Shipping typically makes up 50% to 70% of total eCommerce fulfillment costs, more than labor, packaging materials and warehouse overhead combined, according to fulfillment cost benchmarks compiled by shipping calculator site CalcBee. Yet most retailers still estimate that cost using static rules: a flat rate per product category, a default box size, or an average based on last quarter’s shipments. Those rules do not account for how a specific order actually fits into a box. A merchant who underestimates that fit either eats the difference in margin or surprises the customer with a shipping fee that does not match what checkout promised.
Faire, the wholesale marketplace connecting independent retailers with brands, is treating that mismatch as a solvable problem rather than a fixed cost of doing business. Faire expanded its shipping collaboration with UPS in May, delivering an average 18% reduction in shipping rates across top carriers for orders processed through Ship with Faire, building on a 20% reduction the company introduced the year before, Faire announced. More than 5 million shipments moved through Ship with Faire in 2025, nearly 15,000 packages a day, and that volume is what gives the company enough leverage to negotiate rates typically reserved for much larger retailers, Faire said, and pass the savings down to independent brands, Faire said.
Checkout Shipping Estimates Still Rely on Static Assumptions
The deeper problem Faire and similar platforms are working against is prediction, not carrier pricing. A retailer’s checkout page has to quote a shipping cost before the order is packed, which means the system is guessing how many boxes an order will require and how efficiently the items inside will fit. Fulfillment operators typically handle that guess with fixed assumptions, such as one default box size per product or an average shipping cost drawn from historical order data rather than the actual dimensions of the current cart, according to cost guidance published by fulfillment provider Fulfillrite.
That approach breaks down as soon as a cart contains an unusual mix of products. Shopify apps such as Boxify and DimenAI now use dimensional, AI-assisted bin-packing to solve exactly this gap, calculating the smallest box or combination of boxes that will hold everything in a customer’s cart before quoting a rate, rather than defaulting to a single box size for every order, according to each company’s own product documentation.
That gap isn’t just a vendor talking point. A 2025 study from researchers at Southern Arkansas University, using a random forest machine learning model trained on product weight, volume and fragility to predict optimal packaging configurations, found that traditional box-sizing methods created significant waste from over-packaging and poor alignment between product dimensions and shipping strategy, the researchers reported in the peer-reviewed journal AI.
An Oversized Box Can Cost More to Ship Than the Item
The stakes of getting that estimate wrong are built into how carriers price packages. Dimensional weight, the pricing method UPS, FedEx and USPS use to charge for the space a package occupies rather than just its actual weight, can bill an oversized box for a light item at nearly three times its physical weight, according to shipping calculation guidance published by ShippingEasy.
A retailer that quotes shipping based on an inaccurate box-fit assumption either underprices the true cost and absorbs the loss, or overprices it and drives the customer to abandon the cart. Forty-eight percent of U.S. online shoppers who abandon a cart cite unexpected costs revealed at checkout, primarily shipping fees and taxes, as the reason, according to research from The Baymard Institute.
Getting the box-fit estimate right before checkout is a conversion question as much as a fulfillment efficiency one. As more of eCommerce fulfillment becomes automated end to end, the accuracy of the shipping estimate shown at checkout is becoming a competitive variable in its own right, alongside price and delivery speed.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
