Add a thin layer of felt to a shoe sole and the import duty falls from 37.5% to about 12.5%. Same shoe, same factory, same production line — one small change. That is the whole value of tariff engineering: it does not depend on arguing with Customs. It depends on the product itself.
Let me run the numbers first, because that is where this story starts. Suppose you import 100,000 pairs of shoes a year at a landed cost of $20 a pair. At 37.5%, the annual duty bill is $750,000. At about 12.5%, it is $250,000. The difference is $500,000 a year. What does the felt cost? Felt plus labor, a few cents a pair — tens of thousands of dollars a year at most. Any finance person who sees that math sits up straight.
This is not a new trick. The U.S. Supreme Court settled the principle in 1882, in Merritt v. Welsh — the sugar case. An importer has the right to design a product so that it falls under a lower duty rate. Customs classifies goods based on their actual condition at the time of import. What the product is when it crosses the border is what determines the rate. You designed the product; the duty follows the design. That is legal. That is tariff engineering, not tax evasion. There is a line between the two, and I will show you exactly where it runs.
Tariff Engineering in One Shoe: The Felt on the Sole Was Worth $500,000
Converse did this, and Customs ruled on it formally. A shoe with a rubber sole is classified as athletic footwear: 37.5% duty. Converse added a layer of felt to the sole. The sole's principal material became textile, the shoe was reclassified as a textile-soled slipper, and the duty dropped to about 12.5%. Customs issued a formal ruling — HQ 965751 — confirming the classification.
Pay attention to the nature of the change. The felt is genuinely attached to the sole. The shoe goes out the door exactly like that. Nobody peels the felt off after import. The change is real, permanent, and part of the product. That is the first condition for tariff engineering to work: the modification has to live on the product. It cannot be a prop stuck on for inspection.
Columbia played the same game with an even smaller change. A women's shirt carried a 26.9% duty rate. Columbia added a small zippered pocket at the waist — one pocket — and the shirt moved into a different classification at about 16%. At a landed cost of $15 a shirt, that saves roughly $1.60 a shirt. Ship a few hundred thousand shirts a year and you are looking at hundreds of thousands of dollars. The pocket holds things. It is a real, functional feature, not a decoration drawn on for show.
Put the two cases side by side and the pattern is clear. The tariff schedule is an extremely fine net: thousands of headings, rates ranging from zero to tens of percent. Shift the product design slightly toward a lower-duty heading and the classification follows. But the order matters: read the tariff schedule first, then design the product. Do it the other way around — build the product, then hunt for a lower rate — and you are gambling, not engineering.
How I Run the Numbers on Tariff Engineering: Three Tables Side by Side
Most companies do not ignore the practice. They just never run the numbers to a point where someone can approve the project. When I cost this out for a client, I always lay out three tables.
The first is the duty-gap table. List your main product's current duty rate, the rate it could reach after redesign, and the gap between them. Multiply the gap by annual import value and you get the theoretical saving. If that number is under 1% of annual import value, park it — the redesign may cost more than it saves.
The second is the redesign-cost table. How much to change the mold, change the material, change the packaging — and whether the change hurts the product or makes customers think less of it. Converse's felt barely changed how the shoe wears; Columbia's pocket arguably added a selling point. Those are good changes. If cutting the duty means mangling the product until customers walk away, the duty saving gets eaten alive by lost sales.
The third is the risk-cost table, and this is the one people most often skip. The biggest cost of the practice is not the mold. It is getting the classification wrong. If Customs rejects your classification, back duties plus interest are just the starting price. Worse is the time cost: a classification dispute can drag on for a year or more, goods stuck at the port, cash flow choked. Plenty of companies never put that line in the spreadsheet at all.
Lay all three tables side by side and the conclusion writes itself. Big saving, small redesign cost, controllable risk — go. Missing any one of the three — hold off.
Ford's $365 Million Is the Price of the Red Line
Tariff engineering is legal, but it has a red line. Ford paid $365 million to show the whole industry where that line is.
Ford imported Transit Connect vans with seats installed, claiming they were passenger vehicles. Passenger vehicles pay 2.5% duty; cargo vans pay 25% — a 22.5-point gap per vehicle. The problem was the word "claiming": the seats were removed after import. The vehicle was a cargo van from start to finish. Customs ruled it a disguise, not a design. The final bill was $365 million.
That is the red line: the change has to be real and it has to stay on the product after import. Something stuck on temporarily and stripped off after clearance is, in legal terms, an artifice — a disguise. Under 19 USC 1592, penalties can reach 20% of the goods' value for negligence, 40% for gross negligence, and 100% for fraud — plus back duties and interest. One hundred percent means the entire shipment is forfeited. You gave the goods away. Ford's case did not even reach the ceiling.
So when I judge whether the strategy is viable, I ask three questions. First, is the change genuinely part of the product, still there after import? Second, does the change serve a real function, or is it glued on purely for classification? Third, if a Customs officer stood in front of you, could you walk through the reasons for the change, one by one, without flinching? If you cannot answer any one of the three, leave the plan alone.
Before You Touch Anything, Get a Binding Ruling
The safest way to run the practice is to apply to U.S. Customs and Border Protection for a binding ruling before you change anything. Send the product sample, the design drawings, and your classification reasoning, written out clearly, and wait for the formal ruling. Until it arrives, you know how the product should change and what the duty will be. No guessing.
A ruling costs very little. The trouble it saves is real money. I have seen too many companies change the product first, ship first, and argue with Customs later — and end up paying back duties plus fines that cost more than simply paying the original rate would have. Get the order wrong and a money-saving move becomes a money pit.
Before you file, prepare the sample, before-and-after photos, drawings with the change circled, a heading-by-heading classification argument, and a cost calculation showing the change makes commercial sense. The stronger the file, the faster the ruling comes.
One more thing that has to be said up front: tariff rates and classification rules change. Check the latest official guidance before you act. The rates in this article come from published rulings and cases — accurate when written — but verify them again on the day you move. Tariff engineering is a game played by the rules, and when the rules change, the play changes with them.
Now the final tally. Suppose your flagship product has $5 million in annual import value, a current rate of 25%, and a redesign that can bring it to about 12.5%. That saves $625,000 a year. A one-time mold and material change, call it $100,000. Net saving in year one: $525,000. Every year after: $625,000. The return on that investment speaks for itself. Whether it is worth doing, you can calculate.
But get the posture right. The change lives on the product, it stays after import, and you hold a binding ruling in hand. Keep those three, and the practice is a legitimate craft. Break them, and it is a $365 million fine.
By Lao Zhou, Logistics Cost Analyst







