Every few months a tariff headline lands and every brand with an overseas factory has the same reaction: panic, then a group text asking if it's time to move production. Most of that reaction is aimed at the wrong number.
The tariff rate isn't the cost. The tariff rate applied to your actual landed cost structure is the cost, and those are rarely the same conversation.
A headline rate and a real duty bill are different documents.
News coverage reports the rate that applies to a country or a broad product category. Your invoice reports a duty calculated against the classification code your specific product falls under, at the value declared on that specific shipment.
Two brands importing what looks like the same knit polo can land on different duty rates because one used a fabric blend that falls under a different HTS code than the other. A few percentage points of spandex changes the classification. Nobody's cheating the system — the system is just more granular than a headline can be.
That gap is where a lot of bad decisions get made. A brand reads "tariffs on China are going up" and starts sourcing conversations before anyone's pulled the actual code their product ships under.
The real question is landed cost, not duty rate.
Landed cost is unit cost, plus freight, plus duty, plus any broker or handling fee, all the way to your warehouse door. A tariff increase moves one line in that stack. It doesn't automatically move the total by the same percentage, because freight and unit cost don't move with it.
This is the calculation that actually decides whether a factory relationship still makes sense. A five-point tariff increase on a garment where duty is already 40% of landed cost is a real problem. The same five points on a garment where duty is 8% of landed cost barely registers.
Nobody should be reacting to a rate. Everyone should be modeling landed cost before and after, for their actual product, at their actual volume.
Reshoring isn't free. It's a different cost stack.
The instinct after a tariff hike is always the same: bring it home, or at least bring it closer. That instinct is right sometimes and expensive other times, depending on what's actually being made.
Domestic and nearshore production trades duty exposure for labor cost, and for most apparel categories, labor is still the bigger number. A brand that moves a program to save on tariffs and lands on a factory with double the labor cost per unit hasn't solved the math. It's moved which line item is the problem.
Where domestic production genuinely wins: short runs, fast turns, and anything where a four-week ocean transit plus customs clearance would blow the calendar anyway. Where it doesn't: high-volume basics where labor cost dominates and a tariff hit still comes out cheaper than the alternative.
Multi-country sourcing is a hedge, not a slogan.
"Diversify your supply chain" has become a thing brands say without doing. Actually doing it means qualifying a second factory in a second country before you need it, not after a tariff notice makes the first one suddenly too expensive.
Qualifying a backup takes a real sample cycle, a real fit check, and a real production run at some point before it's the only option on the table. Brands that treat this as insurance, paid for in advance with a small test order, are the ones who can shift volume in weeks instead of months when a rate changes.
The ones caught flat-footed are always the ones who only had one factory and one country, and called that a sourcing strategy.
What actually belongs in the plan.
A tariff classification review before quoting, not after committing to a fabric. A landed cost model that updates automatically when a duty rate changes, so a headline can be checked against real numbers in an afternoon, not a week of spreadsheet archaeology. And at least one qualified alternative factory, in a different country, that's actually made your product before it's the only bid you've got.
None of that prevents a tariff increase. All of it prevents a tariff increase from becoming a crisis.
The rate is going to move. It always does. The only real question is whether your sourcing plan was built to absorb that, or built to be surprised by it.