One Million Robots

Du Chen

On December 7 of last year, Amazon UK successfully tested its first delivery drone, carrying a TV set top box and a bag of chips (well, crisps since they were British) to a consumer’s home, and people cheered the arrival of smart, automated shipping.

But for JD.com, that was already child’s play.

The summer before Amazon’s display of drone prowess, JD.com completed a test of its VT1 delivery drone between the cities of Suqian and Xi’an. These are not like the other vertical takeoff drones you usually see. The more resemble Harrier jets, lifting off vertically to then switch to winged flight. The VT1 can carry up to 200kg of cargo, far more than Amazon’s drone capacity of five pounds (or about 2.3kg).

Recently, JD.com’s CTO Zhang Chen spoke in Silicon Valley, where he revealed that JD.com has already gotten clearance for drone delivery in six different provinces in China.

JD.com has also already deployed driverless delivery vehicles on four Chinese college campuses. They all have different model numbers, but they are usually about as tall as a human, have multiple compartments, are guided by computer vision to read their surroundings, and navigate between buildings according to where they are likely to meet the least traffic. Users can enter a code to retrieve their packages, making the vehicles into something like mobile mailboxes.

The development on these drones and driverless vehicles were started not long after JD.com itself, and are just one example of how JD.com is pulling ahead of Amazon and Alibaba when it comes to technology. Zhang says that machine learning is driving every part of JD.com’s operations.



One of JD.com's drone carriers

Since accepting investment from Tencent and being brought into the fold of its ecosystem, JD.com has been able to draw on the data of 930 million WeChat users to refine its machine learning technology and optimize its supply chain, logistics management, sales services, and fintech.

In its marketing, Zhang explains that JD.com is able to tailor strategies for affiliated brands according to data they take in. For high profile, high conversion products (like iPhones), JD.com recommends presales. Low profile, high conversion goods like smart hardware: gin up interest with product launch events. High profile, low conversion products are a better fit for discounts. And low profile, low conversion goods just need to be cleared as quickly as possible.

Consumer behavior can influence everything from production to sales, as Zhang explains in the example of a humble mouse: “We discovered that a lot of people were searching for a ‘mute mouse’, but there was no such product on the market, so we handed that data over to Logitech, and they produced one. Every week, its sales numbers increased by the thousands, leading to a huge increase in revenue for that year, all just because they had data that other brands didn’t.”

Zhang points out that global retail is undergoing a massive transformation. American retail is depressed, while Chinese retail has been shaken up by the wave of “new retail.” These changes have been precipitated by movements in technology, altering how consumers engage with what they buy. In the past, someone might meander through a Best Buy, spot a TV they wanted, maybe find it a little overpriced, wander around looking at others, come back, ask a sales clerk if there weren’t any discounts available … Getting a consumer’s wants and a brand’s offerings to match up exactly took effort and a little luck. Now, online, JD.com can help brands find precisely who is willing to buy what in which price range, and offer the appropriate discounts.

Zhang feels that, even as retail changes, the things that will never lose their place are price, efficiency, and buying experience, and successful retailers have to attend to all three.

Zhang says: “A few days ago, reps from Google paid us a visit and asked if we had something like Amazon’s Prime service. I said, ‘92% of our orders are filled either the same or next day, what would we want with a Prime service?’” But as JD.com grows, permeating China’s suburban areas, countryside, and remoter regions, how can it maintain that level of speed and efficiency?

Not with humans. This is why JD.com is promoting the use of driverless vehicles. Founder and CEO Liu Qiangdong’s plan is that within 10 years all deliveries should be completed by an autonomous machine. Zhang anticipates that will require a million such vehicles, in the air or on the ground, far more than its current crew of delivery people.

Autonomous delivery vehicles in the cities are still held back by the complexity of travel on the streets and the density of buildings. But that’s not going to stop the march towards unmanned delivery. At the very least, China’s western region still has a huge market awaiting JD.com’s arrival and development.

Jack Ma of Alibaba once argued that JD.com could only be slowed down by the demands of managing its huge staff of delivery people. But JD.com has already decided, it can do without the humans, and race ahead with robots instead.