#21 单条爆款短视频1.05亿播放，AI如何赋能出海营销？对话Leap创始人Will_原文
#21 How Can AI Empower Overseas Marketing? A Conversation with Leap Founder Will on Viral Short Videos with 105 Million Views

2025年07月22日 10:35
July 22, 2025 10:35

发言人1 00:10
Speaker 1 00:10

以前我们是做搜索级优化，是教育这个工作。百度的搜索机现在其实我们是通过大量的短视频，包括图文帖子的内容加上hash care去教育这些算法。
Previously, we focused on search-level optimization, which was about educating this process. Baidu's search mechanism now actually relies on massive short videos, including graphic post content combined with hashtags to educate these algorithms.

发言人1 00:25
Speaker 1 00:25

只要把小爆款，也就是小脉冲，我们叫百万级的小脉冲做好千万级的。通常来说我目前的规律大概是10到20条左右的小爆款，其中会有一条冲向过千万，甚至于说过亿。
As long as we optimize the small hits—what we call "small pulses," meaning million-level performances—to reach tens of millions. Generally speaking, based on my current observations, out of about 10 to 20 small hits, one will break through to tens of millions or even hundreds of millions.

发言人1 00:44
Speaker 1 00:44

其实我们经常做内容的时候，我们会发现就有点像谈恋爱。我经常用谈恋爱说一个例子，如果你每一次的跟你的女生的接触都是奔着结婚去的，也就是说奔着售卖产品，往往这次的这个约会其实对女生的感受是不自然。
Actually, when we create content, we often find it somewhat similar to dating. I often use dating as an analogy: if every interaction with a girl is aimed at marriage—that is, focused on selling a product—the experience often feels unnatural for her.

发言人1 01:07
Speaker 1 01:07

尤其对于高客单产品，如果不及时去对话，很容易被流失。因为美国人他会觉得是不是你这个企业不够大，你的售后服务可能跟不上。所以这个评论的技术维护其实也是要用到ai。肯定是一点多亿的播放，然后TK是六千多万，YT目前是四千多万，其实就是在于。
Especially for high-ticket products, if you don't engage in timely communication, it's easy to lose customers. Americans might think your company isn't large enough or that your after-sales service might be inadequate. So, technical maintenance of reviews also needs to leverage AI. The view count definitely exceeded 100 million, with TikTok at over 60 million and YouTube currently at over 40 million. It really comes down to...

发言人1 01:34
Speaker 1 01:34

就是谁擅长做哪一块，谁就整合进来。但最终考虑的是企业一号位他的这个团队文化的适配度。就是你怎么从一个中国文化变成全球文化，甚至各个国家文化的一个储备，这个其实是最大的挑战。
It's about leveraging whoever excels in a particular area and integrating them. But ultimately, what matters is how well it aligns with the team culture of the company's top leader. The biggest challenge is transitioning from a Chinese culture to a global one, or even adapting to the cultures of various countries—that's the real test.

发言人3 01:51
Speaker 3 01:51

欢迎来到出海相对论。
Welcome to Overseas Relativity.

发言人4 01:53
Speaker 4 01:53

大家好我是J大家好，我是佳瑞。
Hello everyone, I'm J Hello everyone, I'm Jiarui.

发言人3 01:56
Speaker 3 01:56

本期我们的话题是出海营销，尤其是视频营销与AI的结合，以及相关的爆款视频方法论内容与数字营销打法。AI大客户全球化、区域化与本地化等话题。彼得德鲁克曾说，企业有且只有两大基本职能，营销和创新。在营销领域，数字营销无疑是如今的主旋律。根据stadia的报告，2023年全球广告支出总额的70%都流向了数字化渠道，它官网的数据也显示，2024年数字广告市场的广告支出将达到7403亿美元。
This episode focuses on overseas marketing, particularly the integration of video marketing with AI, along with methodologies for creating viral videos and digital marketing strategies. Topics also include AI-driven globalization, regionalization, and localization for major clients. As Peter Drucker once said, a business has only two basic functions: marketing and innovation. In the realm of marketing, digital marketing is undoubtedly the dominant theme today. According to a Stadia report, 70% of global advertising spending in 2023 flowed into digital channels. Data from its official website also shows that digital ad spending is projected to reach $740.3 billion in 2024.

发言人3 02:34
Speaker 3 02:34

那在数字营销中，视频营销自然是重中之重。因为我们的大脑天然处理视觉信息的速度就比文字快得多。麻省理工学院的一项研究发现，只需13毫秒，我们就能识别出自己看到的图像。随着注意力越来越分散，视频就是吸引用户注意的最佳选择。另一项最新的研究也表明，91%的用户会通过观看视频来了解产品或服务的更多信息。同时有82%的人会通过观看视频来决定是否要购买产品或服务。
Within digital marketing, video marketing naturally takes center stage. Our brains inherently process visual information much faster than text. A study by MIT found that it takes just 13 milliseconds for us to recognize an image we see. As attention spans grow shorter, video has become the best way to capture user interest. Another recent study reveals that 91% of users watch videos to learn more about products or services, while 82% rely on videos to decide whether to make a purchase.

