China’s first audio industry

Text | Meng Yue

Edit | Fang Ting

At 10:30 pm on January 17, 2020, Beijing time, the UGC audio community “Litchi” was listed on the NASDAQ stock exchange under the stock trading symbol “LIZI” and the issue price was US depositary shares per share ( ADS) $ 11.

Litchi was founded in 2013 and focuses on “recording and sharing life with sound”, which has accumulated a large number of users and content creators. As of November 30, 2019, the platform has accumulated more than 170 million original audio content; During October-November 2019, the average monthly active users of Litchi exceeded 51 million, and about 5.9 million monthly active content creators , is the largest UGC audio community in China.

Although it is “the first in the audio industry”, the benchmark of litchi is not Himalayan or Dragonfly FM. Compared to the mainstream advertising + knowledge payment + membership model, Litchi’s founder Lai Yilong is more willing to mention comparisons or learning objects is fast.

The financing shortage in 2016, Lai Yilong had to start looking for a clear business model for litchi. The competition for copyright requires ample money and ammunition, and the payment of knowledge is extremely dependent on the big V in the head. Neither is the strength of Litchi. Therefore, litchi never shaken the path of UGC, and eventually gradually explored a set of “voice live broadcast, paid reward” business model.

In the first three quarters of 2019, lychee revenue was 815 million yuan, a year-on-year growth rate of about 46%. Audio entertainment revenue (that is, virtual gift rewards) accounted for 99.1%, which was the largest support point for lychee revenue.

But the business model is too singular and profitability has not yet been achieved, leaving the market still controversial about litchi. However, Lai Yilong said that relying on live broadcasts to reward lychees in 2020 can achieve full profitability, and future live broadcasts will not be the only business model of litchi.

“The quicker do (we) do it.” Lai Yilong said.

The way to broadcast live broadcasts has been opened. Next, Litchi is also considering developing e-commerce business. The latter has set up electric poles such as Simba and Sanda. The selling logic behind it is that the user’s loyalty to the anchor individual can be directly transformed into the trust in the goods they sell. Lai Yilong believes that this point can also be established on audio anchors. In addition, Lichi’s users are mainly sinking channels, and the user structure is similar, so that e-commerce live streaming may have the potential to become the next growth point of Lichi.

  • Why does litchi stick to UGC? Is it forced or helpless?

  • The audio industry is small. How can litchi break through the ceiling?

  • Is the litchi business model too simple, and when can I turn a profit?

  • Who are litchi’s competitors at this stage? What are the characteristics of its users?

    With these questions, I conducted an exclusive interview with Litchi founder Lai Yilong at the listing site of Nasdaq. The following is the actual record of the interview. The content is edited.

    Interview | The audio community

    “Life and death at stake” 2016

    : From the establishment of Litchi in 2013 to its successful listing today, how many important strategic adjustments did it undergo?

    Lai Yilong: The live broadcast is the most important adjustment. It will be tested at the end of 2016 and officially launched in 2017.

    : How did you figure this out?

    Lai Yilong: In 2016, the company’s books were enough for one year. I was very anxious to raise funds, but there was no business model and no one gave you money. The whole audio was not clear at that time. business model. In addition, there was a short winter in 2016. So I went around and ran for a long time without good results, so I was more anxious. Because Jingwei had voted for us every round before, I finally went to Zhang Ying and asked him if he could give me a little more support this winter. He said that self-improvement is the most powerful thing, and you should do your own thing before you say it.

    : It means you do well first, don’t ask me for money first.

    Lai Yilong: What he means is that it’s no use going to anyone in winter. If you do it yourself, people will be interested even in the winter season. Don’t run financing all day, but spend more time on business. After I had talked with Zhang Ying, I also went to Da Gang (Feng Da Gang, current president). Da Gang was about to go at that time, but had not announced yet. I had an appointment with Da Gang in a coffee shop and asked him very stupidly: “What does Zhang Ying mean by this?” He answered me: “Zhang YingThere is always hope in chatting with you, and it will not save you. With his character, he won’t bother to talk to you anymore, so you don’t need to worry, you can work with peace of mind. “

    I understand that if it ’s really difficult in the end, it wo n’t be completely unsupported. I’m not good at financing. I might as well come back and think about the business model.

