Scam or opportunity? Uncovering the truth about digital human live streaming

In recent years, AI has been widely used in all walks of life, even in live broadcasting. In this article, the author briefly introduces the concept of digital human live broadcasting and analyzes its opportunities and future trends. It is recommended for people who are interested in the AI ​​industry.

In the past year, the concept of "digital human live broadcast" has been very popular. AI + digital human is the general trend, and it has been applied in many fields. Especially in the field of live broadcast/short video, digital human virtual anchors have greatly improved the efficiency of enterprises with low investment, high output and long-lasting live broadcast mode. According to statistics, in 2022 alone, the number of digital human-related companies registered reached 948, with a growth rate of 68%.

The results of digital human live broadcast are also very attractive:

Traditional live broadcasting requires hiring anchors, purchasing equipment, and setting up venues. The cost of a standard live broadcast room alone is as high as hundreds of thousands. However, digital human live broadcasting only requires a computer and at most one operator to start broadcasting, and the cost can be controlled at around 10,000.
In addition, imagine that with digital human technology, you can broadcast live on multiple accounts non-stop within 24 hours, maximizing the use of fragmented time to earn profits.
So, what is digital live streaming? Is it as useful as rumored? If so, how can businesses take advantage of it?

1. What is a digital human? Should enterprises join the game?

The so-called digital human, in simple terms, is a digital character image that is close to the human image and created using digital technology.
The creation of digital humans is an extremely complex task that requires three steps: modeling, animation, and rendering. It is achieved through advanced computer programs.
Ultimately, the digital human can restore more than 80% of the appearance of a real person, and whether it is appearance, movement or sound, it can be as natural as in the real world.
At present, digital human live broadcast has been widely used in many fields such as games and technology. Of course, the field where it is most used is the snack and beauty field.
It is reported that after using AI digital humans, the average daily transaction amount of Douyin's Fengweipai pasta live broadcast room was about 163,100 yuan, which is obviously better than the daily transaction amount of 151,500 yuan of real anchors.

In Pinduoduo's Xtep live broadcast room, the digital human anchor's average daily GMV was about 41,000 yuan, and it was on the hourly list of sports brands. Its strength should not be underestimated.

Faced with the surging wave of "digital human live streaming", a fan recently asked me whether companies should enter the digital human market?
I think it is time to join the game. Moreover, now is the best time for enterprises to join the digital human game.
However, this also depends on the industry and size of the enterprise. At this stage, digital human live broadcast may not be suitable for all enterprises.

1. Which companies are suitable to join?

• Game industry: The game industry is one of the earliest fields of application of digital human live broadcast, because the brand image and characteristics of virtual content of games perfectly match the characteristics of digital human live broadcast.
• Beauty and cosmetics industry: Beauty and cosmetics is a field that is highly associated with women. Digital human live broadcast can help brands build a better image through virtual images, beauty techniques, and other means.
• E-commerce industry: E-commerce platforms can use digital live broadcasts to promote products and showcase product features, thereby increasing exposure and sales.
• Food industry: The food industry has the characteristics of multiple categories and low average order value, which is very suitable for digital people to conduct live sales.
• Technology Industry: Digital human live broadcasts can help technology companies showcase their innovative achievements and enhance brand awareness and market share.

2. Which companies currently have limitations?

• Clothing industry: Different from other industries, clothing products sold online usually require the live broadcaster to wear them. At present, digital human technology is not mature enough, and it is not possible to change clothes flexibly. But this does not mean that it will not work in the future. In recent years, there are many companies that have invested in and developed digital human technology. I believe that in time, digital human technology will also be able to support the needs of the clothing industry.
• Knowledge payment: In addition, there are those talented anchors with strong personal characteristics, such as Dong Yuhui of Dongfangzhenxuan. On the surface, he is selling goods in the live broadcast room, but in fact, Dong Yuhui is selling a kind of "culture". For anchors like this who have both knowledge reserves and personal charm, I am afraid that digital people will not be able to completely replace them in the short term.

2. What pain points does digital human live streaming solve for enterprises?

Corporate investment is a rational business behavior that emphasizes the input-output ratio. If the investment cannot bring considerable benefits to the company, then such investment is obviously not worth it. So, what value can digital human live broadcast bring to the company?

1. Digital human live broadcast solves the problem of enterprises having difficulty recruiting anchors

Live streaming requires a series of talents related to live streaming, such as anchors, field control, and operations. However, many small and medium-sized enterprises are facing the following dilemmas: the company's attractiveness is limited, the salaries of talents are too high, and those who ask for low salaries lack experience and ability.

Digital human live broadcast only requires one operator to drive the program, which greatly solves the problem of difficulty in recruiting workers for enterprises.

