CBP pushes Entry Type 13 mail filing deadline to Jan 20, 2027, a 90-day extension

The Analysis

Start with the parcel, because every rule in this business eventually lands on one. A 1.2 kilogram package leaves a consolidator on the Chinese side, is handed to a foreign postal operator, lands at one of the international mail facilities CBP runs with the postal service, and then goes one of two ways. It either leaves the building as an informal entry and goes back into the postal network for delivery, or it is told it does not qualify for informal treatment and has to be entered formally, which is a different job with a different bill. The October 14 announcement does not touch that fork. It moves the date the fork becomes compulsory.

Two dates and one list carry the weight. October 22, 2026 became January 20, 2027, and that is the 90-day extension. The list is what decides whether a postal parcel clears cheaply or expensively: from that date, shipments subject to Section 201, Section 232 or Section 301 duties, or falling under a partner government agency such as the Food and Drug Administration, have to be filed electronically through the Automated Commercial Environment, CBP's ACE system, using Entry Type 13 in order to qualify for informal entry. Miss the filing and you are not fined in the abstract. You are pushed into formal entry.

The sentence most summaries buried sits near the end of the notice. The extension does not restore the duty-free de minimis exemption for international postal shipments, and nothing about the 800 dollar threshold has come back. That is the line that decides whether you spend the next 90 days preparing or celebrating, and the two choices lead to very different Januarys.

Get the tariff number right before you argue about duty relief. That habit is the right one to bring to this notice, because the first wall is not the fee and it is not the inspection. It is the 10-digit HTSUS code. A postal parcel has travelled for decades on a description: phone case, plastic toy, knitted top. A 10-digit code is not a description, it is a legal position, and somebody has to be willing to defend it eighteen months later when CBP pulls the file.

Work that out in hours before deciding it can wait. Assume you ship 300 postal parcels a day at an average declared value of USD 40, sitting on 200 active SKUs. Classifying 200 SKUs to ten digits, with the short written reasoning you will want when someone challenges it, is not data entry. At 20 minutes each, that is roughly 67 hours of competent time, which is about three and a half weeks of one person's calendar. You have 13.

The tenth digit is its own problem. The first six are the international heading, the next two are the US tariff lines, and the last two are a statistical suffix that most product catalogues have never carried. Plenty of systems hold eight digits and stop there. If yours is one of them, you are not doing a data clean-up, you are doing a data build, and 200 SKUs is 200 decisions rather than 200 keystrokes. Budget for the decisions, not the typing.

After the code comes the rest of the filing, and here I would stop thinking of it as a form and start thinking of it as a dataset. CBP listed what an Entry Type 13 filing has to carry: the foreign postal operator's tracking number, the sender's name and address, the 10-digit HTSUS classification, country of origin, value, duty rate and total duty owed. Seven fields. Four of them, tracking number, sender, origin and value, are created upstream of your broker, by whoever packed the parcel and whoever took the order.

Country of origin is the second field people fill in on autopilot, and it is the one most likely to be wrong in a way that matters. The habit in the postal channel is to print the country of despatch, which is not the same thing. Section 301 duties attach by origin, not by where the parcel was posted, and a product assembled in a third country from Chinese components is an origin question, not a label question. If you have not revisited your origin statements since the remedy lists started expanding, this is the window to do it. A wrong origin on a filed entry is a false statement; a wrong origin on a postal label used to be nobody's problem.

Value is the third one, and the argument is always the same: what exactly is being declared. Platform transaction price, invoice value, or a figure net of discounts and freight, and whether shipping sits inside or outside the declared amount. Declared value on a postal parcel has historically been a formality. From January 20 it is an input to the duty rate and the total duty owed, both of which you now have to state in the filing. Pick one basis, write it down, apply it to every parcel, and make sure the figure on the invoice equals the figure in the dataset. If the two differ by a rounding rule, expect the question.

