The BSR model has three data layers: depletion signals from Amazon rank data, replenishment signals from import records, and factory concentration signals from Customs shipment patterns. This piece introduces a fourth: the tariff layer — the structural cost variable that explains why some BSR patterns don’t resolve the way the three-signal model predicts.
In the first four pieces of this series, I described a methodology for reading Amazon supply chains. The short version: Amazon’s Best Sellers Rank measures how fast a product is selling through existing inventory, not how much demand there is for it. Import records tell you when new inventory is arriving to refill what sold through. Customs data tells you which factories are filling which brands’ pipelines. Reading those three signals together gives you a supply-chain view that most Amazon analysis tools don’t provide.
This piece adds a fourth layer.
It is not a new data source. It is a structural cost factor that modifies how you interpret the other three signals — specifically, why some brands respond to a replenishment opportunity slowly, why some factory relationships don’t survive a tariff change, and why a brand going out of stock sometimes reflects margin pressure rather than inventory mismanagement.
That layer is the Section 301 tariff on Chinese-origin goods.
The depletion/replenishment model assumes that when a brand’s BSR is deteriorating — rank drifting upward, inventory thinning — the natural response is to book more containers and refill. In the absence of other constraints, that is what happens. The model works well when the only variable is timing: did the brand book early enough, and is the factory ready?
The depletion/replenishment model assumes that when a brand’s BSR is deteriorating — rank drifting upward, inventory thinning — the natural response is to book more containers and refill. In the absence of other constraints, that is what happens. The model works well when the only variable is timing: did the brand book early enough, and is the factory ready?
Tariffs introduce a second variable: economics. A brand running a 25% Section 301 tariff on a $40 factory cost is paying $10 per unit in duty before ocean freight, US freight, or Amazon fees. On a product selling for $59 with a 15% Amazon fee, the unit economics get tight quickly. When demand is rising and inventory is thin, the correct move is to reorder. But if the fully-loaded landed cost means that refilling at current factory price produces a margin that doesn’t support the next six months of operating costs, brands hesitate. The replenishment signal weakens not because the factory can’t ship, but because the importer isn’t pulling the trigger on the purchase order.
That hesitation is visible in the import records before it is visible in the BSR data. The container stops arriving; the BSR drift that follows is the consequence, not the cause.
Consider the blender category. In piece #1, I described how BSR velocity data for countertop blenders shows a predictable depletion pattern ahead of Q4, with replenishment containers from Indonesian and Chinese factories arriving in September and October to fill the peak.
Consider the blender category. In piece #1, I described how BSR velocity data for countertop blenders shows a predictable depletion pattern ahead of Q4, with replenishment containers from Indonesian and Chinese factories arriving in September and October to fill the peak. That pattern holds in normal conditions.
In 2019, Section 301 tariffs took effect on Chinese-origin food-preparation equipment. Countertop blenders, classified under HTS 8509.40, were caught in this. The rate for most blender subheadings under List 3 was 25%. The brands sourcing blenders from Chinese contract manufacturers suddenly had a $10 per unit cost increase on a $40 factory item.
What happened to the BSR data? Two things. First, brands that had established Indonesian-origin supply chains — which existed before 2019 for several major blender players — saw their cost position improve relative to China-sourced competitors, because Indonesian-origin goods are not subject to Section 301 duties. The import records show a shift: Indonesia’s share of blender container shipments to the US grew between 2018 and 2021. Second, brands that remained China-sourced and could not absorb the margin hit began either raising retail prices (slowing their BSR velocity on the downward slope as consumers price-shopped), or booking lighter container loads to test new pricing levels, which weakened their in-stock position.
Both of those dynamics are visible in the three-signal model — but without the tariff layer, you’d read them as inventory planning errors or demand softening, rather than as margin-pressure responses to a structural cost change.
The tariff layer also modifies how you read the factory concentration signal. In piece #3, I described how Customs data shows which factories serve multiple brands in the same category. A factory in Guangdong that ships blenders to four US brands under different labels is a concentrated point of supply-side risk: a disruption at that factory affects all four brands.
The tariff layer also modifies how you read the factory concentration signal.
