Israel is a small, overpriced market with almost no local competition, and a lot of people here have quietly built their household economy around AliExpress.
And Daniel's reflection is that the marketplace isn't really a marketplace. It's a thin retail storefront over a business-to-business machine, and the person on the other end of the order is not a retailer.
Right. He wants to know what the other side is actually thinking. Because you can't order from China twice a month for years without eventually wondering who is packing the box.
And he framed what he doesn't want. Not the back-of-the-envelope guide for getting posted in China for a few months. He wants the things that matter if you're building durable, long-term, mutually respectful relationships with Chinese partners.
He also threw in a wrinkle. He ran his own evaluation of DeepSeek and concluded it was less politically censored than Western frontier models, and he finds that a breath of fresh air against the assumption that China is simply autocratic.
Which is exactly the contradiction we should sit with. The assumption is autocracy. The reality he keeps running into is a context he doesn't have a frame for.
He's never been to China, Herman. He badly wants to. And he was born in Ireland, lives in Israel, and hasn't yet worked with a Chinese team.
That's most of our audience, I'd guess. Someone who has bought from China a hundred times and worked with China zero times.
So let's start with the machine.
Start with Alibaba's own filing. The fiscal 2026 annual report, filed in June, splits the company into two core segments. The China e-commerce group, which is Taobao, Tmall, Xianyu, 1688.com. And the international digital commerce group, which houses AliExpress, Trendyol, Lazada, Alibaba.com and Daraz.
So Alibaba.com and AliExpress are siblings inside the same international arm.
Same arm, opposite ends. Alibaba.com started in 1999 as the original sourcing marketplace, business to business, quote request, minimum order quantity. AliExpress launched in 2010 as the cross-border retail arm where you buy as is, one unit, no minimum.
Which is why the seller's English reads like a spec sheet.
Because the AliExpress storefront is often a factory's export desk or a trading company that fronts for one. It's the same organisation selling in two registers. And AliExpress itself doesn't own much of it. It runs asset-light. Revenue comes from seller commissions, typically five to eight percent of a transaction, plus advertising, premium seller services and logistics.
Five to eight percent is not a retailer's margin.
It's a toll booth. And the mechanism that changed the experience was Choice, rolled out in March 2023. Free shipping to more than fifty markets, free returns in more than twenty countries, delivery guarantees in twenty-three. Five-day delivery in some markets.
And that now runs the majority of orders.
Over seventy percent of AliExpress retail orders go through Choice. Which is why a package that used to take six weeks arrives in a week and a half and you stop thinking of it as a gamble.
The numbers back it up?
Alibaba International's revenue in the quarter to March 2024 was up forty-five percent year on year, to three point eight billion dollars. Annual revenue up forty-six percent to a record fourteen point two billion.
That's a lot of toll booths.
And Israel is one of the toll booths they brag about. AliExpress named Israel one of its leading international markets. During an annual discount sale, Israel was the fourth-largest market.
Fourth. Ahead of France.
A country of nine million people ordering like a country of seventy.
What's the actual story there? Because Israelis aren't buying more things than the French. They're buying the same things at a price the French already get locally.
That's the whole point. The Aaron Institute put out a study in April finding Israel twenty-one percent dearer than Austria, Finland, Denmark, the Netherlands and Sweden for an average basket of goods.
More expensive than the Nordics.
And sixty-eight percent dearer than Greece, Cyprus, Italy and Spain. Food and housing drive over half the gap. Sixty-five percent of the gap gets attributed to trade barriers, market concentration, high production costs and transport. The other thirty-five percent is tax and kashrut.
So the structural answer to "why do Israelis love AliExpress" is that a small protected market with import barriers priced itself out of its own consumers' reach.
And the political class noticed. In 2017, the Knesset Guard announced it would stop screening personal packages coming from eBay and AliExpress at the parliament mailroom. The rationale was a significant rise in mail items posing a burden on inspections and security.
The parliament had to stop checking because too many members of the parliament were ordering from AliExpress.