The sender field is where I would expect the first rejections to come from. On a marketplace order, the name on the parcel is often a consolidator, a trading company or a fulfilment house, not the merchant whose listing the buyer clicked. File a sender that does not reconcile with the commercial reality of the sale and you have built a mismatch that surfaces later, at a post-entry review, at a refund claim, or the first time a buyer asks for a return. Documents that do not reconcile are not a cosmetic problem. Everything downstream of them is trouble.

Bigger than any individual field is the question of who may file at all. CBP was explicit: only the owner or purchaser of the merchandise, or a licensed customs broker designated by the owner, purchaser or consignee, may file an Entry Type 13 entry. The postal operator cannot file it for you. The platform cannot file it for you. A forwarder cannot file it unless a licensed broker is properly designated somewhere in the chain. In a channel where four parties touch a parcel and each of them assumes another one handled the entry, that is the most expensive sentence in the document.

Here is the part the headlines did not write, and it is the part I would build the next three months around. The extension moves the date on which Entry Type 13 becomes mandatory. It does not create a grace period in which the old duty-free habits return, and CBP said that in plain words. What the notice also does not do is say what the fallback is, between now and January 20, for a postal parcel that owes Section 301 duty. CBP says current international-mail entry procedures remain in place.

At the mail facility, the current procedure for a parcel carrying trade-remedy duty is not release it the way we always did. The realistic answers are file Entry Type 13 anyway, or go formal, and both cost more than the habit. A shipper who reads 90-day extension as 90 more days of the old way is betting margin on a sentence that is not in the notice. The gap between what a document says and what a facility does with it has cost me more afternoons than I care to admit.

Put a number on that gap. Keep the 300 parcels a day assumption and 90 days is 27,000 parcels. If 15 per cent of them, so 45 parcels a day, carry a 301 classification or an agency interest, roughly 4,050 parcels are exposed before the rule even takes effect. Assume formal entry runs USD 30 a parcel more than informal in broker handling and documentation, and take your own broker's quote rather than my number; that is about USD 1,350 a day, or roughly USD 121,500 across the window. That is what it costs to read a delay as a holiday.

Who carries it. Not the large importers, because they do not move meaningful volume through the post, and they have brokers, bonds and somebody whose actual job is the entry. The exposure sits with the small and mid-size direct-to-consumer shipper, the marketplace seller, and the overseas warehouse that tops up stock in small parcels because that is what the cash flow allows. A shipper with 300 parcels a day does not have a customs department. He has whichever person in the office also does the buying, and that person is about to become a classifier.

The partner agency limb deserves more attention than it is getting. The notice names the Food and Drug Administration, the Department of Agriculture and the Environmental Protection Agency as examples, which means the affected population is wider than the remedy lists alone. Food-contact articles, cosmetics, dietary supplements, certain electronic products, anything with a plant or animal derived component, anything sitting inside pesticide or chemical control. If a line in your catalogue has ever generated a recall notice or a marketplace compliance request, treat it as in scope until you have checked. This is the limb that catches the seller who correctly concluded he had no 301 exposure and stopped there.

The boundary that keeps getting blurred is the channel. This is a rule for merchandise arriving through the international postal network. It is not the rule for express carriers and other non-postal channels, which run their own informal and formal entry processes. I have sat in the same meeting a dozen times where someone shifts a lane from post to express to avoid a requirement, or from express to post to chase a cheaper rate, without checking which entry regime the new channel puts him into. It depends on the port, and it depends on the channel. Check before the move, not after the first rejection.

Then the unglamorous part: where the data physically comes from. Four of the seven fields are created at pack time, in a warehouse, inside a system built to print labels rather than to file customs entries. Somebody has to decide what the sender value in that system means, who owns the SKU-to-classification table, and what happens when a new SKU appears that nobody has classified. I would give that table a named owner and a weekly review rather than leaving it to whoever notices a rejection. In my experience a classification table rots quietly for about four months and then starts costing money by the hour.

Timing, in the order the dates actually matter. The window is about 13 weeks, and the critical date inside it is not January 20 but December 7, the comment deadline on CBP's separate October 8 proposed rule on low-value imports. That proposal would extend Entry Type 13 data requirements to nearly all international-mail packages other than letters and flats, and would add new filing and bonding requirements. If it lands anywhere near its proposed shape, the population that needs this data goes from a subset to almost everything postal. Prepare for the January 20 scope and ignore the December 7 docket, and you will do this work twice, the second time without the runway.

