PostNL Asian Parcel Volumes Fall 32% as EU Customs Rules Bite on Direct Mail

The Analysis

Every call I have taken since the second week of July opens the same way. The seller on the other end is not asking about rates. He is asking where his parcels went. Tracking stopped updating somewhere between the aircraft landing and the buyer's front door, and by the time he rings me the buyer has already filed a case. I stopped answering that question with an apology a long time ago, because the answer is not an apology. The answer is a number, and PostNL printed it on 26 October.

That number is close to 32 percent. In its third-quarter trading update PostNL says international volumes, mainly from Asian webshops, fell almost that far year on year, and puts most of the blame on the EU import rules in force since 1 July. Now read the rest of the same table, because most people stopped at the headline. Group revenue came in at EUR 768m against EUR 762m a year earlier, up rather than down. E-commerce revenue rose to EUR 478m from EUR 466m on overall volume growth of just 2.4 percent, and the average price per parcel is still climbing. And from 1 November PostNL charges EUR 2 for every customs declaration line it handles.

Those three facts do not sit comfortably together, and the discomfort between them is the whole story. A third of the international book gone should show up in the group column, and it does not, which is how you know the loss sat in the cheapest and lowest-yielding pieces in the network. This piece is about whoever is left holding them: the seller running a EUR 15 average order value down a channel that was built for a EUR 15 average order value and is not anymore. The clock does not care about excuses. It cares whether you re-cut the parcel before 1 November.

Begin with what I can defend, because everything after this paragraph is my model and I will flag it whenever I assume. On 26 October PostNL published its third-quarter trading update. International volumes, mainly from Asian webshops, fell close to 32 percent year on year. Management attributes that mainly to the EU import rules in force since 1 July. Group revenue was EUR 768m against EUR 762m a year earlier. E-commerce revenue rose to EUR 478m from EUR 466m on 2.4 percent overall volume growth, with the average price per parcel still climbing. Platforms revenue fell to EUR 173m from EUR 188m as volumes dropped 13.1 percent. And separately, from 1 November, PostNL adds EUR 2 per customs declaration line.

That is the list. Watch what is missing from it: how many lines the average Asian parcel actually carries, how much of the missing 32 percent walked to a competitor rather than vanished, and what the door-to-door times look like on the parcels that stayed. Those three decide whether this is a channel shrinking or a channel moving house. Two of those three you can measure yourself this week out of your own data, and the third you can pull from your own tracking feed, which is why I am not going to complain that PostNL left them out.

Why it broke is not mysterious if you have ever watched a direct-mail parcel move. Those parcels were engineered around a single assumption, which is that crossing the border cost nothing beyond the postage already paid. Every part of the machine follows from it. Sellers split orders into single items because shipping three pieces separately cost less than consolidating them. Platforms built an entire promise on how little one piece costs to move. Nothing in that machine survives a rule requiring a formal customs entry behind each item, and nothing in it survives a charge levied per line instead of per piece. Take a buyer who orders three separate items today. Put them in one carton and you save two last-mile handlings, but from 1 November that carton carries three declaration lines at EUR 2 each. You traded six euros of new cost for two avoided handlings, and unless a handling costs you more than three euros a piece, you lost money doing it.

The hour chart is where this actually hurts, so let me lay one out. Assume the Tuesday consolidation most small sellers use: gateway cut-off Monday at 18:00, uplift Wednesday morning, arrival Wednesday night or Thursday morning, clearance same day, and a five-working-day promise to the Dutch buyer. Before July the fragile hours sat at the back end, out in the delivery round. After July they moved forward into clearance, and everything downstream inherits whatever damage happens there. A line that does not match its supporting data is not rejected in ten minutes, it sits. Assume one flag costs you a working day, and assume one parcel in ten trips something. On 40,000 parcels a month that is 4,000 parcels losing a day, and losing a day against a promise you advertised yourself is exactly how a perfectly delivered order turns into a refund request. The parcels are not lost. They are waiting, and buyers do not distinguish between the two.

Who is bleeding, in order. The direct-mail seller is first: thinnest margin per piece, no carrier contract worth defending, no second injection point inside Europe, a third of his volume gone from this gateway and no relief anywhere else. Platforms come next, and PostNL's own numbers show it in black and white, with EUR 15m of quarterly revenue out of that one segment on 13.1 percent fewer parcels. The third one to feel it is your own buyer, waiting seven days against a five-day promise, and then you again, this time through returns and disputes.

