The emergence of Internet e-commerce has brought great convenience to people’s lives. With the trend and development of media, more e-commerce related content forms, such as live broadcast and short video, have emerged one after another, bringing higher technical requirements and challenges.

The following content is compiled from LivevideoStack’s interview with Wu Guilin (Liang Shu), head of Alibaba’s Taobao Technology Content Center.

Q: Hi, Liang Shu. It’s my great honor to invite you for an interview. Talk to LivevideoStack readers about yourself and your team, right?

Liang Shu: Hello, I’m Liang Shu. At present, I am in charge of all the technical teams related to content business of Tao department, including the group’s content middle platform and upper core content business, such as Taobao live broadcast, Taobao new product “browse”, and the technical team of independent innovation APP “DianTao”. At the same time, there are a lot of horizontal content middle platform products, such as the underlying network transmission, codec, and the underlying AI big team, etc.

Our team basically took shape in 2019. Taobao in the graphic era to the mobile era between the transformation of a large number of upgrades. With the advance of multimedia technology, we hope Taobao can truly complete the transformation from the traditional mobile era to the multimedia era. Lately, I’ve heard a lot of talk about “mediating,” which is something my team has always wanted to do. Perhaps many people will have a question: “the industry has done a good job in short video APP, Taobao why do this thing, what is the purpose?”

In my opinion, video or live broadcast is just a media tool. If you want to build a product around the tool, each family has its own way of doing it. For e-commerce, the typical change brought by live broadcasting is the upgrading of shopping guide. In the era of picture and text, simple description of goods is given, while in the era of video multimedia, we can better describe goods with the help of video or live broadcast to help users “plant grass”.

Many years ago, the dream of Taobao was “shopping on Taobao”, and the future goal direction was “living on Taobao”. And “media” can help us achieve the goal of “living on Taobao” in a real sense. Taobao not only allows shopping, but also covers clothing, food, housing and transportation. We hope that all the things that can be seen, beautiful and helpful to life can be realized and solved on Taobao. This is what we want to do to be different from other products, which is the direction that our team has been working on, and also the direction that Taobao is upgrading to media.

Q: From the mobile era to the multimedia era, “living on Taobao” is a very good vision. When Douyin and Kuaishou are making short videos, we can see that their main means of cash is through e-commerce. Taobao is from e-commerce to mobile terminal, and then to short video live broadcast. Then, how to view the relationship between short video and live broadcast and e-commerce? Is it traffic, social or something else?

Liang Shu: In my opinion, should not be the flow. In 2016, everyone was competing for the show live broadcast, which was very popular at that time. Taobao Live firmly chose to take the lead in e-commerce shopping live broadcast centering on our advantages. Maybe now people think that e-commerce live broadcast is particularly popular, but in 2016-2018, everyone’s focus is still on other types of live broadcast. It was not until “Double 11” in 2019 that e-commerce live broadcast and Taobao live broadcast were really paid attention to by everyone.

And we can see that it’s periodic. Taobao spent a lot of time to do the underlying infrastructure, to do new products. The same is true for short videos. The production of short videos in the era of graphic and text is very difficult and the cost is relatively high. With the development of The Times, technology is getting better and better, production costs are getting lower and lower, and the content created is getting better and better. At this point, we hope to help consumers in more new ways to make shopping decisions. Live streaming is a great way to grow grass, but there are many more.

In the past, we can understand the understanding of Taobao as that we only open Taobao when we want to buy something. However, in the future, we hope that in all aspects of clothing, food, housing and living, no matter whether the public has specific shopping needs, they can go to Taobao first, generate interest through the content, and then make purchase.

Taobao’s vision of content is to “make life better,” which can be understood as being able to recommend something helpful for your future life when you don’t know what to buy.

Taobao’s core strength is that it has a high value user base that is truly willing to shop. What we need most is to find the core users of Taobao, and actually move forward around the user group step by step. The most important point to do short video is from supply to consumption. We need to have firm confidence, and do the best of all content production and supply around the user group. Only by slowly accumulating in the consumption side can we achieve a balance, so as to really help consumers.

An important part of this is to be patient and from the user’s point of view. Multimedia is the direction of the future, Taobao “grass” and “lifestyle”, are in the development of media, but for the future may produce more changes, is not clear, but can be sure that the future of multimedia and video focus is expression, and the core of Taobao is e-commerce.