    In fact, when I created litchi, I went to the United States to investigate. The sound is nothing more than three forms. The first is music, that is, Spotify; the second is audio books, Amazon ’s Audible had billions of dollars in revenue in 2013, but it has a prerequisite that Amazon has mastered a large number of published book rights in the United States, and the copyright owner, Amazon platform owner, and producer have cooperated very well. However, in China, the copyright in the publishing industry is very fragmented. I roughly estimate that about half of them are web articles. The copyright is in the hands of companies like reading. It is difficult to make audio books without copyright.

    The last one is Podcast. At that time, there were already 1 billion users in the world, but there was no business model. The revenue for a year depends on third-party advertising, which adds up to tens of millions of dollars.

    After inspection, I firmly stated that I want to do podcasts, but I must not advertise. I have to charge users. This is what I wanted to understand from the beginning, but I have not figured out how to charge.

    We tried a lot of methods in 2016, and even walked through idol mode. I also went to Japan to study AKB48 to see how their theater did it. Can we also make our anchor into an idol. Back to learn AKB48, hold handshake meetings, release photos, release records, and do radio dramas, but found that moving hard is not enough. We have hundreds of thousands of anchors. It is impossible to go offline. CD photos are too heavy. It is difficult to scale.

    The actual gift is not easy to make. I naturally thought of the virtual gift, but the team was very skeptical, because there was no model of purely rewarding the sound. At that time, knowledge payment was considered the best audio. Monetization model. At that time, I looked for celebrities for courses, but finding celebrities completely dilutes the features of our platform. You are pushing celebrities, and grassroots people put them with them, and certainly no one buys them, so I have not considered paying for knowledge.

    We first tried recording virtual broadcasts for virtual gifts, and for the first time we received 30,000 yuan a day. According to my experience on the radio, if you add interaction, then 300,000 a day may be able to do it, and I have the bottom of my heart. At that time, everyone felt that it was because they didn’t know the value of sound interaction. When I was on the radio for more than nine years, the audience sent text messages in order to interact with the anchor, one dollar each, and received many messages every day.It is one of the important sources of income for radio stations. So all the right decisions are actually based on the bottom line.

    : So in the end, I chose to go live with audio. What is the difference from live video?

    Lai Yilong: Compared with live video, the biggest feature of live audio is its strong interactivity. In fact, the key to the success of Litchi in products is the change in two small points: One is that the interaction between our live broadcast is full-screen. Many previous products were modeled on video and a background The photo shows the dialogue in the lower left corner; the second is that the audience can talk to the anchor on Mai, we are the first product with this feature on the mobile terminal.

    Also, sound is not the same as video, one is mainly hormonal or entertainment, and the other is mainly emotional. This has also been verified. Emotion is the most popular content on the litchi platform, and female users also account for a higher proportion.

    : In addition to live sound, are there other important nodes?

    Lai Yilong: If it is important after 2016, there is another, We have really done a comprehensive AI intelligent distribution. This thing is actually The challenge is very big, because the distribution is not accurate at the beginning, you will face the problem of user loss. After we persisted for a year, slowly the user growth rate came back. In fact, everyone knows that AI is the future, but it is still difficult for you to be firm or say that you have the confidence to fully use it. Coupled with the AI distribution of sound, there is no one in the world who can learn from us , no one has done this.

    : Ca n’t learn the distribution algorithm for videos?

    Lai Yilong: Video and sound are completely different. Smart The key to distribution is three points. The first is understanding of the content. , including keywords, etc. A video has become very mature. Some open source technologies such as Facebook and Google have been disclosed. The difficulty of sound is that the content contains a lot of emotions. The same sentence has a different tone, and the emotions expressed are completely opposite. Therefore, we need to analyze people’s emotions, speed of speech, and voice characteristics.

    The second is understanding of people , which is the analysis of user portraits.Home is almost the same; The third is understanding of human behavior , which requires constant feedback from users. Why is the distribution of short videos good? Because the feedback from the user is fast enough, please like it if it looks good, and pass it if it doesn’t look good. But the sound user may be immersed in it for a long time and will not give any feedback, so the algorithm will be slow to train.

    : How long did it take for this algorithm to reach a more mature state?