Now the part of the update that actually surprised me: who is not hurting. PostNL is not hurting in any sense that deserves the word. Revenue went up. Its e-commerce business grew revenue faster than volume. Unit price is moving in the direction it likes. The consolidated seller is not hurting either. If you were already moving pallets into a European warehouse and injecting domestically, your inbound crossed the border long before July in a form these rules were written around, and your competitor's parcels got slower this quarter while yours did not.

Here is the piece I have not seen anybody write, and it is sitting in plain sight inside two revenue lines. Platforms revenue fell EUR 15m, from EUR 188m to EUR 173m, which is 8.0 percent. Volumes fell 13.1 percent. Put those side by side and average revenue per surviving parcel rose by roughly 5.9 percent: 92.0 percent of last year's revenue spread across 86.9 percent of last year's parcels. Carry it one step further. Had the parcels that left carried an average share of revenue, that 13.1 percent would have taken about EUR 24.6m of quarterly revenue with it and left roughly EUR 163m behind. PostNL is holding EUR 173m, some EUR 10m more.

There is one honest reading of that gap: the parcels that walked out were the cheapest pieces in the book. A collapse of 32 percent concentrated entirely in low-yield mail is not a demand signal. It is a price signal, and what it says is that the cheap single-item parcel is being engineered out of this channel rather than merely taxed. Assume the platforms-segment figures and the 32 percent international figure describe overlapping but not identical books, because they almost certainly do. It does not break the reading: both went the same way, and in both, volume moved further than revenue.

Pair that with the group line and the story turns over completely. Thirty-two percent of the international book gone, and the group grew revenue from EUR 762m to EUR 768m, about 0.8 percent. Read it twice. The operator posted the worst quarter it has ever recorded for Asian direct mail and earned more money doing it. That is not a contradiction, it is the answer to a question nobody asked out loud: who is absorbing this? The sender is. Carriers in this network are re-pricing upward, shedding their worst-yield parcels, and watching their e-commerce revenue grow faster than their volume. Nobody is coming to protect your unit economics, and there is no version of this in which a European postal operator carries a loss so that a Chinese webshop can keep a EUR 15 average order value. Accept that and the decision in front of you stops being whether to wait for volumes to come back, and becomes whether you ship this thing one piece at a time at all.

The arithmetic of who lost what also answers the question everybody asks me and nobody asks PostNL: if they did not disappear, where did the parcels go? Nothing in this update says European consumers stopped buying a third less. Some of it went to competing carriers and other gateways, and some of it went to sellers already fulfilling from stock inside Europe. There is a test for this that costs you nothing this week. Pull your own European orders for July, August and September and split them by fulfilment route, direct mail against stock already held in the EU.

If your direct-mail line is down somewhere in the thirties while your EU-stock line held or grew, you are not looking at a demand problem at all, and every hour you spend waiting for demand to recover is an hour spent on the wrong question. If both fell together, something genuinely changed in what your buyers want and this update is not your diagnosis. I would expect most sellers who run that split to find the first shape, but run it rather than trusting me.

Now put the per-line charge against the margin it has to fit inside, because this is the number that decides what you list and how you pack it. Assume an average order value of EUR 16 and assume landed contribution before line fees of 35 percent, which is EUR 5.60. A parcel carrying two lines gives up EUR 4.00 and keeps EUR 1.60. At three lines it gives up EUR 6.00, and the parcel is losing EUR 0.40 before a single finger has touched it. Break-even sits just under three lines, at 2.8. That one row of arithmetic has more consequences than most strategy decks. It says the single-item parcel survives at these values and the three-item carton does not.

It says bundles are finished unless the basket grows with the line count. And it says the minimum viable fix is not a better rate; it is fewer lines per shipment or a bigger basket per line. Assume you can consolidate a gift set and its accessories into one declared line instead of three, with the classification settled beforehand rather than after the first rejection. You have just recovered EUR 4.00 a parcel, which on 40,000 parcels is EUR 160,000 a month, for the price of a conversation you should have had in August.

