HS codes for small importers: how classification actually works

Somewhere on your supplier's proforma invoice there's an "HTS code." Treat it as a hint, not an answer: it's often the code used for the Chinese export declaration, built for a different customs administration's purposes, and you can't rely on it for a US entry. What you can rely on is the method — a fixed code structure and six rules of interpretation that CBP applies the same way every time. Learn those and you can classify a simple product yourself, sanity-check what suppliers send you, and know exactly when to hand the problem to your broker.

(The basics — what an HS code is, the lookup tools, and why misclassification is expensive — are covered in the customs duties guide. This article owns the structure and the method.)

The shape of the code: chapter → heading → subheading → suffix

A US entry code is ten digits, built in layers:

The first six digits are the international HS — the same everywhere, which is why a code quoted by a factory in China and one quoted by your broker should at least agree that far (Flexport, "What are HTS codes?"). The US adds four more digits for the 10-digit HTS used at entry, administered by the US International Trade Commission. Enforcement sits with Customs and Border Protection, which publishes its classification guidance in the Informed Compliance Publications — worth knowing, because the practical upshot is that CBP's reading of the rules, not your supplier's, is the one that counts at the border.

The GRI: six rules, applied in order

Classification isn't a keyword search. It's the General Rules of Interpretation (GRI) — six rules applied hierarchically, in order, and you stop at the first one that resolves the question (CTP Inc. has a good plain-English walkthrough).

GRI 1 does most of the work. Classify by the terms of the headings themselves and the section and chapter notes — which are legally binding parts of the tariff, not commentary. Before you argue about which heading "fits better," read the chapter notes: exclusions there kill candidate headings before any deeper reasoning starts.

The other rules extend coverage rather than compete: GRI 2 handles incomplete or unfinished goods (classified as the complete article when they have its essential character) and mixtures; GRI 4 and 5 are narrow fallbacks for goods with no specific heading and for cases and packaging. For finished consumer goods you'll mostly live in GRI 1, 3, and 6.

GRI 3 breaks ties — when goods could fall under two or more headings:

• 3(a): the most specific heading wins over the one that covers a broader class of goods. A heading naming the actual product beats a catch-all.
• 3(b): if specificity doesn't settle it, classify by essential character — the character of the material, role, bulk, value, or function that determines what the good essentially is. This is the rule that decides most composite products.
• 3(c): last resort — when goods can't be distinguished by (a) or (b), they go to the heading that appears last in numerical order.

GRI 6 repeats the exercise at subheading level. The same logic — specificity, essential character — applies within each heading to pick the right subheading and, in the US, the right statistical suffix. This is where most 10-digit errors happen: importers get the chapter and heading right and guess the tail.

A worked walk: the silicone dog-treat pouch

Take a fresh product: a silicone dog-training treat pouch with a belt clip, sold as an accessory for walkers. No definitive code printed here — the point is the method. Your broker confirms the final digits against the current HTS.

Step 1 — chapter candidates. The pouch is silicone-bodied. Chapter 39 (plastics) covers articles of plastics and silicone. But a pouch worn on a belt also smells like chapter 42 (articles of leather; travel goods — which covers some bags and similar articles), and if a textile panel dominates, chapter 63 (other made-up textile articles) enters the frame. Three candidate chapters; GRI 1 doesn't resolve it yet because the notes don't clearly exclude any of them for a composite good like this.

Step 2 — GRI 3(a), specificity. Is there a heading that names this thing? No — no heading says "dog treat pouch." The candidates are generic: articles of plastics (3926, "other articles of plastics"), possible bag-type headings under 42, possible made-up textile goods under 63. No candidate is specifically the product, so specificity alone doesn't settle it and we move on.

Step 3 — GRI 3(b), essential character. Ask what gives the pouch its character across role, bulk, value, and function:

• Role and function: it's a washable food container that clips to a belt. The container role — holding treats, wiping clean — is performed by the silicone body.
• Bulk: the silicone forms most of the physical volume; the clip and any strap are small relative to the body.
• Value: the silicone molding is typically the cost driver; hardware is minor.