2. Digital human live broadcast solves the problem of high cost for enterprises to keep anchors

A person in charge of a company that uses digital humans for live broadcasting mentioned: “A major advantage of digital humans is that they greatly reduce the labor costs of the company.
At present, the average monthly salary of a single anchor on the market is as high as 20,000 yuan, which means that even without counting the live broadcast team, the company's annual expenditure just to recruit one anchor is as high as 240,000 yuan.

The modeling fee for a digital anchor is about 10,000 yuan, plus various fees including system usage fees, it is only 100,000 yuan a year. If the company is large and has 20 anchors at the same time, using digital people to replace real anchors can save the company 2 million yuan in expenses a year.

3. Digital human live broadcast solves the problem of enterprises’ difficulty in retaining anchors

For companies, training anchors is a hard job. For a new anchor to enter the industry and become familiar with it, the company needs to spend a lot of time and money on him.

Even if the company finally creates a phenomenal IP, the uncertainty of live-action anchors is still too great for the company. They may become tired, sick, depressed, or even quit at any time.

Compared with real people, digital people's performance is very stable. They do not get sick and do not leave their jobs, which greatly reduces people's management costs.

4. Another major advantage of digital people is "off-peak traffic"

Currently, many businesses’ live broadcast rooms use a combination of real people and digital people. During the prime time in the evening, real people appear on the screen to promote products.
Late at night or during some periods of daytime when traffic is low, digital humans are used to fill the idle traffic, grab more users that real anchors cannot reach, and increase corporate profits.

3. How much does it cost to broadcast digital life live?

The essence of digital humans is a technology-driven track. From the initial cartoon-like paper figures with almost no conversational ability to today's stage with multiple portrait styles and the ability to communicate with users 24 hours a day, digital human technology has gone through multiple versions of development and iteration.

Today, there are multiple versions of digital human models with different prices and functions on the market for brands to choose from.

Recently, our team has visited a number of companies that develop digital human technology. According to the information we have learned, there are currently three levels of digital human technology used in the market. I have compiled their corresponding prices and effects for your reference.

Tier 1: The production of virtual digital people is mainly based on character image + Vocaloid voice synthesis, generating virtual digital people with beautiful appearance, such as Luo Tianyi;

• Reference price: No modeling fee, 1,000 yuan/month
• Effect: Recorded broadcast, stiff movements, single scenes, and the risk of being blocked by the platform.

Tier 2: The core technology of virtual digital humans is 2D/3D modeling + real-time motion capture + real-person dubbing, such as virtual digital singers Ling Qiu and Mu Xi, and beauty expert Liu Yexi;

• Reference price: Modeling fee is about 10,000 yuan, and then 5,000-10,000 yuan/month
• Effect: The movements are natural and the scenes are real, but it still requires a real person to drive it.

Level 3: Virtual digital humans can not only predict lip movements, but can also express themselves through text-synthesized speech. The entire virtual digital human’s face, body, voice, and movement are all generated by AI. It has the characteristics of real-life texture, high recognition, and strong interactivity. It is a "hyper-realistic virtual digital human" that is closer to human form.

• Reference price: Modeling fee is about 20,000 yuan, and then 10,000 yuan/month
• Effect: Natural movements, realistic scenes, highly close to the model itself, and highly interactive.

From the perspective of price, it is still relatively expensive for small and medium-sized enterprises to enter the field of "digital human live broadcast". The basic annual fee is more than 100,000 yuan, not counting the invisible operating costs and hardware requirements.

Therefore, a popular way of playing in the market is that enterprises find service providers to purchase digital human services in the early stage to reduce operating costs. After the test results show that digital human live broadcast is indeed suitable for them and achieves the ideal input-output ratio, they will consider building a full set of digital human supporting facilities.

However, we need to remind all companies that digital human live streaming is essentially just a tool. It can solve efficiency problems, but it cannot solve traffic problems well.

Therefore, don’t think that you can make a lot of money by buying a digital person to live broadcast and then give up, “letting the digital person live for 24 hours without anyone broadcasting.” This will not only fail to achieve the effect that the company wants, but will also have a certain risk of warning and account suspension.

Therefore, in the early stage of digital human application, enterprises still need to arrange 1-2 operators to monitor the traffic and drive the live broadcast in real people in the background, and gradually turn to unmanned live broadcast after the data is stable.

4. Final

Your understanding of new things today will determine your future competitiveness. The digitalization of the industry is an inevitable trend, and digital human live broadcast is an important part of the digitalization process.

Faced with the surging "wave of digital humans", if we cannot embrace change at a faster pace, we must at least learn and understand it.

Just like the previously popular GPT, the emergence of digital human technology will ultimately eliminate not humans themselves, but those who do not know how to (refuse to) use new technologies.

Author: Yan Tao Sanshou