Are BMW and Xiaomi about to kill the car dealer?

Both are betting on the agency model - BMW Group converting its own network, Xiaomi Technology entering Europe with an agency model run through eight dealer groups it just signed. Everyone is embracing agency. Almost nobody has made it work. That gap is the whole story.

Let me make the case for agency first, because it should be the right future model.

Under the classic model, the maker sells to the dealer, the dealer sells to you. Under agency, the maker sells to you directly - the dealer becomes a paid agent, and the price, the data and the customer relationship belong to the manufacturer. One consistent price. Clean first-party data. No discount war. When the car is becoming a software-and-service product, owning that relationship is the whole game.

So why does it keep stalling? Volkswagen Group even pulled back. The reason given is IT problems, which is only a symptom.

A dealer network is not just a sales channel. It is a vast, distributed shock absorber. A thousand dealers quietly hold inventory, flex on price, run delivery logistics, handle the angry phone call and carry local working capital, all in parallel, all locally, without head office orchestrating.

Agency moves every bit of it to the centre. HQ must price in real time, in every market, hold and allocate national inventory, deliver to the customer's door, and answer directly. Work once spread across hundreds of independent businesses now lands on one central organisation, which was not built to serve millions of customers one at a time.

That is the real reason agency stalls. Not the software, but the organisation behind it.

The evidence is everywhere. Chinese brands entering Europe have tried every mix of direct and agency selling and none has truly cracked it at scale. And the one company that made direct-to-customer work, Tesla, did not convert a legacy organisation. It built the whole machine from scratch, as a tech company, with its own logistics, pricing, service, all vertically integrated.

Which is what makes Xiaomi interesting. It has zero experience in the European car-retail market. But it has deep experience in consumer electronics, it already runs an agency model at home, and it has real superiority in data and tech - exactly the muscle agency actually demands.

So here is the open question I cannot yet answer. Will that tech strength let Xiaomi make the agency model work where every traditional carmaker has failed? Nobody in the legacy industry has cracked it. Tesla only managed it by being a tech company first. With a clean-sheet EV brand and no dealer legacy to unwind, Xiaomi might just have the right starting point.

Agency is not the wrong model. It is a data-and-organisation challenge disguised as a retail one. Which is why the company that finally masters it may not be a carmaker at all.