Every axis points the same way: essential character of a plastics article. That lands the walk plausibly in heading 3926 — other articles of plastics — rather than the leather-goods or textile chapters. Had the product been built differently (a textile pouch with a silicone lining, say), the same reasoning would point the other direction, which is exactly why method beats keyword lookup: two visually similar pouches can classify differently.

Step 4 — hand off the tail. The correct 6-digit subheading and the four US statistical digits depend on construction details this walk doesn't settle, and the US tail changes with statistical revisions. Your broker confirms the final digits — and if you're self-filing, you confirm with them before entry, not after.

Don't trust the supplier's code — verify it independently

Worth repeating with the reason attached: the code on a Chinese proforma is usually the export code for China's declaration system. The first six digits will often match the international HS, but "often" is doing heavy lifting, and the 10-digit tail never transfers — it's US-specific. The legal responsibility for classification at entry sits with the importer of record, not the factory. So run the method yourself for anything you import repeatedly, or have your broker verify it, and treat supplier codes as a starting hypothesis. The duty rate that falls out of the code feeds straight into your landed cost, so an unverified code quietly corrupts every margin calculation built on it.

When a binding ruling is worth the trouble

For genuinely ambiguous goods — or anything you import repeatedly at meaningful volume — CBP offers binding rulings: written classification decisions issued by CBP for a product described in enough detail, which the agency applies at entry for that good. The request process runs through CBP's eRulings system (as of October 2026; check CBP's site for current submission mechanics and turnaround, which varies). You don't need one for ordinary goods a broker classifies daily. The trigger is repetition plus ambiguity: the same hard-to-classify item, entering month after month, where a surprise reclassification at exam would reprice a year of entries.

Where this fits in your import workflow

Practically: classification is a once-per-product task with a long shelf life. Do the GRI walk (or the broker call) when you first commit to a SKU, record the full 10-digit code with your product record, and reuse it on every entry — proforma, ISF (which needs the 6-digit minimum), and final customs paperwork should all carry the same code. Divergence between documents is one of the classic exam triggers, which makes consistency nearly as valuable as correctness.

Common questions

What's the difference between HS and HTS codes? The first six digits are the international HS — universal across countries. The US adds four more digits for the 10-digit HTS used at entry, administered by the US International Trade Commission (Flexport). Same product, same first six digits worldwide; the country-specific tail is where entries are actually filed.

Can I use my supplier's HTS code for US customs? Treat it as a hypothesis, not an answer — it's usually the code for the export declaration in the supplier's country, not US entry. Verify independently or have your broker confirm. The importer of record owns the classification either way, penalties included.

What is GRI 3(b), essential character? When two headings could fit a composite good, classify by what gives it its essential character: the material, role, bulk, value, or function that defines what it essentially is. A silicone-bodied treat pouch with a small belt clip classifies as a plastics article because bulk, function, and value all point that way.

What are the six GRI rules, in order? GRI 1: classify by heading terms and chapter notes. GRI 2: incomplete goods classify as complete. GRI 3: when two headings fit, most specific wins, then essential character, then last heading. GRI 4: default for the truly unclassified. GRI 5: cases and packaging. GRI 6: the same rules at subheading level. Your broker confirms the final digits.

Do I need a CBP binding ruling? Not for ordinary goods. Consider one when the same ambiguous item imports repeatedly at volume, or when competent classifiers genuinely disagree — a binding ruling fixes CBP's classification for the described good and removes the reclassification argument at exam. Requests go through CBP's eRulings system.

Next step

Pull the last entry for your highest-volume import SKU and decompose the 10-digit code yourself: chapter, heading, subheading, suffix. If you can't reconstruct why it landed there, send the product's construction details — materials, proportions, function — to your broker and ask for the reasoning, not just the code (how brokers work and what to expect from one). Do it once per product and the answer holds for years.