    Lai Yilong: I don’t dare to say mature at this time, but it started to take effect in a year.

    : How to judge the positive effect?

    Lai Yilong: The click-through rate is increasing, and the retention is better.

    “Finally, litchi is a cross-terminal sound service platform, live broadcasting is just a process”

    : Litchi said in the prospectus that he is the UGC community, and it is the instagram version of instagram. Does Litchi position himself on the audio or community track?

    Lai Yilong: We are still audio, because in the end we need to cross-terminal, we are a voice service provider, now benchmarking instagram for easy understanding, because litchi is also cut in from tools, the earliest is to provide users with recordings Beautification, noise reduction and other functions.

    : Do you think there will be another product with the same ins in China? In fact, many products are still benchmarking like this before.

    Lai Yilong: Actually, if Facebook comes out, it doesn’t make any sense for you to become another Facebook. New ways of socializing and interacting will emerge, and seizing new opportunities is even more important.

    3 6 氪: The outside world will think that litchi looks more like a live voice product. Is litchi in your current state of satisfaction?

    Lai Yilong: I think positioning Litchi as a voice broadcast company may be a misunderstanding of the outside world. I told a colleague about a vision. I said that someone who was driving on the road one day was definitely not the traditional FM radio station. It should be a networked, interactive, and very versatile voice platform. That’s how our slogan “Redefined Radio” was. So to this day, we have actually done only 30% of this goal.This form is also a process.

    Have you seen the movie “her”? This is a movie that inspired me a lot. The arrival of 5G will definitely bring very big changes. In the future, as described in this movie, there will be scenes of sound interaction in wireless headphones, in cars, and speakers at home. Then it needs to be powerful behind it. Content supply. So Litchi has been accumulating, and it is a large group of people behind it, not some fixed content. As I said above, Litchi is a cross-terminal voice service platform.

    : So the core resource of litchi is the anchor?

    Lai Yilong: Yes, the core is the anchor. We have about 5 million anchors active in the recording every month, of which hundreds of thousands are live broadcast anchors. Why do we insist on UGC mode? It’s because we want to interact. Now you may not see the power of these voice anchors, but if you are covered by instant voice interaction from home to car to personal scenes, you will definitely feel it.

    : So it sounds that the future of litchi is closely related to changes in the terminal?

    Lai Yilong: Indeed it is. Now everyone will think that the audio market is relatively small. In the United States, if the ceiling of video is $ 1 trillion, the audio may be less than $ 100 billion. But once various terminals are interactively covered, sound has the incomparable advantage of video, and we have been waiting for this day.

    “Force a hole, make a hole”

    : You just mentioned that litchi has only achieved 30% of the goal so far. In addition to the external factors mentioned above, what else does litchi need to do internally?

    Lai Yilong: The first thing is to serve the anchor, and let more people participate in the content contribution of the voice. Another is innovation. In the past, I called on the radio. In our case, it was Lian Mai. Before, I wrote or texted. Now I type in the public screen. The nature of the sound will not change, but the method must be continuously innovated. This is What we have been doing.

    For example, we made a sound card in the “Litchi Lab” to identify whether your tone is “Yuejieyin” or “Lolitayin”, etc. Later we copied a lot of products to do, but Behind this is that we have accumulated sound analysis capabilities for several years, which cannot be surpassed in a short time. On this basis, we also launched “Sound Dating” because the effect is betterLater, it was independently made into a “squeak” product.

    We are also exploring a lot of sound innovations on smart speakers with Baidu. Although we do n’t make terminals ourselves, interactive sound innovation on terminals will be one of our main directions in the future.

    : Why not touch the terminal? Himalayan made smart speakers.

    Lai Yilong: Not our strength, nor our mission. We ultimately want to provide content across terminals, and it doesn’t make sense to produce our own terminals.

    : “Squeak” seems to have just returned to the shelves after being removed at the end of July. What is your plan for this product?

    Lai Yilong: At that time, it was actually an industry adjustment. During this time, we have been improving the content supervision level. For example, we can use voiceprints to identify which types of voiceprints are dangerous and which types Is safe. At that time, the data rose rapidly, which was also unexpected, indicating that the user of billing for voice dating was paid, and we will definitely continue to do this. (Due to version reasons, at present, some Apple users may not be able to search for “squeak” in the App Store)

    : Is the multi-product matrix also based on the consideration of diversification of business models? Because more than 95% of the current revenue comes from live users.

