Tariff Impact on Ecommerce Prices: A 200 Store Study Design

Every founder I talk to has a theory about what tariffs did to prices. Almost none of them have checked what their five closest competitors actually did.

That's the gap this post is about.

Here's the short answer on the tariff impact on ecommerce prices, from the best data that exists. Harvard Business School's Pricing Lab tracked daily prices on 359,104 products at five large U.S. retailers. Between March and October 2025, imported goods rose about 6.2% above their pre-tariff trend and domestic goods rose 3.6%. Retail pass-through reached 24% of the applied tariff after seven months. A New York Fed study puts total pass-through at about 26%. And at the border, the Kiel Institute found foreign exporters absorbed only 4% of the bill. So nearly all of the tariff lands on the U.S. side, but only about a quarter of it shows up on the shelf within seven months. The rest sits in somebody's margin.

Here's the catch. Those studies measured Walmart-sized retailers selling general merchandise. Nobody has published the same measurement for independent DTC brands on Shopify, where one founder sets the price and margins are a different animal.

So this is a study design, not a study. We haven't run it. There are no invented results below. What you get is the published research, the method for tracking 200 DTC stores, and a 10-store version you can run on your own competitors starting this week.

The tariff hits your landed cost at full strength. Your price only moves if you move it. The study is about who moved, by how much, and when.

What does the research say about tariff impact on ecommerce prices?

Four studies are worth your time. I opened and read each one. Here's what they found, in plain English.

Harvard's Pricing Lab (Cavallo, Llamas and Vazquez). The paper, Tracking the Short-Run Price Impact of U.S. Tariffs, is dated January 30, 2026, with data through October 1, 2025. It links daily prices from five major U.S. retailers to each product's country of origin and its 10-digit trade code. Imported goods rose about 6.2% relative to their pre-tariff trend. Domestic goods rose 3.6%. The average applied tariff in their sample hit 23.8%. Retail pass-through reached 24%, adding about 0.76 percentage points to the all-items Consumer Price Index. An earlier NBER version reported 20% through September, so the number was still climbing.

The New York Fed (Amiti, Heise and Weinstein). The Anatomy of Tariff Pass-Through into Consumer Prices, revised September 2026, finds about 26% of the tariff increase passed through to consumer prices. 64% of that comes from imported goods getting pricier. The other 36% is indirect: U.S. producers paying more for imported inputs, and domestic brands raising markups because their imported competitors got more expensive. That indirect part takes nine to twelve months to show up.

The Kiel Institute. Analyzing more than 25 million U.S. customs records covering nearly $4 trillion of imports from January 2024 to November 2025, Kiel found foreign exporters absorbed only 4% of the tariff burden. The other 96% was paid by U.S. importers and consumers.

Canada's retaliatory tariffs (Cavallo, Kostyshyna, Kryvtsov and Vieyra). Using daily prices from seven Canadian retailers, the February 2026 study found tariffed goods rose 6.3% by August 2025. Two details matter for DTC. When Canada removed the tariffs on September 1, 2025, prices unwound rapidly. And pass-through was substantially larger when products were labeled "Tariffed" at the point of sale.

Study	Data	Headline finding
Cavallo, Llamas, Vazquez (Harvard, Jan 2026)	359,104 products, 5 U.S. retailers, daily	Imports +6.2% vs trend, domestic +3.6%, 24% retail pass-through
Amiti, Heise, Weinstein (NY Fed, Sep 2026)	CPI items linked to import and producer data	About 26% total pass-through, 36% of it through domestic goods
Kiel Institute (Jan 2026)	25 million+ customs records, about $4 trillion	Exporters absorbed 4%, U.S. side paid 96%
Cavallo, Kostyshyna, Kryvtsov, Vieyra (Feb 2026)	7 Canadian retailers, daily	Tariffed goods +6.3% by Aug 2025, fast unwind after removal

Put those together and you get a clear picture. The tariff is almost fully paid at the border by U.S. importers. Retail prices absorb roughly a quarter of it within seven months. The gap between those two numbers is margin, and the Harvard team estimates U.S. consumers carried up to 43% of the tariff-driven cost after seven months, with the rest absorbed mostly by U.S. firms.

That "U.S. firms" line includes you.

Pass-through is a polite word for a hard question: of every tariff dollar, how much did you hand to the customer and how much did you eat?

Which categories saw the biggest price increases?

Here's the chart everyone asks for. It's the closest published thing to "median price change by category," and I want to be precise about what it is.