Q: At present, we can see that Taobao’s products in e-commerce and short video include “Fumao” and “Diantao”. Could you give us a brief introduction of the overall ecological structure of Taobao and the changes of product forms in the past period of time?

Liang Shu: In terms of the overall product form, there must be a centralized content center at the bottom. The content center is responsible for the content at the bottom, auditing, etc.

At the same time, in front of the production side, the first is short video. We have recently launched the “pro-shooting” product, which is oriented to merchants, and can be used as the production and supply method of the whole short video. The second one is Taobao, which is not a marketing-oriented platform for merchants, but more focused on “planting grass” or “lifestyle” content publishing products. In terms of the whole front desk product side, our core of short video is “browsing” and “DianTao APP”. “DianTao APP” is a new way to upgrade content e-commerce, which is a multi-media oriented product. In terms of product form, it is mainly reflected in Taobao Live in DianTao APP and Taobao APP.

This is at present treasure the integral product form that arrives from supply, production to consume. There are many technologies involved in this, such as content shooting. We have the SDK of “Tao Pai”, which can realize the shooting of various gameplay, and all business lines can use it. At the codec level, as a common code base, S265 can provide compression and decoding functions for all internal codec requirements involved in the business;

In terms of live broadcasting, we have the GRTN network transmission at the bottom, including the middle station at the bottom, the basic SDK at the bottom and all the technologies involved in the production and consumption at the top.

Q: Faced with the fierce competition in the current industry market, do you have some anxiety? What is the biggest challenge you are facing now?

Liang Shu: I think the biggest challenges are mainly in two aspects. For the outside, the competition is relatively fierce, of course, this will cause a certain amount of anxiety. So we really need to think from the user’s point of view.

The first is what has been mentioned from the year before last until now: how to achieve data understanding and user-driven growth of front end products. How can a consumer-oriented product understand the user? The user has the perceptual and the rational division, the perceptual thing is often difficult to measure. But rational things can be learned and understood. Therefore, data is not the only criterion, we will find problems through data, through technology and data driving business to iterate quickly, really understand the user, let the user feel that you are slowly getting better. As long as I have enough patience, I believe I will find my own way for the product eventually.

The second aspect is the introduction of talents. The real content, or to understand the content, understand users, understand technology. We also brought in a lot of technical talent last year, but it’s still not enough. Whether in codec, or in short video shooting, we have a great demand for talents who understand content, consumption algorithms and so on.

Of course, there may be a lot of anxiety in this process, such as pressure from outside. But we want it to really go back to the basics, where every video, every live stream, finds its place in a different context. A lot of things change over time, whether it’s video clarity, codec, tool play, content delivery, all sorts of technical aspects, and over a period of time, the effect is completely different. In fact, more concerned about the partners may be more anxious, impatient. If we have a little patience, more understanding of users, do not need to see how much other people’s users, on the hand of the existing part of the real user group to meet the good, in fact, there is nothing to worry about.

Q: We can see that most of the products like “Fumao” and “Diantao” are for domestic users. What services, products and technical capabilities are available overseas?

Liang Shu: For example, overseas businesses are mainly Lazada, AE and ICBU. In the overseas one of the most important goals is the distribution of points, Tao Department and Ali cloud together in the global distribution of points, mainly follow the business to carry out, where the business is, where the point is.

So far, we still do very simple and pure technical support for the whole overseas market. For live broadcast, we will support it mainly through some technologies developed by ourselves, in the form of platform and partly through SDK. We will export some core technologies and let each business line build by itself, including coding technology and underlying communication technology. For short videos, we will build the content center of the whole group, including the shooting, uploading and processing of short videos. Business parties only need to connect and use the existing platform without worrying about transcoding, video processing, video CDN and other issues.

In the future, we hope to export some practices of “wandering around” or Taobao live broadcast in China not only as the underlying technology, but also as the core experience. The better video understanding techniques, short video gameplay, or video content construction techniques for output.

Q: What do you think are the biggest challenges overseas today?

Liang Shu: There are two challenges. The first is the complexity of the global network environment, which requires us to continue to strengthen cloud computing abroad, especially the layout of network infrastructure. The second point is how to fully deploy the whole set of our proven architectural patterns overseas to form a global and integrated architectural system.

# Technology layout, facing the future of multi-media.

Q: Basic technologies such as codecs, network transmission, content production and consumption, as well as new technologies such as XR immersion experience in the long run, what specific technical layout will Taobao have and what upgrades may be brought to the video experience of Taobao in the next step?