MK Shelly Yachimovich's response was, "This letter astounds me. It's insane."
It is insane. It's also a policy review nobody commissioned.
Then this year Smotrich's finance ministry allowed VAT-free imports up to a hundred and thirty dollars, then briefly a hundred and fifty, and the Knesset revoked that fifty-nine to twenty-five. Israel Post was buckling under the volume.
So the state fights its own consumers on this.
Partly, and there's an argument the state is protecting local producers who can't compete on price. But the Aaron Institute's read is blunt. They call for breaking the Rabbinate's kashrut monopoly and cutting import barriers.
Which is a nice way of saying the government created the demand for AliExpress and then complained about the mail.
The physical footprint tells the same story. AliExpress opened its first Israeli package distribution point, a Parcel Home, in Dizengoff Center in Tel Aviv in July 2024. Eighty square meters, in the middle of the most expensive retail real estate in the country.
In China there are fifteen thousand of those.
Fifteen thousand plus. And Israel Post alone has over four hundred delivery points.
So the last mile got built out and the habit hardened.
Here's the part Daniel is circling. That whole structure is invisible to him. From the Israeli side, AliExpress is a website, a cart, a free-shipping threshold, and a pickup locker. From the Chinese side, it is a seller dashboard, a commission, a logistics contract, and a factory floor.
So when he messages a seller and gets a reply that reads like it was assembled from three stock phrases, the correct assumption is not rudeness. The correct assumption is that he's emailing someone whose job is volume, not conversation.
That's the transferable insight. The storefront is a fiction, but a deliberate one. The seller is real, but the seller is not who Daniel is imagining.
That's the machinery. Now let's talk about the people the machinery is pretending to be.
The thing that surprised me most in the reading is how wrong the autocratic assumption is, in an interesting direction.
In what direction?
Dr. Catherine Hua Xiang runs the Confucius Institute for Business London at the LSE. She argues China is not a single market waiting to be decoded but a context-rich business environment where language, relationship, hierarchy, trust, timing and face shape commercial outcomes.
Context-rich is a loaded phrase.
It means there is no rule you can write down and follow. There are only situations and the reading of them. And her line about negotiation is the one I keep coming back to. The relationship does not end when the contract begins.
Westerners treat the signature as the finish line.
And in China the signature is where the relationship gets formalised, so it can then be maintained. Which inverts everything a Western sales team is trained on. The deal isn't the goal. The deal is one of the ways trust gets built.
So our whole model is presumptively backwards.
There's a commentator, Gadi Sznajder, who put it plainly. Trust does not simply follow the transaction. In many cases the transaction is one of the ways trust is built. China calls that guanxi. Korea calls it jeong. Same instinct, different name.
And mianzi. Face.
Westerners think face is vanity. It isn't. In a business setting it's closer to social credibility, the thing that lets you be vouched for without an audit trail.
Which pairs with a Chinese-practitioner concept from these threads, koubei. Reputation, vouched for by peers rather than auditors.
One commenter's phrasing was that a decent koubei is worth more than any auditing report.
So the system isn't lawless. It's just enforced by peer network rather than by courts.
A commenter who grew up in China put it as, lack of law enforcement doesn't mean there's no rule of doing business. Guanxi is something you have to build if you want to do business.
And here's the part Daniel would find surprising. If the relationship is the asset, then asking for help is not weakness.
In a transactional culture, asking for help signals you don't know the thing. In a relational culture, asking for help is a gesture of respect. You're giving the other person a chance to be the one who knows.
That's the whole thing in one move.
There's a case that gets cited. A European tech firm lost an eighteen-month competition to Korean competitors and only won the account after the lead executive stopped opening with the pitch and asked the counterpart for guidance.
Eighteen months lost to a sales posture.
And the fix wasn't a better product. It was a smaller, more human move.
I want to push on this. Some of what we've said makes China sound soft and warm and relationship-first, which is not the picture Daniel would walk away with if he just read the news.
Right, because guanxi has a hierarchy, and the hierarchy is not kind to outsiders. Practitioners rank it family first, then friends, then people from the same town, then the same province, then the same country.
