For eCommerce Directors and fulfillment operations leaders at retail brands feeling DIM surcharges and damage returns in the P&L.
US retail returns hit $849.9 billion in 2025, roughly 15.8% of all retail sales. About one in four of those returns is a damaged or broken item. (National Retail Federation, 2025; Claimlane returns analysis, 2025)
That is one side of the margin story. The other side is that shipping itself now consumes 15 to 20% of ecommerce revenue, and carriers no longer charge by weight alone. UPS, FedEx, and USPS all bill on dimensional weight, which means an oversized box can cost 30 to 50% more than its actual weight implies (PARCEL Industry, 2025). As of August 2025, FedEx and UPS began rounding up every fractional inch before applying the DIM formula on small package services. Half-inches that used to be free now cost real money.
Stack the two together and a pattern emerges. The same fulfillment step (a packer eyeballing a box selection and tossing items in) creates both problems. Pick the wrong box and the carrier surcharges you. Pack it wrong and the item arrives damaged, then comes back. Reverse logistics already runs 20 to 30% of original product value, and more than 30% of returned units cannot be resold as new (Claimlane, 2025). Cartonization plus structured pick and pack data is not a nice-to-have anymore. It is the single workflow where ecommerce margin is most often lost.
Why this matters: labor is already 65% of a warehouse’s operating budget, and pick/pack is 50 to 70% of that labor (Stacia Americas, 2025). When the packer’s decision is unguided, the cost lands in three places: carrier invoices, returns processing, and the labor hour itself. The platforms that have solved this treat the pack station as a system, not a workstation.
What teams ship today vs. what the math actually requires
What most fulfillment platforms ship today | What a DIM-priced, returns-heavy reality requires |
|---|---|
Packer eyeballs box selection from a shelf | 3D bin-packing recommendation against a defined box catalog |
Carton dimensions re-keyed at label time | Recommended dimensions pre-populate label generation |
Fragile, orientation, hazmat captured on sticky notes | Structured shipment attributes, scoped to line item or shipment |
Carrier picked from a default, not the rate | Rate shopping against actual package dimensions |
Compliance audit means archaeology in emails | Read-only attribute record preserved with the shipment |
Three failure modes follow directly from the left column.
The DIM tax compounds invisibly. A packer who picks a 12 by 12 by 12 box when a 10 by 10 by 8 would do is paying for air. One box does not matter. Ten thousand boxes a week do. Cartonization deployments combined with right-sizing have been shown to cut transportation costs by up to 15% and packaging material by 35 to 40% (Numina Group, 2025). For a mid-sized retailer, that is seven figures.
Damaged goods come back, and they do not come back the same. Roughly a quarter of returns are damaged items, and 12% of customers explicitly cite shipping damage as the return reason (Claimlane, 2025). When the platform has no way to flag “ship this upright” or “no stacking” at the line item level, the packer guesses. The vase ships sideways. It comes back broken. You eat the cost twice: once on the outbound, once on the reverse leg.
The pack station bottleneck eats throughput. Manual walking and packer guesswork can absorb more than half of total picking time. Cartonization combined with voice picking and pick-to-light has been shown to lift picking productivity 35 to 50% and packing efficiency by up to 100%, because packers stop doing carton selection and secondary inspections (The New Warehouse, 2025).
The throughline: the pack station is being asked to make three decisions (which box, how to orient, which carrier) without the data or the math to make any of them well. Fixing that starts inside the fulfillment workflow itself.
How we think about this at Kibo
Structured shipment data, captured where the work actually happens
Operational data only matters if it lands at the moment the decision is made, not in a spreadsheet someone fills in after the fact. The fragile flag, the orientation rule, the hazmat compliance entry: these have to be required fields inside the fulfillment workflow, not best efforts in a sidebar.
Our Shipment Attributes feature lets admins define custom fields with type (text, list, yes/no, date), required vs. optional status, and editable-vs-read-only behavior. Every new shipment inherits enabled attributes automatically. If a required attribute is missing at the final step, the system blocks completion and identifies the gap. Once the shipment is fulfilled, attribute values lock as a permanent record, which is the difference between an audit trail and a post-mortem.
Scoping data to the level where it actually applies
A shipment containing a vase and a book does not want “Handle with Care” smeared across the entire order. The fragile rule belongs on the vase. The stackable rule belongs on the book. If your platform forces one value per shipment, you are losing operational precision on every multi-line order.
We built attribute scope into the model directly. Each attribute is configured as shipment-level, shipment-item-level, or both, with independent values per line item. The packer working a multi-line shipment sees “Orientation = UPSIDE” on the vase and “Stackable = Yes” on the book, with no overwrites and no manual splitting. The values stay tied to the line item forever, which is what compliance audits and carrier disputes actually need.
Cartonization that feeds the label, not the packer’s intuition
The right shape of a packing decision is straightforward: define the box catalog once, let a 3D bin-packing algorithm choose, and let those dimensions flow straight to the label and the carrier rate engine. The packer is in the loop for judgment calls and overrides, not for math the system can do faster and better.
Our Cartonization engine reads product dimensions and weights from the catalog, evaluates them against the box types you configure in Location Group settings, and returns a concrete recommendation at the Print Packing Slip step: which box, which items go in it, and how to distribute across multiple boxes when needed. A 3D visualization shows the arrangement and volume utilization before the packer commits. The recommended dimensions then pre-populate the Prepare for Shipment step, which feeds Rate Shopping against real-time FedEx, UPS, and USPS rates using the actual package, not a default. The packer can override at any time. They just no longer start from zero.
The takeaway
The $849.9 billion in returns and the 15 to 20% shipping cost ratio describe the same underlying problem from two angles. Both get worse when the platform leaves the packer to make three high-stakes decisions without help. Cartonization gives the math. Shipment attributes give the data. Rate shopping gives the carrier comparison. None of the three works as a bolt-on, and that is why most retailers running them in middleware are still losing the same margin every week.
If your pack station is still running on packer judgment and sticky notes, that is where the seven-figure leak is. Talk to Kibo about fulfillment operations.