These are raw price changes on imported goods at five large U.S. retailers, from March 4 to October 1, 2025, from Table A5 of the Harvard paper. They're index changes, not medians, and they're not DTC stores. Treat them as the baseline a DTC study would be compared against.

Furnishings and household 5.94%

Miscellaneous goods 3.81%

Food and beverages 2.34%

Recreation and culture 1.88%

All sectors 4.28%

Source: Cavallo, Llamas and Vazquez (2026), Table A5, imported goods, raw price change March 4 to October 1, 2025, five large U.S. retailers. Not DTC data and not medians.

Furnishings and household goods led the pack. That's the category most dependent on Chinese imports, and it's where a lot of DTC lives: bedding, decor, kitchen, storage. Against the pre-tariff trend, imported furnishings ran 9.24% higher. Electronics barely moved because of broad exemptions. Food was mostly domestic.

Drill one level deeper and some narrow categories went wild. Relative to trend, carpets and floor coverings were up 41.54% on imported goods. Household textiles up 7.87%. Glassware, tableware and household utensils up 7.94%.

Country of origin mattered too. Goods from China rose 3.68% in raw terms against an applied tariff of 28.3%. Goods from Mexico fell 0.27%, with an applied tariff of just 1.4% thanks to USMCA exemptions. Same retailers, same months, very different bills.

One more finding I'd tape to the wall. Within each category, the cheapest quarter of imported products rose roughly 1.5 percentage points more than the most expensive quarter. The authors read this as thin-margin products passing through more of the cost. A premium DTC brand with fat margins has more room to absorb. That's a choice, not a law.

Where do tariffs stand in October 2026?

The rules moved twice this year, so any study that runs now needs the timeline right. Here's where things stand from the sources I checked today.

1. February 20, 2026. The Supreme Court declared the IEEPA tariffs unconstitutional. Per the Penn Wharton Budget Model, about $166 billion had been collected under that authority, and CBP had certified about $107 billion in refunds through August 21, 2026.
2. Late February 2026. A 10% global tariff under Section 122 replaced them.
3. July 24, 2026. Section 122 expired. New Section 301 tariffs of 10% or 12.5% on imports from 60 economies took its place, with exemptions for certain products.
4. Effective rate. Wharton puts the U.S. effective tariff rate at 6.7% as of July 2026, up from 2.3% in January 2025. Steel and aluminum products remain the most heavily tariffed category at 40.5%.

And the CBP trade remedies page, last modified October 1, 2026, still lists active Section 232 actions on autos, metals, timber, semiconductors and pharmaceuticals, Section 301 actions on China, Brazil and the forced-labor list, Section 338 duties on Canada, and the IEEPA refund process.

If you want what this does to a single parcel, I broke that down in the de minimis rule change for DTC brands. This post is pricing research, not customs or legal advice. Confirm your own product's rate with your broker.

Here's why the timeline matters for a study. It gives you three natural experiments in one year. A tariff went away (February 20). A flat 10% replaced it (late February). A different structure replaced that (July 24). Did DTC prices move at each date? Did any go back down? That's testable.

Why don't the big studies answer the question for a DTC brand?

Because a Shopify brand prices nothing like a big-box retailer.

Think about who sets the price at a big-box chain. A pricing team, a vendor contract, a promo calendar set months out. Now picture a $40,000-a-month bedding brand. One founder, one Shopify admin, one late-night decision.

Four differences make DTC its own question.

Margins are fatter. Plenty of DTC brands sell at 3x to 5x landed cost. The Harvard team notes that if import costs are half of a retail price, a 10% rise in border costs can only lift the retail price 5% even under full pass-through. For a DTC brand where landed cost is 25% of the price, that same tariff can only move the price a few percent. That's a big deal for how you read the numbers, and I'll show the math below.

Prices are sticky in round numbers. A big retailer can move $19.97 to $20.43. A DTC brand moves $68 to $74, or doesn't move at all. Price changes come in lumps.

The compare-at price is a second lever. Plenty of brands kept the list price and quietly shrank the "sale." Same sticker, smaller discount. A study that only watches the selling price misses that.

Bundles hide the change. A brand can keep the $68 sheet set and raise the price of the bundle, or drop the free pillowcases from it. I walked through why that's often the smartest move in tariff pricing strategy.

So the question for DTC is different. Not "what's the pass-through rate," but "which levers did founders pull, how hard, and did it stick?"

A big retailer passes a tariff through in pennies. A founder passes it through in one nervous decision about a round number.