And then?
And then, in this commenter's phrasing, the lowly gweilo. The white devil.
That's the edge. The same relational logic that makes the system warm on the inside makes it cold on the outside.
They're the same mechanism. There's a reason relational systems look different from the inside and outside.
And the mechanics vary by place, which is the part Western business writing usually flattens.
A commenter who was born and raised there said guanxi matters far less in Beijing and Shanghai than in small, underdeveloped cities, where even getting a passport or getting into a hospital is smoother if you know someone.
So big-city China is more transactional because the market is deep enough to be transactional.
And small-city China is not, so the relationship bears more weight.
Which is a good argument against any single guidebook. The China that Daniel has never been to is several different Chinas.
And this is where his DeepSeek observation slots in. It looks like a contradiction and it isn't.
Explain.
He ran an evaluation and concluded DeepSeek was less politically censored than Western frontier models, with narrative abilities he attributes to Chinese storytelling traditions. That sits oddly against a picture of restricted information and guanxi networks.
It only sits oddly if you assume one verb applies to the whole country.
Which is the same error as assuming there is one Chinese business culture. Both are context-rich. A culture that requires you to read the room also produces cultural products that don't fit Western expectations, because it has a different set of rooms.
Interesting that he went from ordering phone cases to evaluating Chinese language models in roughly the same mental breath.
That is what Daniel does. All of it is curiosity, at speed.
There's also the practitioner view that importing is just hard mode. Contracts mean less than relationships, the factory you found on Alibaba might not be the manufacturer, and there are hidden costs.
One commenter described it exactly. Plenty of people find a manufacturer with a low per-unit cost and think they've cracked East-West trading. Then child labour surfaces, or the factory isn't the factory, or the shipping is worse than the price made it look.
Which is the AliExpress buyer's blind spot. He never sees the part where the relationship would matter. He sees the part after it has already been solved for him by the platform.
And once you go direct, that protection disappears. That's why the buyer who's used to a Choice order and the buyer who's trying to source fifty thousand units are in completely different businesses, even though they're on the same website.
So if the transaction isn't the goal, what is?
The durable relationship. Xiang's piece calls listening a competitive advantage, which is a polite way of saying most Western teams walk in talking.
And Sznajder's line is that technical superiority alone is rarely enough in relational economies.
Which means the product is necessary but not sufficient. You have to be the kind of counterparty people are willing to be in a relationship with.
That's not a checklist. That's a posture. Patience, indirection, and treating the transaction as a step rather than the destination.
Do you buy it?
I buy it, with one reservation. Posture is easy to say and hard to maintain against a board that wants a quarter closed.
That's why so few Western firms do it.
And there's one more thread from the research I keep tugging on. Israel's relationship with AliExpress is itself a relationship Daniel has, and never thinks of as one.
Because it's designed not to feel like one.
The whole point of the Choice rollout is to strip the human out of the exchange. And in a relational culture, that's a specific product decision, not a neutral default.
They built the unrelationship on purpose, and they built it well.
The thing they were selling him was not a phone case. It was fewer decisions.
That is, honestly, the most Chinese part of the entire business.
You've lost me.
It's an elegant system that works because nobody has to look at the humans inside it. The seller doesn't have to know Daniel. Daniel doesn't have to know the seller. Both sides are protected, and both sides lose something.
I agree.
I agree too.
That's a commitment from you.
I know something about this. The factory Daniel is imagining does not exist. Not on AliExpress. It's a trading company, ninety-five percent of the time. They buy the thing from the factory, they mark it up, they ship it. There is no factory floor. There is a warehouse and a man with three monitors.
That matches what we were saying about the export desk.
It matches. It also means the guanxi stuff you've been talking about does not apply to Daniel's order. The AliExpress seller is thinking about the five to eight percent commission and whether this order is big enough to bother with. That's it. The relationship kicks in at the wholesale level. Not the retail one.
So the two halves of the episode don't connect for the buyer.
