In Q2 2024, I put a re-quote in front of our CFO that showed a 21% unit-price reduction on our white corrugated mailer program. Twenty-two thousand units a year, three SKUs. The incumbent was running $0.78 a unit, landed. The new vendor quoted $0.61.

He signed off in nine minutes. Nine minutes, for a decision that took me three months to unwind.

When the first-year invoices reconciled in our cost tracking system, the actual landed cost was $0.94 a unit. Not $0.61. Not $0.78. Ninety-four cents — a 20% overrun against the number I'd promised the CFO, and a 21% "saving" that quietly dissolved into freight, tooling charges, and 900 scrapped boxes.

I've managed packaging spend for six years, roughly $180,000 annually across corrugated, labels, and stretch film — about 340 purchase orders logged. The Q2 2024 re-quote is the one that broke my faith in unit-price comparison. Not because the numbers were wrong. Because I was comparing two numbers that measured different things.

The problem you think you have

If you're sourcing white corrugated boxes right now, the problem in your head probably looks like this: I'm paying too much, or I don't know if I'm paying too much, and I need two or three more quotes to find out.

So you send out a request. "Custom made corrugated boxes, white, 12 x 9 x 4, approx. 20,000/year." You get back three numbers that span 40%. You pick the low one, or you pick the middle one and feel prudent.

That's the surface problem. It's real, it's painful, and it's almost never where the money is.

What's actually driving the number

Here's what took me six years and a very awkward budget review to understand: in corrugated, you are not really buying boxes. You're buying board grade, press time, pallet cube, and lead time — and the box is just the shape those four things happen to take.

1. "White corrugated" is a description, not a specification

This is the one that got me. Two converters can both confirm "white corrugated, same as current" and quote you 30% apart, and both can be telling the truth.

What people call white corrugated is usually one of three things: mottled white (white top liner over a recycled medium, slight speckling, cheapest), bleached white top liner (cleaner face, more expensive), or double-white / fully bleached (white on both liners, the most expensive, and usually unnecessary if you're printing a full-coverage design anyway).

Then there's the flute profile — B-flute at roughly 1/8", C-flute at roughly 5/32", E-flute around 1/16" — and the strength rating, which is where the real divergence hides. A quote written as "32 ECT" (edge crush test, a stacking-strength measure) and a quote written as "200 lb test" (Mullen burst, a puncture measure) are not the same board, even when they're nominally interchangeable for light-duty shipping. If you don't specify one, you will get both, and the price difference is board, not margin.

My 2024 re-quote wasn't comparing like to like. The incumbent was quoting mottled white top, B-flute, 32 ECT. The new vendor quoted bleached white top, E-flute, and I never noticed because neither quote sheet had a field for it.

2. One-time charges are where the comparison breaks

The new vendor's quote had a line item I skimmed past: $650 for plates and a cutting die. Our incumbent had amortized that years ago and never mentioned it again, because it was already paid. On a 22,000-unit program that's 3 cents a unit, which sounds trivial right up to the moment you change vendors again in year two and pay it twice.

Print method matters here too. One-color flexo on a stock-size box needs a plate; four-color process or a photo-quality panel may push you toward digital or litho-laminated, which changes both the tooling and the minimum run. If your design is still in revision, you're paying for tooling on a moving target.

3. MOQ is a cash decision wearing a price tag

The new vendor's price break was real — at 5,000 units per SKU. Three SKUs meant 15,000 units landing at once, roughly five months of inventory for us, sitting on 6 pallets in a leased warehouse at about $14 a pallet position per month in our market (as of early 2024; verify your own rate).

That's not a packaging cost. That's working capital and floor space, and it never shows up on the quote sheet. What I should have asked for was a tiered quote — 1,000 / 3,000 / 5,000 — and then run the math on all three instead of reflexively taking the deepest break.

4. You're not buying square footage. You're buying air.

This is the counterintuitive one, and it's the reason I now treat box dimensions as a logistics decision rather than a packaging one.

Freight is charged on cube and weight, not on unit count. In Q3 2023 we widened a mailer by half an inch to accommodate a new product variant — 12 x 9 x 4 became 12 x 9.5 x 4. Tiny change. It dropped our pallet pattern from 1,200 units per pallet to 960. Same truckload, 20% fewer units. Our per-unit freight cost went up 20% overnight, and nobody flagged it because the box price didn't move at all.

Half an inch. That's the whole story. Before you finalize a custom size, get the pallet pattern and the case count from your converter, not just the dieline. Ask how many fit on a 48 x 40 pallet, how many pallets fit in a 53' trailer, and what the cube utilization looks like. A box that costs 4 cents more but ships 15% denser is usually the cheaper box.

5. Lead time gets charged twice

A longer lead time isn't free — you pay for it in inventory you have to carry and in the flexibility you give up. A shorter lead time isn't free either, because expedite fees and partial shipments are how you buy it.