How would a 200 store study of tariff price increases work?

Here's the design. It's built so anyone with a laptop and a spreadsheet could replicate it, and so nobody can accuse the results of being cherry-picked.

The sample: 8 categories, 25 stores each

200 stores, picked before any price data is pulled. Eight strata, chosen by tariff exposure and how much DTC lives there:

Stratum	Stores	Why it's in the study
Bedding and home textiles	25	Furnishings led the Harvard data, high China exposure
Kitchen and tableware	25	Glassware and utensils up 7.94% on imports vs trend
Home decor and furniture	25	Heavy import share, big-ticket prices
Apparel and accessories	25	Clothing accessories up 12.73% on imports vs trend
Beauty and personal care	25	Mix of U.S.-made and imported, good control group
Supplements and pantry	25	Mostly domestic, the expected low end
Pet products	25	Imported hard goods next to domestic consumables
Outdoor and fitness gear	25	Metal-heavy products touch Section 232

Inclusion rules, set in advance: the store runs on Shopify, sells at least 20 products, has been live since January 2025 or earlier, ships to U.S. customers, and serves its catalog at /products.json. Out: marketplaces, dropshippers with 5,000 SKUs, and anything that isn't a brand.

Within each stratum, aim for a mix of import-heavy and U.S.-made brands. That gives you a built-in control group, the same way the Harvard team compared imported and domestic goods. If U.S.-made candles rose as much as imported candles, the tariff isn't the whole story.

The data: list prices, weekly, from the store's own catalog

Most Shopify stores publish their whole catalog as JSON. As this walkthrough shows, you add /products.json to the store's domain and get every variant with its price, compare-at price, availability and publish date, 250 products per page. Some stores disable it, so expect to replace a few.

For each variant, each week, record:

• Selling price and compare-at price
• Availability (out-of-stock variants distort averages)
• Variant weight or size, to catch shrinkflation
• New and removed products (new launches carry new prices, not changed ones)

Plus, once a month, a manual pass on each store for the free-shipping threshold and any "tariff surcharge" line at checkout.

The baseline and the event windows

Prices before the first event are the baseline. Then measure the change at fixed windows after each policy date: 10, 30, 60 and 90 days. The Harvard event study found affected prices up about 1.0% by day 10, 1.5% by day 25 and about 2.0% by days 60 to 90 after the March 2025 announcement, so those windows line up with how fast prices actually moved.

For 2026, the events are February 20 (IEEPA struck down), late February (Section 122 begins) and July 24 (Section 301 replaces it). A study started today can't go back and collect February prices by hand, which is the honest limit. Any store you snapshot today is a baseline for the next change, not the last one.

The measures

1. Share of products with any list-price change in each window, by stratum.
2. Median price change by category, counting only products that existed in both snapshots. This is the chart a finished study would publish next to the Harvard one above.
3. Size distribution of the changes. How many moved 1% to 3%, 3% to 8%, more than 8%. Lumps tell you about round-number pricing.
4. Compare-at gap change. Did the "sale" shrink while the sticker stayed?
5. Direction after relief. Did any prices fall after February 20? Canada says prices can unwind fast. Nobody has checked DTC.
6. Labeled surcharges. Count stores that show a tariff line. Canada found pass-through was bigger when the tariff was labeled.
7. Dollar pass-through on a subsample. For 40 products where origin, trade code and price are clear, estimate the tariff in dollars and compare it to the dollar price change.

That last one is the measure I care most about. Here's why.

Why does dollar pass-through matter more than percent?

Because a DTC brand can pass through 100% of a tariff and a percent-based study will say it passed through 21%.

Run the math on a store like this, a hypothetical bamboo sheet set. Retail price $68. Customs value $14 a set. A 12.5% tariff on $14 is $1.75.

Now say the founder raises the price by exactly $1.75, to $69.75. That's every tariff dollar handed to the customer. Full pass-through, in dollars.

As a percent of the retail price, that's $1.75 divided by $68, a 2.6% increase. Divide 2.6% by the 12.5% tariff rate and a percent-based study records pass-through of about 21%.

Same decision. Two very different headlines.

Now flip it. Say the founder rounds up to $74. That's $6 more, against a $1.75 tariff. That's about 343% of the tariff in dollars, and an 8.8% price increase. A percent-based study records it as 70% pass-through. In plain English, the founder used the tariff as cover to fix a price that was too low.