Testing does not have to wait for any of that. CBP deployed the Entry Type 13 test to the ACE production environment on September 22, so the pipe is live and you can push real transmissions today. My own target would be classification finished by the end of November, sender and value data locked with whoever packs the parcel in the first week of December, live test transmissions through the second half of December, and a written fallback for the parcels that cannot make informal entry. Ninety days were offered because the agency expects systems to change, not intentions.

Two items belong on the same list and rarely make it onto it. The first is the bond. The October 8 proposal carries new bonding requirements, and CBP's own transition checklist tells importers to review customs bond requirements, which is not language an agency uses casually. If you are clearing today on a bond sized for duty-free postal volume, a rule that puts real duty and fees across 27,000 parcels changes the exposure your surety is underwriting. Call them in November rather than in January, when the rest of the channel is calling too. The second is the fallback: decide in writing now what happens to a parcel whose data you cannot complete, because the answer at the facility will be formal entry and nobody will negotiate it at the counter.

One operational reality the notice does not mention is how examination behaves in the postal channel. It does not behave like the express channel. There is no integrator to chase, the notice path runs through the postal service, and the clock is counted in days rather than hours. If the consignee refuses to pay duty, or the parcel is refused admission, the return leg is slow and largely outside your control. The way to manage that is not to manage it: file complete data so the parcel is never pulled. Every hour spent on classification is an hour not spent arguing with a facility about getting a parcel back.

My colleague Nadia Pryce, who reads this industry in hours rather than weeks, would frame the same 90 days differently: it is roughly 2,160 hours, and not one of them changes the delivery promise the buyer saw at checkout. That is worth holding on to. The operational risk here is not the filing, it is a parcel sitting at a mail facility for nine days waiting on a classification nobody assigned, while the customer's clock keeps running. The clock does not pause for a rulemaking.

Where this could read the wrong way round. If the December 7 comments push CBP to narrow the October 8 proposal, or to phase it, then January 20 is the whole job and the preparation above is proportionate. If instead the proposal lands broadly and quickly with the bonding changes intact, a different question overtakes this one: whether direct postal shipping to US consumers is the right channel at all for low-value goods. I would not call that yet, and I would not let anyone selling you software call it either. The condition that flips it is not the date. It is whether the entry cost per parcel ends up above the contribution margin on a USD 40 order.

So the practical list, in the order I would actually do it. Find your postal lanes first, because most importers cannot name theirs in ten minutes and that is the real starting point. Then take 100 parcels and test all seven fields against them, counting how many you could file today rather than how many you hope to file by January. Then decide per lane and per product family who files, you as owner, the purchaser, or a designated licensed broker, and get the designation in writing. Then run test transmissions. Then write the fallback for the ones that cannot go informal.

Keep the file while you are at it. If you have classified 200 SKUs, keep the reasoning, the date and the name of the person who signed off. A post-entry review eighteen months from now is not interested in your intentions. It is interested in whether you could show, on the day you filed, why that code was right. That is one line per SKU, written now, by whoever made the call. Reconstructed later by someone who was not there, it is worth far less, and the difference shows up exactly when you cannot afford it.

One more thing that belongs to the same 90 days and gets discovered too late: the commercial side. If part of your range moves from informal to formal because the data cannot be completed, the landed cost per unit changes, and so does whatever your marketplace listing promises about delivery. That is not a customs problem, it is a repricing problem, and repricing takes longer than classification. Start the two together, because finishing the codes in December and discovering in January that the basket no longer works is the expensive version of doing the work.

None of this is technically hard, and all of it is date-sensitive. This shipment can be saved, but the classification has to be decided and the filer designated before January 20, and the bond conversation has to happen this quarter. If you take one line from the notice, take the one about de minimis rather than the one about the extension. One of them is a schedule. The other one is the money.

↑ Back to the key points

— By Derek Xu