Four Tariff Regimes in One Year: Protecting Margin When Your Landed Cost Won't Sit Still

Your Cost of Goods Changed Four Times This Year

If you import anything — finished goods, components, packaging — the number you plug into your margin spreadsheet has moved repeatedly in 2026, and not on a schedule you set. Here is the actual sequence.

• February 20: The Supreme Court ruled in Learning Resources, Inc. v. Trump that IEEPA does not authorize the president to impose tariffs, invalidating the reciprocal tariff program that had been in place since April 2025. Refunds began processing on a mass scale.
• February 24: A 10% global surcharge under Section 122 took effect, signed the same day the Supreme Court ruled. Statutorily capped at 15% and 150 days.
• July 24: The Section 122 clock ran out at midnight and Section 301 forced-labor tariffs of 10% or 12.5% on 60 economies took effect at 12:01 a.m. — covering locations that account for nearly all imports into the United States. Unlike Section 122, these have no statutory expiration.
• August 22: A 50% duty on roughly $20 billion of annual Canadian imports landed under Section 338 of the Tariff Act of 1930 — covering goods regardless of USMCA eligibility. Canada's dollar-for-dollar countermeasures follow on September 8.

That last one deserves a closer look, because it's the whole problem in miniature. The Section 338 duties were proclaimed on July 20 with an effective date of August 19. On August 18, they were suspended for three days while negotiations continued. Talks collapsed, and the tariffs took effect automatically on the 22nd. If you were routing inventory, your landed cost depended on which side of a date that moved twice with four days' notice your container cleared customs.

Meanwhile the net effect on the average household is still being revised in real time. The Tax Foundation currently estimates 2026 tariffs at an average tax increase of $840 per US household, down from $1,000 in 2025 — specifically because the struck-down tariffs "have not been fully replaced." That number is a moving target too.

The Problem With Committing to a Percentage in Advance

Here's where this collides with how most stores discount. A sitewide sale is a decision made well before it runs. You pick "20% off everything" in planning — six weeks out, maybe ten — and you pick it by looking at your current margin and deciding how much of it you can afford to hand back.

The trouble is that "your current margin" is an input with a shelf life now. A merchant who set a 20%-off Labor Day sale in early July was pricing against a 10% Section 122 world. By the time that sale ran, the goods were arriving into a 12.5% Section 301 world with no expiration date attached. The percentage was locked. The cost underneath it wasn't.

Merchants describe this better than any policy summary can. One small business owner sourcing from Vietnam, China and Mexico told the National Retail Federation during last year's swings:

"Because of the tariff volatility… 40% one day… 145% the next, my vendors don't know what their costs will be, so I don't know what mine will be."

That's not a sourcing problem. That's a pricing problem, and it hasn't gone away. A Harvard Business School survey this spring found the fog is genuinely disorienting: nearly half of respondents misjudged the tariff rate then in effect on imports from China, many by wide margins, with some owners describing calling suppliers weekly just to verify pricing.

The behavioral response is already visible in the data. Netstock's 2026 tariff research found 73% of SMBs say tariff uncertainty has pushed them to plan inventory further ahead than before — 29% significantly so. Merchants are lengthening their planning horizons on inventory. Most haven't done the same thinking about the discounts they've already committed to on that inventory.

A Blanket Discount Compounds the Error in Two Directions

A flat percentage is a blunt instrument in a volatile-cost year for two separate reasons, and they stack.

It ignores which products actually absorbed the hit

Tariff exposure is not evenly distributed across your catalog. The Section 338 action covers specific annexes — dairy, alcohol, motor vehicles, and a long tail that reaches products like wine, hockey sticks and cement. Section 301 rates split between 10% and 12.5% depending on origin. Some of your SKUs are made domestically and absorbed nothing at all.

A 20% sitewide discount applies the same margin giveaway to the SKU whose landed cost just jumped 12.5% and the SKU whose cost hasn't moved since 2024. On the first, you may now be selling below the number you'd have set as your walk-away price. On the second, you're leaving money on the table for no reason.