Tariff per set $1.75

Raise A (exact pass-through) $1.75

Raise B (round up to $74) $6.00

Hypothetical $68 sheet set, $14 customs value, 12.5% tariff = $1.75 per set. Raise A = +$1.75 to $69.75. Raise B = round up to $74.

This is the thing the big studies flag and can't fully measure: the Harvard team notes that percent pass-through can look low even when dollar pass-through is substantial, and measuring it in dollars needs cost data most researchers don't have. A DTC study can get closer because landed cost is a few line items you can estimate. If you haven't rebuilt yours since the rules changed, start with the landed cost calculation worksheet.

If your landed cost is a quarter of your price, a 12.5% tariff is a 3% price problem. Plenty of founders reacted like it was a 12.5% problem.

What would wreck the study?

Six traps. Each one has a fix built into the design.

Sale noise. A store runs 20% off for a weekend and your weekly snapshot lands on it. Fix: track the compare-at price separately, and use the most common price across a four-week window as the "regular" price.

Variant churn. A brand retires the old sheet set and launches a "new" one at $79. That's a price increase the matched-product method can't see. The Harvard team hit a version of this too: their index didn't link new and old models, so markdowns on older items show up as steady declines. Fix: flag replacements by title and image, and report them separately.

Shrinkflation. Same $34 price, 8 ounces instead of 10. Fix: record variant weight or size in every snapshot and compute price per unit.

Survivorship. Stores that went under during the tariff year drop out, and they're probably the ones that ate the most margin. Fix: keep dead stores in the sample with an exit date. They're data.

Origin mistakes. Calling a U.S.-made product imported wrecks the control group. Fix: use the product page's "made in" line first, AI second, and hand-check a sample.

Confusing correlation with cause. Prices rise for lots of reasons: freight, ad costs, a new supplier. Fix: the U.S.-made control group, plus the event windows. If imported and domestic prices in the same category move together on the same dates, look elsewhere.

None of this is exotic. It's the same discipline you'd apply to reading your own profit per product, just across 200 stores instead of one.

How is AI changing the way prices get tracked?

Here's the part I find genuinely new. The Harvard study itself runs on AI.

For products without a findable country of origin, the team used generative AI models to identify where each product was made. Against products with known origin, that approach correctly split domestic from imported goods 88% of the time and named the exact country 85% of the time. They then used an AI model to classify each product into a 10-digit Harmonized System trade code, walking down the category tree one level at a time, and pulled the tariff rate for each code from public data. That's how a three-author team tracked origin and tariff exposure on 359,104 products.

You can run the same pattern on a much smaller scale.

1. Origin tagging. Paste a competitor's product descriptions into an AI assistant and ask it to tag likely country of origin with a confidence level. Check every low-confidence tag by hand.
2. Trade code drafts. Ask it to draft a 10-digit code for your own top 20 products. Then hand the list to your customs broker. A draft is a starting point, never the filing.
3. Change detection. Each week, give it last week's catalog snapshot and this week's, and ask for a table of every variant whose price or compare-at price changed, with old price, new price and percent change.

That third one is the whole competitor-tracking job, in minutes instead of an afternoon.

Here's the honest limit. AI is fast at sorting and spotting. It can't tell you why a competitor moved, and it can be confidently wrong on origin. The Harvard numbers say roughly 1 in 8 domestic-versus-imported calls was wrong. Build the hand-check in.

AI made a 359,104-product price study possible for three researchers. It makes a 10-competitor study possible for you, tonight.

How do you run a 10 store version on your own competitors?

You don't need 200 stores to make better pricing decisions. You need the 10 a buyer compares you against.

Start by finding them. If you aren't sure who your real Shopify rivals are, this will map them for you.

Then:

1. Pick 10 stores that sell your hero product's closest substitutes. Mix import-heavy and U.S.-made brands if you can.
2. Snapshot each catalog at /products.json into one sheet: store, product, variant, price, compare-at price, available, date.
3. Repeat every Monday. Same day, same time, so sale weekends don't drift into the data.
4. Have AI flag every change. Price, compare-at, new products, removed products.
5. Tag each changed product as likely imported or U.S.-made, and note the nearest tariff date.
6. Add your own store as the 11th row. You're part of the market your buyers see.
7. Read it monthly. Who raised, how much, and did they hold the price or quietly bring the discount back?

After eight weeks you'll know something no published study can tell you: what your buyers' other tabs cost now.

What does this mean for your own pricing?