    Lai Yilong: I agree with Huawei Ren Zhengfei’s sentence about the business model. “Strengthen a hole, make a hole” is to put your strength in the same place, and the benefits come from the same place. I think “concentration” is very important at this stage, and it is precisely because we are focused enough that we can achieve a revenue scale of more than 1 billion. I can say that live broadcasting is the only voice that is growing and exceeds 1 billion. Business model.

    But this does not mean that the future live broadcast is the only business model of Litchi. All business models in China are nothing more than live broadcast, games, advertising, e-commerce I think advertising is difficult to scale up to sound, because sound ads are not clickable. You wo n’t click when you listen, so it ca n’t be quantified. But the game of sound is a good way to try. Love and producer are “sound + develop”. We are also doing some incubation. E-commerce including sound is also worth exploring..

    : What about e-commerce for sound?

    Lai Yilong: Do whatever you want.

    : May need some visual stimulation to sell goods? Anchor expressions, interaction between live broadcasts, and more.

    Lai Yilong: Actually the earliest radio stations will also sell goods. I think at least it can be explored. In fact, the fast hand is also the realization of “fan culture”. Trusting the anchor and paying for the anchor will only pay for it. Litchi still has this atmosphere. We have always been emotional content is more popular, the deeper emotional binding between the anchor and the user, the atmosphere is more tender.

    : Is there any time when you can expect to be profitable?

    Lai Yilong: We can achieve profitability in 2020. In fact, we have basically evened out in 2018, only a strategic loss of several million. In 2019, a lot of money was invested in the sharing of anchors, AI and overseas strategies, which led to losses. Our losses are not because of business losses, but because of new investments. With the expansion of our income scale, I think we can achieve full profits in 2020.

    : Or will it mainly rely on live broadcasting?

    Lai Yilong: Yes, the live broadcast will grow rapidly, and our profit model will not change much in the short term.

    : The share of the anchor has been rising, will it put pressure on our earnings?

    Lai Yilong: At this stage we still want to support a large number of anchors below the waist. We increase subsidies for anchors, so that more small anchors can make money, and they are willing to broadcast live in Litchi for a long time and become a professional full-time anchor. Only with more high-quality anchors can user growth. If we stop subsidizing, the financial statements will immediately look good, but it is not the right approach at this stage. At a certain stage, we will definitely consider reducing or stopping the subsidy.

    : What income level is considered a professional anchor level?

    Lai Yilong: A monthly income of 5,000 yuan is considered to be a line. Many of our anchors are in third-tier or lower-tier cities, so 5,000 yuan is basically a normal salary level.

    : If the entire platform only looks at one indicator, which one do you think is the most important?

    Lai Yilong: is a sticky user, our user is L37, which is 3 times in 7 days. L37 users now account for 50% to 60% of our daily lives.

    Neither Himalayan nor fast-hand tremolo is a competitor

    : Who do you think are litchi competitors at this stage?

    Lai Yilong: First of all It’s definitely not Himalayan, but we haven’t figured out who the competitors are. Himalayan is very different from us in all aspects, whether at the business level or the user level. I think the bigger challenge for us now may come from the music platform , such as NetEase Cloud Music, also has anchor stations and a good sense of community. In the future, especially after the popularity of smart speakers and other terminals, the giants will definitely exert more power.

    But in my opinion, the entry of giants is also good, it can make the audio penetration rate higher. If the giants come in to be healthy competition, it is good for this industry. You can see Douyin and Faster. This kind of healthy competition is the explosion of the entire short video industry. I hope the end result will be like this.

    For litchi, we have formed our own UGC ecology. This takes time, patience, and accumulation. The relative advantage is that the moat will be higher. Even if it is a giant to do it, it also takes time and accumulation, not so fast.

    : Are the Douyin and Quick Hands essentially taking your market?