In piece #3, I described how Customs data shows which factories serve multiple brands in the same category. A factory in Guangdong that ships blenders to four US brands under different labels is a concentrated point of supply-side risk: a disruption at that factory affects all four brands simultaneously. That concentration is visible in the import records.
Tariffs add a migration dynamic to that picture. When Section 301 duties hit a Chinese factory that serves multiple US brands, those brands face the same incentive: find a way to move production to a non-tariffed geography, or absorb the duty. The factory that served four US brands in 2018 may have seen some of those brands quietly begin sampling from Vietnamese or Indonesian alternatives by 2020. The import records show that migration as new factory relationships appearing for those brands — smaller initial shipments, testing the new source, not yet replacing the Chinese volume but running in parallel.
Reading the factory map over time — not just the current state — shows which brands have successfully diversified their factory concentration risk and which are still running single-source from a China-origin factory under full Section 301 exposure.
Watch three patterns: BSR deterioration with no import gap, signaling a demand-side or listing-quality problem where tariffs don't apply; import gap without demand softening, signaling supply disruption possibly from a factory, financing, or tariff pressure; and a factory relationship change with brand persistence, the tariff-migration pattern where the brand reprices its supply chain.
Pattern 1: BSR deterioration with no import gap. The brand’s BSR is drifting but the import records show containers arriving on schedule. This is not a replenishment failure; it is a demand-side or listing-quality problem. The tariff layer doesn’t apply here.
Pattern 2: Import gap without demand softening. The brand’s BSR was stable, containers stop arriving, BSR deteriorates. The question is why the containers stopped. If demand is still strong (Google Trends stable, review velocity maintained, adjacent brands still selling), the most likely explanation is a supply disruption — which could be a factory issue, a financing issue, or margin pressure from tariffs producing a pause in purchase orders. The tariff layer is relevant if the brand is China-sourced and operating in a high-duty subheading.
Pattern 3: Factory relationship change with brand persistence. The brand’s BSR position holds, but the factory sourcing visible in import records shifts — new factory, different port of origin, sometimes a new country of origin. This is the tariff-migration pattern. The brand is repricing its supply chain. Watch whether the new factory achieves the same shipment cadence as the old one within two to three container cycles; if the cadence weakens, the migration is still in progress and the replenishment risk is elevated.
Not every BSR deterioration is tariff-related. A brand that goes out of stock in February in the blender category is probably not responding to March tariff pressure; it may have simply under-bought for Q4 and sold out. The tariff layer is most useful as an explanatory variable when the import gap and the demand picture are inconsistent — when you.
Not every BSR deterioration is tariff-related. A brand that goes out of stock in February in the blender category is probably not responding to March tariff pressure; it may have simply under-bought for Q4 and sold out.
The tariff layer is most useful as an explanatory variable when the import gap and the demand picture are inconsistent — when you see a brand that looks like it has demand, looks like it has a factory relationship, and is still not restocking. In those cases, the question of what the brand’s landed cost math looks like given the current tariff rate on its product category is often the missing piece.
If the fully-loaded landed cost at current factory pricing and Section 301 rates produces a unit economics picture that’s underwater at the brand’s current Amazon price, the brand is making a rational decision not to restock — it’s waiting for the price to clear, or for the factory to renegotiate, or for a new tariff exemption that isn’t coming. That’s not bad inventory management. That’s a structural constraint that the three-signal model alone won’t diagnose.
In the next piece, I’ll walk through a specific category where the tariff layer is doing significant explanatory work right now — and show what the full four-layer read looks like in practice.
Subscribe below to get it when it publishes. And if you have a category where the BSR/import data isn’t giving you a clean read — this is often why — use the snapshot form to send it over: https://tally.so/r/Y5xlOW?source_channel=substack_sc101_05
Notes on sources
The tariff rate figures in this piece reference the USITC Harmonized Tariff Schedule database (https://hts.usitc.gov/) and USTR Federal Register notices implementing Section 301 tariff actions on Chinese-origin goods (Lists 1–4, 2018–2025). Blender import sourcing geography shifts (Indonesia vs China, 2018–2021) are based on US Customs bill-of-lading data available through ImportYeti and ImportGenius. Specific brand or factory attributions from Customs data are not included in this piece; the pattern described is a category-level structural observation.