UPS Started Charging for Air. Then It Started Firing Its Biggest Customer.

Picture two packages on the same brown conveyor belt. One is a dense, two-pound box of bolts. The other is a giant carton holding a single throw pillow and a lot of air. They weigh nothing alike, but they take up the same room on the same truck driving the same route. For most of UPS's history, the bolts paid more, because UPS priced by weight. Then it stopped. The shift from 'how heavy is it' to 'how much of my truck does it eat' is the quiet hinge on which UPS's entire pricing strategy turned - and it explains why, in FY2025, the company could shrink its volume and grow its revenue per package at the same time.¹

The story everyone tells is that UPS just raises rates every December and passes through fuel costs - a reactive, inflation-chasing toll. That isn't what happened. Over roughly a decade UPS rebuilt the unit it charges for, layer by layer, from a blunt weight scale into a system that meters cubic space, delivery difficulty, and even the time of year. The base rate became the least interesting number on the invoice.

How a package stopped being priced by its weight

The pivot point is dimensional weight - the practice of charging by the volume a package occupies rather than the scale it tips. Popular write-ups date this to 2015, and they get it twice wrong. UPS's own filing puts the expansion to all Ground services on December 29, 2014, days before the year turned.⁵ And it wasn't a debut: dimensional weight had long applied to UPS Air and International, and to oversized ground packages over three cubic feet since 2007.⁶ What changed in late 2014 was scope - every ground package, regardless of size, now got measured by the room it stole.⁵ That is the difference between tweaking a price and redefining the product you sell. UPS was no longer selling pounds of transport. It was selling cubic feet of a finite, moving warehouse.

On top of the space charge, UPS stacked the rest of the architecture. Fuel became its own surcharge, repriced weekly off Department of Energy indices rather than smeared into the base rate.² Residential delivery - more stops, fewer packages per stop, the costliest geometry in the network - got its own line. And then came time. UPS introduced peak/demand surcharges in May 2020 as an explicitly temporary lever to absorb the COVID eCommerce surge, starting with residential ground and air.⁷ Treat that origin story as a footnote and you miss the whole move: an emergency measure quietly became a permanent pricing pillar, extended in duration and expanded in scope.⁷ Each layer monetized something the old weight scale couldn't see - air, complexity, and the calendar.

The year UPS chose yield over volume - on purpose

All that machinery built toward a single decision that arrived under CEO Carol Tomé: stop chasing packages. The phrase the press loves - 'better, not bigger' - is attributed to Tomé's remarks to investor analysts, not to any official UPS strategy document; the filings themselves talk about 'revenue quality,' the 'Network of the Future,' and a 'Transformation Strategy.'⁸ But the math is unambiguous. In FY2025, U.S. Domestic average revenue per piece rose 7.1%, to $13.21 from $12.34, while average daily volume fell 8.6% to about 17.5 million pieces.¹ Read those two numbers together: UPS deliberately handled fewer boxes and made more on each one. A network built to fill trucks was now being run to fill them well.

The clearest proof is who UPS fired. Amazon was its single largest customer at 10.6% of consolidated FY2025 revenue - and UPS announced it would cut that volume by more than half by June 2026 against 2024 levels, calling the Amazon fulfillment-center outbound flow not profitable or a healthy fit for its network.⁴ No company walks away from a tenth of its revenue by accident. The same filing shows where UPS is steering instead: SMB penetration past 30% of domestic volume and healthcare logistics revenue above $11 billion⁴ - exactly the segments its FY2024 10-K names as the ones that value an end-to-end network and will pay for it.³ The surcharge architecture and the Amazon exit are the same strategy at two scales: charge for what the network actually costs, and refuse the volume that doesn't clear the bar.

Isn't this just a giant raising prices in a captive market?

The fair objection is that 'sophisticated repricing' is a flattering name for market power: a duopoly with FedEx, a December rate hike like clockwork, and a fuel pass-through dressed up as strategy. There's truth in it - the December 2025 round was another roughly 5.9% net increase across Air and Ground.² But the volume number is what breaks the simple story. A pure price-gouger keeps every package and charges more; UPS is actively shedding its biggest account because the price it could get didn't cover what that traffic cost the network.⁴ That's not raising rates in a captive market - it's the opposite, walking away from revenue it could have kept. And note the honest limit: the 5.9% headline increase is not the same thing as the 7.1% revenue-per-piece gain. The 7.1% blends the rate increase with favorable customer and product mix and fuel effects¹ - which is exactly the point. The yield came from changing the mix, not just the menu. You can argue UPS got lucky that healthcare and SMB pay better than fulfillment-center bulk. You cannot argue it didn't choose them.

UPS spent a decade learning to charge for air, for difficulty, and for the calendar - and then used that pricing apparatus to do the thing that looks insane from a distance: hand back its largest customer. The throw pillow and the box of bolts are still on the same belt, but UPS now knows precisely what each one is worth to it, and it has decided it would rather move fewer of the right ones. The blunt instrument became a scalpel, and the most expensive lesson it taught the company was that the most valuable package is sometimes the one you decline to carry.