It applies to shoppers who never needed it

This is the cost we've written about before in the hidden cost of blanket discounts, and tariff pressure sharpens it. A sitewide code discounts every order in the window, including from the customers who had their card out at full price. If 60% of your sale-period revenue would have converted anyway, 20% of that 60% is pure donated margin — in a year where you have materially less margin to donate.

A Floor Is a Different Kind of Number

The alternative isn't "stop discounting." Demand is real and price-sensitive shoppers are real. The alternative is changing what number you commit to in advance.

With negotiated pricing, you don't set a discount percentage. You set an accept price: the actual dollar figure, per product, that you're willing to let that item go for. List at $120, set your floor at $90. A shopper who clicks "Make an Offer" and submits $95 gets an instant accept. One who offers $88 gets a single counter at your floor. One who offers $40 gets declined outright. The full accept-counter-decline logic is three plain rules, and none of them can settle below the number you typed.

Three things follow from that, and they matter specifically in a volatile-cost year.

It's an absolute number, not a ratio. "20% off" means something different every time your COGS moves; $90 means $90. When the Section 301 rates landed and your landed cost on an item went from $62 to $68, your old 20%-off price silently got worse. A $90 floor didn't move — you can look at it and immediately see whether it still clears your new cost. This is also why product-level floors beat percentage-based collection rules for your most exposed SKUs, a distinction worth understanding before you configure anything: see percent off vs. percent of.

It's per product, so it maps to actual exposure. You can hold a tight floor on the imported line whose duty just changed and run a genuinely aggressive one on domestic overstock, in the same store, at the same time. A sitewide percentage cannot express that difference at all.

It only fires for shoppers who asked. The offer widget doesn't touch the price for anyone who buys at list. The discount reaches only the segment that was price-sensitive enough to negotiate for it, and each accepted offer resolves to a single-use, single-order discount code — not a reusable code that outlives your cost assumptions by leaking to an aggregator site.

The honest caveat: you maintain the floor

Worth naming plainly, because it's the thing most tools in this space overstate. Lury does not read a cost feed and recalculate your floors automatically. There's no landed-cost integration quietly adjusting your numbers when a proclamation drops. You set the floor and you keep it current.

That's a real ongoing task — and also, deliberately, the point. Automatic repricing off a cost input is exactly the kind of mechanism that ends up selling inventory at a number nobody reviewed. What you get instead is one field per product, legible enough to audit in an afternoon. When a trade action lands, you open your exposed SKUs, compare each floor against current landed cost, and adjust the ones that need it. A maintenance habit, not a black box.

A Practical Rhythm for the Rest of 2026

1. Segment your catalog by tariff exposure, not by margin. Which SKUs cross a border, and under which action? That list is shorter than your full catalog and it's the list that needs attention.
2. Set floors from current landed cost, not list price. Start from what the unit actually costs you today, add the margin you need, and that's the floor. Working backwards from a percentage off list reintroduces the exact problem you're solving.
3. Treat trade announcements as a review trigger. When a proclamation, court ruling or effective date lands on a category you import, that's the prompt to reopen your floors on the affected products — the same way you'd reopen your reorder points.
4. Let the exposed SKUs stay off the sitewide sale. If you still run seasonal promotions, exclude the products whose cost basis is actively in flux and let negotiation handle those instead.

One footnote on the regulatory backdrop: scrutiny of pricing that shoppers don't see or agree to keeps growing. An offer a shopper actively submits, at a number they chose, sits in a very different conversation than a price adjusted invisibly around them — a distinction we unpack in coupon fatigue in 2026.

The Bottom Line

In a year with four distinct tariff regimes, an effective date that moved twice in four days, and active litigation over two of the statutes involved, the riskiest thing about "20% off everything" isn't the 20%. It's the everything, and the fact that you committed to it weeks ago against a cost basis that has since changed.

A per-product floor doesn't make trade policy predictable. It just means the number you're defending is a number you can actually see, on the products where it matters, adjustable the moment your costs move — and it only ever gets used by the shoppers who asked.

Lury starts at $4.99/mo with a 14-day free trial on every plan. Turn it on for your most tariff-exposed collection, set floors against what those units cost you this week, and stop pricing your promotions against last quarter's assumptions.