Our new vendor quoted 6-8 weeks. Our incumbent was at 3-4. We absorbed the difference by ordering earlier, which meant more cash tied up, which meant the finance team's version of "savings" and mine were never going to match.

What this actually costs (the part nobody puts in the spreadsheet)

When I finally sat down and reconstructed the year, the invoice was the smallest line.

  • Scrap and rework: 2.1% of units in 2023 came back with print misregistration or weak glue joints. On 22,000 units that's roughly 460 boxes, plus the labor to inspect, repack, and re-ship. Call it $1,400 all-in including the labor hours.
  • Administrative drag: every PO cycle on a custom program — spec confirmation, artwork approval, sample sign-off, receiving inspection — runs about 3.5 hours of somebody's time. At 14 cycles a year, that's 49 hours I never billed to anyone.
  • Customer chargebacks: this is the one that scares me. Two of our retail accounts assess compliance penalties for late or mislabeled shipments. One missed pallet in 2023 cost us $850. That single event was larger than the entire first-year "saving" on the re-quote.
  • Carrying cost: the extra inventory from the MOQ tier tied up roughly $9,000 in working capital for five months.

Add it up and the $3,740 of theoretical unit-price savings became a net loss of somewhere around $1,100 — before counting my time.

The lowest quoted price is a hypothesis. The landed cost is a fact. I now refuse to compare vendors on anything but landed cost, and I make them itemize to get there.

What I do differently now

I'm not going to pretend this turned into a sophisticated system. It turned into a one-page document and a rule.

The one-page spec sheet

Every RFQ now goes out with the identical spec sheet attached, and I require vendors to quote against it or formally note a deviation. It has eleven fields: inside dimensions (inside, not outside — this trips people up constantly), board grade with ECT or Mullen stated explicitly, flute profile, liner type (mottled white vs. bleached white top vs. double white), print method and colors, color target, closure style, case count, pallet pattern, target annual volume with quarterly breakdown, and required delivery window.

Eleven fields. It takes twenty minutes to fill out and it has eliminated roughly 90% of the apples-to-oranges quoting I used to wade through.

The landed-cost RFQ

I now ask, in writing, for a quote stated as delivered cost to our dock, all fees itemized separately. That means: base unit price at each volume tier, one-time tooling charges (plates, dies, proofs, samples), freight per shipment with the assumed origin and mode, any rush or split-shipment fees, and the stated lead time with the date the quote was issued. If a vendor won't itemize, that's information too.

The TCO line I actually compare is five numbers wide: base price + setup and tooling amortized over the expected run + inbound freight + expected scrap/rework + inventory carrying cost. It takes about fifteen minutes in a spreadsheet. It has never once matched the ranking I would have gotten from unit price alone.

The cube test

Before any custom dimension gets approved, I ask for the pallet pattern in writing. If a half-inch change costs me 240 units per pallet, I want to see that number before the die is cut, not after the first truckload arrives.

The golden sample

We sign and date a physical sample, keep one in the office, and photograph it with calipers on the inside dimensions. First-article approval against a signed sample is the cheapest quality control I've ever bought. It cost me nothing and it would have caught the flute substitution in about four seconds.

Regional math: California and Seattle

Geography is a real cost line, and it's one people skip. If you're sourcing custom corrugated boxes in California, the converter density in the southern part of the state means more bidders and more freight options, but you're also competing for press time with very large accounts, so lead times can stretch in Q4. In the Seattle, WA market, the bid list is thinner and a meaningful share of board comes in from outside the region, which shows up in both lead time and freight — and it means a converter with a plant or a distribution point closer to you can beat a lower-priced bid from farther out once you load the freight in.

My rule now: don't compare a quote from a plant 40 miles away to one 900 miles away without the freight line filled in. Ever. The distant quote wins on paper about half the time and loses on landed cost most of those times.

I'd also flag one thing on claims: if a supplier advertises recycled content, compostability, or food-contact suitability on a white corrugated box, ask for the documentation before it goes on your packaging artwork. Labeling rules around recyclability claims have tightened in several states, and "eco-friendly" is not a certification. Verify anything that ends up printed on the box.

Where this leaves me

It took me six years and about 340 purchase orders to understand that the number at the top of a quote is the least reliable number on the page.

Looking back, I should have sent the incumbent's full board spec to the new vendor word for word instead of writing "same as current." At the time, I thought "same as current" was a specification. It isn't — it's a wish. But given what I knew then, and given that nobody had ever shown me a dieline or a pallet pattern, my choice was at least a reasonable one made with bad inputs.

The upside on that re-quote was $3,740. The risk was missing a Q3 launch window and eating a chargeback. I kept asking myself whether $3,740 was worth potentially losing a retail account over. I told myself the answer was obviously yes, because the number was on a spreadsheet and the risk was in my head.

Now the risk is on the spreadsheet too. Same column as everything else. That's the entire change.