Here's the thing. The research says retail prices absorbed about a quarter of the tariff in seven months, and that cheaper products passed through more. Read that from the founder's chair and it says two things.

One: your competitors have probably raised less than you fear. The big-retailer numbers are single digits, and DTC brands with fat margins had even more room to absorb. So the "everyone's raising prices" story may be giving you cover that isn't there.

Two: eating the tariff is a choice with a price tag. Every dollar you don't pass through comes out of margin on every order, forever, unless you fix it somewhere else.

So run the math on your own page before you touch the price. Here's a hypothetical. A store sells a $60 product at a 2.0% conversion rate. Revenue per visitor is $1.20. On 10,000 visitors, that's $12,000. Raise the price 5% to $63, and say conversion rate dips to 1.9%. Revenue per visitor is $1.197. On 10,000 visitors, that's $11,970.

Same revenue, near enough. But every one of those orders now carries $3 more margin. That's the move most founders are too scared to test.

The calculator shows revenue, not profit. Subtract your new landed cost before you decide. And if the drop in conversion rate is bigger than that, the problem usually isn't the price. It's a page that never explained why the product is worth it. I wrote about how buyers judge "expensive" in Shopify price anchoring.

Here's what a page that can carry a higher price looks like in real numbers. One Amazon product we moved to Shopify sold at $17.95 on Amazon with $3.83 of net profit per sale. On Shopify, priced at $24.95 with a page built to sell a bigger order, average order value reached $101.38 and net profit hit $48.27 per sale, after paying for ads and inventory. See the full case study numbers. Real client numbers, not typical results, and not a promise of what your store will do.

That's the polarity. A page that can't defend its price turns every tariff into a margin leak. A page that can turns a tariff into the best excuse you'll ever get to charge what the product is worth. If you suspect the leak is bigger than the tariff, read how one store found a $4,200 monthly leak inside $60K in sales.

Tariffs didn't create weak pricing. They exposed it.

And if you're planning a Black Friday markdown on top of a tariff-squeezed margin, run the discount depth math first. Thin margin plus deep discount is how a record revenue weekend loses money.

A note on method

Nothing in this post is a result from our own study. We have not run the 200 store study yet. The research numbers come from the four studies linked above: the Harvard Pricing Lab paper by Cavallo, Llamas and Vazquez (data through October 1, 2025), the New York Fed staff report by Amiti, Heise and Weinstein (revised September 2026), the Kiel Institute customs analysis as reported by Fortune, and the Canadian retaliatory tariff study. The 2026 tariff timeline comes from the Penn Wharton Budget Model (September 9, 2026) and CBP's trade remedies page (modified October 1, 2026). The bamboo sheet set and the $60 product are hypothetical, built to show the method.

FAQ

How much have tariffs raised ecommerce prices? At five large U.S. retailers, imported goods rose about 6.2% above trend from March to October 2025, and domestic goods 3.6%. Nobody has published the DTC version yet.

What is tariff pass-through? The share of a tariff that shows up in the customer's price. Harvard estimated 24% at retail after seven months. The New York Fed found about 26% including domestic goods.

Do retailers pass the full tariff on to customers? At the border, the U.S. side paid 96%, per the Kiel Institute. At the shelf, only part of it shows up, because the tariff hits only the imported share of the retail price.

How fast do online prices change after a tariff? About 1.0% by day 10, 1.5% by day 25 and 2.0% by days 60 to 90 in the Harvard data. Indirect effects on domestic goods take nine to twelve months.

How can I track competitor price changes from tariffs? Snapshot 10 competitors' /products.json catalogs every week and have AI flag every price and compare-at change.

Did prices fall after the Supreme Court struck down the IEEPA tariffs? No published U.S. study I've found measures it yet. In Canada, prices unwound rapidly after tariffs were removed in September 2025.

What to do next

Open a sheet tonight. List the 10 stores your buyers compare you against, pull each one's /products.json, and save the snapshot with today's date. Do it again next Monday. In two weeks you'll have the first real answer to the question everyone's guessing at: what your market actually did with its prices.

---

Book Your Profit Audit

Bring your hero product's price, landed cost and conversion rate to a free profit audit, and we'll show you whether your page can carry a tariff-inclusive price or is leaking buyers at the old one. Then we'll show you how to rebuild a high-converting product sales page in less than 15 minutes.

Book Your Profit Audit →

Or go here to check it out → revenueflows.ai

P.S. Eating the tariff bleeds margin quietly. A page that earns the new price keeps the margin and the buyer.