    Lai Yilong: I don’t think so, I think it is supplementary, and even in disguise makes the demand for audio stronger. Judging from the data, the user usage time of our platform is indeed increasing, which has increased by 20-30% this year. Because short videos are very lively, people will feel empty after the lively time. At this time, emotional consolation is needed. Audio is a good takeover, especially for users below the third and fourth tier cities with more time. We have done a lot of user research, and they have all mentioned this.

    : Is litchi users sinking?

    Lai Yilong: Very sinking. We have many users in the 7th and 8th tier cities, in the county seat. They are more lonely and confused.

    : Was litchi positioned in the sinking market from the beginning, or did it naturally sink?

    Lai Yilong: For users, we have neverThere are preset positions and positioning. Sinking or more young female users are naturally formed.

    : Do you want to sink?

    Lai Yilong: We want to sink. Litchi makes popular and universal products, and young people in towns are the majority of users.

    In the Spring Festival of 2018, I rode a motorcycle from Panzhihua, Sichuan, and ran from county to county, about a week. I observe their lives in every county. I was particularly surprised that the restaurants in each county would do live broadcasts, selling lamb and kebabs, and become a channel to promote restaurants there. These are incomprehensible to us. Let’s take a look at the public comments when we eat. They aren’t. They’re brushing which one sells today and which one is fresher.

    The same goes for the clothes sellers. Those aunts were walking around wearing clothes. A group of people were watching her live broadcast, and some people came to the store after seeing the live broadcast at home. You can hardly find these brushes on the fast hand, but if you are local, you can see them in the recommendation.

    This is what I observed at the beginning of 2018, and the fast hand has become a local delivery platform in the county seat. The audience of Kuaishou has always been these small town youths, and they are very valuable. So Must go to the middle of the masses and see with your eyes what you originally thought was absolutely different.

    : Did the rise of the fast hand enlighten you? They are the products of young people in towns.

    Lai Yilong: Faster is more extreme and sinking than us. Fast-handed this market, frankly, I saw it in 2010. The product of the last venture, the “183 real-name community” aimed at famous agricultural workers, wanted to be this market. But the way we do it is from the offline, one by one factory to find them to register, it is too slow, because of the restrictions of the times, there is no way, smart phones were not popular at that time. Later, the shape of this product was a bit like WeChat. The IM of acquaintances was social. As soon as WeChat came out, I felt that there was no chance.

    But this experience changed my perception of the circle. The circles in China are very large, and the differences between each circle are beyond our imagination. So do n’t use your own aesthetics and cognition to guess users. This will be very helpful for me to make litchi later.

    At the time, I lived around Foxconn for two years. I also took Liu Qin (morningFounding partner of Xing Capital) distributed leaflets at the entrance of Foxconn to observe what mobile phones the workers used and what software was loaded in the mobile phones. It was found that everyone used the same kind of cottage machine.

    I took Liu Qin to the mobile phone store next to it. There were more than 300 copycat machines in it, and there was an unattended 2,000 OPPO function machine. I said that if there are six or seven hundred feature phones with better quality than the cottage, they will definitely be sold. Later, Liu Qin voted for Xiaomi, and within a few years Xiaomi killed the cottage.

    Say back to fast hands, fast hands products are very simple, I told our product team every day to keep it simple, because users do n’t want to learn. But it is really difficult to make it simple. You have to make up your mind to remove some unnecessary things. In this regard, we still have to learn like quick hands.

    : Do you meet Su Hua a lot?

    Lai Yilong: There are many intersections. When I was doing AI distribution, I talked to him a few times. What he said above is to do intelligent distribution. It is nothing more than understanding the content, understanding the user, and understanding the user’s behavior is what he told me. He is a science and engineering man. I am more objective and calm, so I am more emotional.

    : Do you want to stay biased or more rational?

    Lai Yilong: This is actually a difficult problem. Sensibility will bring me some frustration. For example, I want to use a Braun opener for the recording interface, but it is not necessary.

    It took me a long time to balance what I liked and what users really wanted. Sometimes content creators can easily fall into the “chauvinism of content”, which means that my aesthetics are leading, users need to obey, and in the end, the product often becomes the product manager’s own service. In order to avoid this, we must do a comprehensive AI distribution and completely abandon the personal aesthetics of the editor.

    But lychee is an emotional product, and I can’t completely get rid of sensibility, so I need to find a balance between sensibility and reason, which is a long-term study.