Transcript of Keynote Speech Given by G7 Founder and CEO Xuehun Zhai
at GeekPark Innovation Festival 2019
BEIJING, China, January 21, 2019 – GeekPark IF is a 3-day festival covering all things digital in China. Now in its 9th year, this industry event is cross between SXSW and TechCrunch Disrupt, focusing solely on the promotion of entrepreneurship and innovation in China. This year, G7’s Founder and CEO, Xuehun Zhai, was invited to provide one of the keynote addresses. Bringing his 20+ years of logistics and technology experience, Zhai stressed the importance of maintaining a productivity focus and highlighted how G7 has brought a 21st Century Industrial Revolution to China’s logistics industry.
Translation of Zhai’s remarks on January 18, 2019, Opening Day of the GeekPark Innovation Festival:
JACK ZHANG: (Founder and President of GeekPark): Next we have an entrepreneur, Xuehun Zhai. He has been tirelessly promoting efficiency in the logistics industry. I like to call him a technology entrepreneur who spans four major eras: the Information Era, the Internet Era, the Mobile Era, and now the Internet of Things Era. He has consistently pushed the logistics industry forward. I have known him for 20 years and he really is an old warrior.
I always thought of G7 as just a company that puts lots of sensors vehicles so I can know the transport status and driver’s behavior. This would improve efficiency by providing real-time feedback. But this year, he again has surprised me with his new ideas and latest thinking.
What is it? Let’s welcome him to hear his insight.
XUEHUN ZHAI: When I met Jack, we were only 30 years old….
I particularly like this topic that Jack gave me: The Evolution of Productivity Under Industrial Intelligence. When it comes to evolution, there must be past, present, and future. So, let me walk you through it.
Let’s start with how I used to provide technical services for the logistics industry.
You may have not felt the evolution of the logistics industry over the past 20 years, but just think about it. Just now Jack mentioned that you could send a parcel from Beijing to Shanghai for 6 or 7 yuan. Over the past 20 years, logistics has not increased in price. On the contrary, it has decreased --- from 20 yuan to 6-7 yuan.
All things considered, logistics is built on the backs of many, many people. I remember when my daughter was born 16 years ago. At that time, we had a full-time helper and we paid her 300 yuan. Today’s helpers are paid at least 5000 yuan. Our basic labor force cost has increased at least 15x.
Here’s another one. If you drove 20 years ago, gasoline at that time would cost you 1-2 yuan per liter. Today we’re at 7 yuan per liter, an increase of at least 3x.
Resources and labor have risen several times over. Logistics, on the other hand, hasn’t just not increased, it has dropped. What this means is that this industry has been operating incredibly efficiently, in extreme, over the past 20 years.
Let me show you a picture that basically explains what I have just explained. In the past 20 years, whereas labor and energy prices have risen tremendously, how is it that logistics continues to provide us with very cost-effective services?
This is the G7 highway [a highway in China that was formerly the Silk Road]. I took this shot when I drove it with my team last year. I don’t know if you’ve ever taken the G7. This picture is at the northernmost end of the road. If you keep going another 200 km, you’ll have left China. I guess most of you don’t have any particular impression of Mt. Wula, but the highways there are stunning.
Because cars are fast and fuel efficient, their costs are much lower. For those of you around my age, you’ll remember this truck. 20 years ago in Beijing, you could find this: Beijing 120. This truck didn’t have a full trailer and could hold 10 square meters. Now, look at trucks today. They can hold over 100 square meters. Moreover, as your packages travel in this truck, they are safe from damage. So, we have better trucks. We have better roads. Overall quality has improved while the cost of our transportation has been fundamentally reduced. All for 6 or 7 yuan.
During this period, G7 and I did something interesting. Assuming good road quality, if you want to build a road network with sufficient national coverage, you will need 5000 truck routes. Every day, you need 15,000 trucks running those routes. How do you dispatch those trucks? How do you know the trucks are arriving on time? 8 years ago, G7 built a tool for this.
Have you been to the airport and seen those large flight tracker signboards? 20 years ago, I had the opportunity to help Beijing Airport build this flight tracker. I brought this same concept to the freight industry and used the simplest GPS to link a truck to its truck lane. 5,000 truck lanes or 15,000 trucks. It doesn’t matter. Each truck automatically knows whether it is on time or not.
With this, a company with a national network like Yunda can dispatch its national network in a timely fashion. This small management tool gives everyone the capability for everyone to manage their transportation. It is by this small tool that G7 has the opportunity to serve most of China's best logistics companies, such as JD.com, SF Express.
Just now Jack mentioned the main theme of our industry. What makes me think that this is the main theme?
My WeChat has over than 4,000 friends. More than 3,000 are owners of logistics companies. Seven or eight have listed companies valued over 10 billion. All successful logistics company owners in the past 10 years have done one thing without exception: established a punctual fleet. Today, you give them goods. Tomorrow, they deliver it. They are just doing this one little thing. Maybe the time or the method is different, but their product is the same.
So, in the past, “industry intelligence” development meant improving roads and vehicles. Now, we have added a new dimension: The Punctuality KPI. Just by including the means to calculate and permutate The Punctuality KPI, the industry has spawned million… billion dollar companies.
I’ve just covered the past. Now, let’s cover the Internet of Things era. Recently, we have seen the emergence of AIOT (Artificial Intelligence Internet of Things). Take a closer look at how G7 is working this in today.
This screen shows the main, real-time, real-data at the moment on the G7 platform. Last year, Zhou Hang asked me how many cars are on my platform. I said that there are 700,000. Today, there are 900,000. Zhou Hang commented that our progress is slow: “it will take a while before you to get to 7 million….”