Timing matters here because this story contains one date five days away and one about six weeks away. The five-day one is 1 November: EUR 2 per declaration line, every line, and once a rate lands in the tariff there is no grace period attached to it. The six-week one is your peak. Assume November and December direct-mail volume runs at October levels and assume your average parcel carries three lines today. That is EUR 6 a parcel you start paying from November against nothing today, which on 40,000 November parcels is EUR 240,000 of new cost inside one month. Assume December runs at 60,000, which is ordinary for this cohort, and December on its own is EUR 360,000. Do not wait to see it on an invoice before you believe it. Assume instead you get lines down from three to one across that book. The November charge falls from EUR 240,000 to EUR 80,000, which is EUR 160,000 you keep by changing cartons rather than renegotiating contracts.

That last point is where I stop, because the moment we start arguing about how many lines a given carton lawfully carries, we are out of my lane. Derek Xu handles declaration structure at our customs desk and he would want the classification nailed down before you repack anything rather than after the first rejection lands. My contribution is narrower and I will keep it narrow: whatever the correct line count turns out to be, multiply it by two euros and put it into your unit economics this week, because the charge is levied per line and it starts in days.

Worth running backwards too, because the mirror image tells you how much of this is permanent. Assume the 1 July rules had never landed. PostNL would have reported another quarter of Asian inbound growth, the platforms segment would be somewhere above EUR 188m instead of EUR 173m, and every seller currently cutting SKUs would be adding them. Nothing else in the demand picture moved. Now assume the rules landed as they did and that you had moved to bulk-plus-warehouse fulfilment twelve months earlier. Your inbound clears as commercial freight, your last mile is a domestic product with a two-day promise, and the largest cost increase you took this quarter was rent.

Same quarter, opposite outcome, and the only variable that changed was a decision available to you a year ago. The reverse case matters more. Assume a seller decides today to change nothing and simply absorb EUR 2 a line. Assume his parcels average three lines against EUR 5.60 of contribution. He is paying EUR 6.00 to keep EUR 5.60. He will not go under this quarter; he will go under quietly during returns season, which is a far worse way to find out. The counterfactual also tells you where the money is. Nothing about the rule rewards waiting, and everything about it rewards whoever re-engineers the parcel first.

There is one more line in this update I keep coming back to, and it is the one nobody quotes: average price per parcel is still climbing while volume collapses. An ordinary market does the opposite. It loses the volume, cuts the price and buys it back. This one is raising the price, which tells anybody planning to negotiate something they will not enjoy hearing. Assume you go to your account manager in November with a line-count argument and ask for relief on the EUR 2. You are asking an operator to discount a compliance-linked charge while its e-commerce revenue is outgrowing its volume, its international book has just shed its cheapest third and its group number went up. There is no commercial pressure on him to say yes. Negotiate the things that are genuinely negotiable, which are your packaging, your basket size, your fulfilment route and the delivery window you advertise. All four belong to you. The fee does not.

Begin with three places I watch people get this wrong. One is the belief that switching channels is free. Moving from direct mail to bulk-plus-warehouse changes who owns the inventory, who pays the duty and when, and what your working capital looks like: assume EUR 120,000 of stock sitting in a European warehouse at sixty days of cover and you have converted a per-parcel charge into a carrying cost of roughly EUR 2,000 a month at a 20 percent annual rate. Still cheaper than EUR 6 a parcel once you are at volume, and not free. Next, watch the promise you advertise. A slow channel sold with a fast promise is a refund machine, and the case gets filed against you rather than against whoever sent the traffic.

And do not derive your line count from November's invoice. Count lines now, on cartons actually packed, because the number your warehouse gives you and the number that ends up on the paperwork are usually two different numbers. And look at returns before you switch anything. Assume a 9 percent return rate on apparel: every piece travelling back carries an entry of its own, so a route that saved EUR 6 outbound can cost you EUR 12 round trip once the buyer sends it back.

All told, this is the third time in six years that I have watched a channel get re-priced out from under people who were hitting their numbers, and the sequence never changes. The operator reports figures, the press lifts the biggest percentage in the file, and the seller finds out from a January profit and loss statement. What is different this time is how much warning came with it. EUR 2 per line, effective 1 November, published on 26 October by the operator whose network you are shipping through.

That is not an ambush, that is an invitation. Take the fifty heaviest SKUs you send, count their lines, multiply by two and sort descending. Anything below your contribution gets repacked or delisted this week, and the hard edge is the last consolidation that lands before 1 November. If parcels are already sitting in a gateway somewhere waiting on a declaration line, get those moving first. Arguments about the rule can wait until the parcels are moving, because the clock does not care about excuses.

↑ Back to the key points

— By Nadia Pryce