. (Talk to us)
FAQ
Does Amazon's cost structure tell me anything useful at my size?
The absolute numbers, no. The direction, yes. When the largest shipper in the country says distance and consolidation moved its shipping line more than automation did, that is a signal about which problems are worth your attention this quarter.
Should I add a second warehouse to shorten zones?
Usually not below about 25,000 orders a month, and not for postage alone. A second node adds safety stock, transfer freight, and a second count of everything. Run the math on a single central node first. Our FAQ covers how we think about it.