So far today, G7-equipped trucks on this illustration have run over 80 million kilometers. Each blip you see means something has happened at that point. It may be a good thing. For example, a vehicle has departed a warehouse. It may be a bad thing. For example, this cargo has a high-temperature alert. Your frozen lamb is now at 7.3 degrees. It could be a low-temperature alert. Your milk is now frozen. This alert shows that the driver has closed his eyes for more than two seconds. The concept is that it takes a mere 2 seconds for a truck to drive off the road.
So everyone can see that 132 of the 900,000 vehicles on G7 have collided or rolled over today. A total of 2.95 million liters of fuel have been consumed.
All of this is using the Internet of Things to see the living "breathing" China's logistics industry. That is to say, there are many dangers right now. There are many problems with our goods. There are departures. There are arrivals.
I can be proud to tell you that I have been all over the world, spoken to people on the internet, logistics, and the Internet of Things industries. Such a comprehensive system, that covers such a large country, and uses the Internet of Things - different kinds of data, connected together, to solve these problems - in having this, G7 stands alone. Also, 900,000 vehicles all tracked on a single third-party independent platform. Is G7’s progress really so slow? Each vehicle still needs a device. Each customer still needs service.
What I just showed you is data generated just now. But we are not only trying to generate data, we also want to change things. For example, the high temperature alert you just saw.
Recently, we took a small step, to collaborate with one of the leading refrigeration unit manufacturers in the world. By connecting the platform with the refrigeration unit, we can control the temperature without assistance from the driver or the truck. Too high or too low, or no matter how many trucks, what temperature is needed, I can control the temperature here [shows phone]. If it is fruit and I want to see it, I can look at it, in real-time --- the fruit in the cargo hold. This is the first step we have made. For the first time in this industry, we can not only see but also control.
We have also done some AI. Last year we made a safety robot that mimicked the safety management procedures of the human team.
This curve shows the performance of the first month after our robot security officer went online. In its first month, the safety system reduced unsafe risky behavior time by 91%.
I am particularly proud to say that in 2018, the G7 safety system served 20,000 trucks, that’s 20,000 drivers, 24x7. Until today, no one on our system has been killed because of the poor driving. Maybe you find this unremarkable, but did you know that for the 20 million trucks in our country, there are on average 50,000 accidents and 20,000 deaths per year?
We used AI last year to analyze drivers' expressions, driving styles, weather, and other conditions to determine if the driver was in a dangerous situation. If so, we actively intervened and called him, first via automated call. If that wasn’t enough, then with a staff member. We call this: human + AI.
In the end, we saw drivers sleepy, playing mobile phones, watching movies. I’m serious, a driver would drive a 60-ton truck... and watch movies... on the highway. So, after deploying human + AI, we built a team of customer service agents. But we found that the employees needed to have frequent breaks, and rotations. As you can imagine, the monitoring job was quite harrowing. Frankly speaking, with regards to introducing autonomous driving and regulations, I am the most supportive. You people have no idea how unreliable human drivers are.
In addition to collecting data and creating an AI, how could we go one step further? We wanted to change the product itself. How could we improve it?
I am very pleased that we had another good pilot test last year. Take a look at this.
Last year, we tried to make a 48’ long IoT device, that’s 17.5 meters. Why would we want to do that? Because if you ever go to a loading dock, you will find many people holding papers, recording how many tons of goods are in each load. People need to do this because production objects can’t speak for themselves. Again, you need human managers because objects aren’t able to manage themselves.
So, our team said, “let’s combine the manager and object in one.” This is our attempt. The results aren’t bad. This 48’ trailer was just launched in the latter half of last year and already it accounts for a double-digit share of the entire trailer market, more market share than any traditional trailer company has seen in the past 18 years.
The logistics industry is accepting this device. It hopes to transform objects into devices that can have intelligence, operate, manage themselves, and calculate their own accounts. This is a huge improvement in efficiency.
I would like to give you one last example. Take a look at these photos. Those are our technicians on the scene. See that shirtless guy? He is the product manager of this beautiful trailer system.
Look at his buddy below, stacking those boxes. That one is the algorithm expert of our company's best-embedded software. He is handling the truck algorithm. The guy above him? He’s our company's best platform algorithm expert. We say we want to help the industry. Being innovative doesn’t mean just sitting in a bright, fancy office. To engage in new things, we need to solve problems on the ground.
It used to be a truck and a trailer. The truck has no intelligence, algorithms, and data. So it needs a human to drive it. But humans have a lot of bugs. They get sleepy. They like to play. So we need more humans to monitor those humans. Do you know how many humans it takes to look after drivers like these? I just said China has 20 million trucks. There are 5 million people in China to manage these 20 million trucks.
Of course, even these managers aren’t completely failsafe, because sometimes they have faulty memories. Sometimes they remember incidents incorrectly. So customers turn to G7 to build a management tool. Although our speed is quite good - it has improved a lot - the sheer number and variety of managers, operators, and devices actually hinder our efficiency.
Everyone says that the next era will be autonomous driving. I think that for logistics, it’s not just autonomous driving. It is turning this truck, AI, and driver into an automated network-based platform, integrating management tools, people and devices. This will improve safety at least 20-fold. In addition, the efficiency can be increased by at least 10-fold.
That means those 5 million people, managing 20 million cars now, will drop to 500,000 in the future. The rest can do something more interesting. Like, spend more time listening more to Jack, discuss things of the future.