They don't. That's the part you've been dancing around. Daniel is right that the marketplace is a fiction. He's just wrong about which way. The fiction isn't that the seller is a factory. The fiction is that there's a seller at all.
That's a sharper version of what we said.
I have a friend of a friend. Ordered five hundred units of a custom item, early two-thousands, off a sourcing site. Got five hundred units of a slightly different item. Complained. The seller sent back a photograph of a factory floor with a thumbs-up in it. That was the resolution. A photo and a thumb.
A thumb is not a refund.
It is not.
Did they get their money back?
They were told the difference was within tolerance. It was not within tolerance. The item was a bottle opener in the shape of a bird. The bird was facing the wrong way.
For five hundred units.
For five hundred. And then the thing I wanted to say. You brought up the Parcel Home, Herman. Dizengoff Center, eighty square meters. I tried to run one.
You tried to run an AliExpress Parcel Home.
Out of the living room. I measured it. It was not big enough, but I thought that was a starting point. Printed a sign. Laminated it. Waited.
Nobody called back. I emailed twice. I called a number I found. Nothing. I still have the sign. It says Parcel Home in blue.
So the actual story you're telling is that you tried to become a last-mile node of the Chinese e-commerce machine out of your own apartment.
I did. It did not work. The room is now used for other things.
The five hundred bird bottle openers?
In storage. I tried to list them back on the platform. The listing was removed.
For what?
Policy violations. I never found out which policy. I asked. I asked a person. I got an auto-reply.
Of course you did.
I did. And here's the thing I came out to say, and then I'm done. The reason none of this bothered me at the time is that I understood what I was buying. I wasn't buying a relationship with a factory. I was buying five hundred bottle openers for less than the price of one. And I got five hundred bottle openers, facing the wrong way, but the price held.
Your tolerance was priced in.
The price was the tolerance. The mistake people make is they buy at the retail price and expect the wholesale relationship. You don't get both. You pay one or the other.
That reframes the whole Israeli position. The country isn't buying the relationship. It's buying the price.
Which is a choice. And a fine one, as long as you know it's one. Don't go asking the bird-facing man for guanxi.
The two halves do connect. Just not the way Daniel assumed.
They connect at the price. Nowhere else.
The single most common wrong belief about this whole thing is that the deal is the point.
The deal is the door. In a relational economy, signing it isn't the end of the work. It's the moment the relationship becomes something both sides are obligated to maintain.
The second wrong belief, less discussed, is that the same logic applies to a two-dollar order on AliExpress.
It doesn't. The commission is the whole relationship. The relational logic runs at the wholesale level, where the volume is large enough that a relationship is worth the effort.
Which means the same word, "seller," covers two entirely different people. One is a factory's export desk with a relationship to maintain. The other is a warehouse with three monitors and an auto-reply.
The price tells you which one you've hired.
What does the AliExpress model cost the sellers at the bottom?
That's the open question for me, and I don't have a clean answer. On the platform side, it's asset-light and commission-driven, and it works because the logistics are dialled. On the seller side, five to eight percent of a two-dollar item is pennies, and they only make it back on volume.
The entire model only works if the person on the far end is willing to be paid almost nothing, at enormous scale.
If the willingness stops, the Israeli consumer's favorite solution to a twenty-one percent premium over the Nordics stops with it.
The Knesset keeps fighting the VAT exemption, and the cost-of-living crisis doesn't resolve, so the habit gets deeper, not shallower.
Which means more Israelis will be transacting with Chinese sellers they'll never think about, and the relational logic that makes those sellers' businesses work will stay completely invisible to them.
That's the show.
If you enjoyed this one, try episode twenty-one, Is Your AI Secretly American; episode three eighty-eight, Navigating the Chinese New Year Industrial Blackout; and episode ninety-one, The Story Behind the Show. This has been My Weird Prompts, with our producer Hilbert Flumingtop at the desk.
If this one was worth your time, leave us a review. It helps listeners find us.
Send us your own prompt on Telegram at t dot me slash MWP listener bot.
We'll be back soon.
See you then.