Liang Shu: To sum up, we will make layout in three directions at present:

First, coding. From 264, 265 to 266, Ali, including Taobao, will continue to invest a lot in new coding methods. Whether it is video or live streaming, picture quality and clarity are very important, including 360 panorama, among which coding is a very important one to save costs and improve picture quality.

Second, just as mentioned 3D, XR and other cloud rendering. In the past, the algorithm was divided into terminal and cloud. In the future, we will combine the terminal and cloud processing for the algorithm at the multimedia level. Like 3D algorithms on the end of the cost may be relatively high, we will do it in the cloud; If some algorithms are difficult to implement on the end, I will also put them into the cloud for processing. In this way, the delay of processing will be very high. How to realize the cloud-to-end transmission quickly is a problem, which needs the support of some technologies at the edge computing and RTC level. In fact, this is similar to the delay of live broadcasting, similar to the audio and video conferencing technology. This is also an aspect of our recent layout. In the future, the combination of end-cloud is a problem we must solve.

Third, electricity live there are a large number of video content, before the understanding of multimedia content is often done through the label content classification and recognition, and recognition level at electricity goods, with the help of a large number of electrical business content, we hope to achieve more real-time, effective content understanding and recognition, later as long as see the short video content, would identify what goods in real time, That’s what we’ve been investing in for the last few years.

We hope to solve similar problems through a variety of technologies, to reduce the cost through technology, to make the experience better. This is an ideal, and we are working hard to further realize it.

Q: Multimedia is a trend of e-commerce in the future. Can you assess the current progress of multi-media of Tao Department and what stage it is in?

Liang Shu: I think it’s still in the stage of gradual infiltration. We have done a good job in live streaming, but we have not done enough in short video, which is actually what we have been doing since last year. In the aspect of live broadcasting, our technology has been relatively mature, such as low delay technology, codec and so on. Recently we have also been doing some new experiments, such as 3D live streaming and XR live streaming.

When merchants release new products, they have a high demand for live broadcast. Generally, they will set up a good site offline, which is actually a waste of cost. So from last year, we started to make 3D scenes, hoping to make the technology similar to film production accessible and stable to the people. Material rendering can be carried out in the cloud by means of green screen and software production, so as to achieve a live broadcast effect closer to the real scene. We hope to combine with Ali cloud, through technical means to achieve rendering in the cloud, after rendering through a series of processing, to provide a simple effect preview, preview confirmation can be directly broadcast. For merchants, the release of products in the past need to spend a lot of costs offline, but now only need a very low cost online can be completed, and the scene is flexible, can choose the corresponding scene according to the different actual products.

In fact, this is a very good new technology, but the difficulty lies in the fact that the technology requires very high performance of the machine. How to transfer the technology of film production and realize it more simply? For example, the most recently mentioned is cloud rendering, which can provide powerful machines and computing power in the cloud. The benefit is that it can be recycled and the cost is relatively low. This is the first technology.

The second technique is to solve the problem of latency. How to let the broadcast people see their scene in real time, including the identification of characters, interaction and so on. The future is likely to be immersive viewing, allowing consumers to interact in real time on a live stream.

Using technology to change the world.

Q: A lot of people in technology go down the road of development, probably because of a variety of accidental factors, I wonder what made you choose the path of technology?

Liang Shu: In the 2000s, when QQ and Facebook came out, I also hoped that people like Ma Huateng and Mark Zuckerberg could change the world by writing code by themselves. Tencent has successfully connected people closer through two communication software, while Facebook has connected people through social communication.

When I was in college, I liked to toss and turn, especially some products of that era, which touched me a lot. Sometimes I get a little defeated, and some thoughts come up, like why I can’t make these things. Young and ignorant, if you will. But also at this stage I began to build their own websites, their own books to download a series of things, and then slowly on the road to the Internet. I probably prefer to study the preferences of consumers, and I hope that the things I make can really change the lives of consumers. So sometimes I feel very happy, in this era, people who know technology have an advantage, and can make more things different through programs, algorithms and many other technologies. When people communicate with you about technology, such as face recognition, you know the principle behind it. Sometimes you wonder if you can try to change something. This is what I like to do.

Q: Have you thought about any goals for this year? Work or life.

Liang Shu: The most important thing in my work is that I hope our new product “Shou Shou” can reach the expected target this year. Be patient and find your real target audience. There is also “DianTao” APP, which can truly “discover good content, good goods and good prices” at the level of content e-commerce.

LivevideoStack: OK, thank you Liang Shu.