发言人3 03:07
Speaker 3 03:07

随着深圳市AI引发的全球生产力革命，从流程自动化、超级个性化到永远改变创意生成的过程，生成式AI也将在新时代的营销各个环节中扮演重要的角色。麦肯锡的一份报告就预测，生成式AI仅在营销领域，每年就可为全球带来约4630亿美元的生产力提升。福布斯也在一篇名为2025年将改变你业务的六大AI营销趋势中写道，如果说内容为王，那么视频仍然是其中的瑰宝。AI驱动的视频制作工具使大规模创建个性化有针对性的视频成为可能。而可购物视频和直播等新形式也为品牌开辟了令人兴奋的可能性。用户生成内容是一股强大的力量，但创建和分享这些内容的平台和工具正变得越来越复杂。到2025年，创建、优化和分发视频内容的能力将不再取决于技术能力，而更多的取决于战略应用。
With Shenzhen's AI-driven global productivity revolution, from process automation to hyper-personalization and permanently altering the creative generation process, generative AI is set to play a pivotal role in every aspect of marketing in the new era. A McKinsey report predicts that generative AI alone could boost global productivity by approximately $463 billion annually in the marketing sector. Forbes also highlighted in an article titled "Six AI Marketing Trends That Will Transform Your Business by 2025" that if content is king, video remains its crown jewel. AI-powered video production tools now enable the large-scale creation of personalized, targeted videos. Meanwhile, new formats like shoppable videos and live streaming are opening up exciting possibilities for brands. User-generated content is a powerful force, but the platforms and tools for creating and sharing such content are becoming increasingly sophisticated. By 2025, the ability to create, optimize, and distribute video content will no longer hinge on technical prowess but rather on strategic application.

发言人3 04:08
Speaker 3 04:08

我们今天有幸邀请到的嘉宾，就是将AI的各项能力应用在视频和数字营销中，持续打造爆款营销视频，带动品牌增长，并且立志成为AI驱动的全球数字导购的leap的创始人will。如标题所说，截至目前will领导的leap曾打造过单条ins短视频1.05亿播放量，单条tiktok短视频6140万播放量的记录。在youtube的数据表现也非常好，并且leap能够始终维持相对稳定的爆款率。
Our guest today is Will, the founder of Leap, who has harnessed AI's capabilities in video and digital marketing to consistently produce viral marketing videos that drive brand growth. Leap aspires to become a globally recognized AI-powered digital shopping guide. As the title suggests, under Will's leadership, Leap has achieved records such as a single Instagram short video with 105 million views and a single TikTok short video with 61.4 million views. Their performance on YouTube is equally impressive, and Leap has maintained a relatively stable rate of producing viral hits.

发言人3 04:43
Speaker 3 04:43

Vivo本人也是出海营销的老炮了，虽然他实际年龄并不大。之前在国内大数据赴美上市第一股国双和出海移动营销第一股汇量都当过多年的高管，为众多的出海全球化企业拿到过响当当的成果，这也是他的博客首秀，期待今天的交流。回到惯例，我们先请威尔和大家打个招呼，也简单介绍一下你的经历，还有现在leap在做的事情。
Vivo himself is a seasoned veteran in overseas marketing, despite his relatively young age. He previously served as a senior executive for many years at Gridsum, the first Chinese big data company to list in the U.S., and at Mobvista, the first overseas mobile marketing company to go public, achieving remarkable results for numerous global enterprises expanding abroad. This marks his blog debut, and we look forward to today's discussion. As per tradition, let’s first invite Will to greet everyone and briefly introduce your background, as well as what Leap is currently working on.

发言人1 05:09
Speaker 1 05:09

谢谢。大家好，我是will，我的名字比较好记，因为那个姓付will will就是这个will be rich，所以我取这个名字也是希望未来有一个很好的愿景。然后我的这个经历也比较简单，我的经历分成三段。第一段经历就是在国内营销。随着百度的崛起，然后一直在百度、小红书、知乎，也就是围绕国内的内容流量的这个生态。第二段经历就是我把国内的这个流量的玩法。因为国内在这个移动互联网时代是运营手法是走的比较前的，我们就把国内这个运营的流量手法放到海外，所以我加入了这个汇量，然后在汇量过程中，就是见证了这个剔透的惊喜。所以第三股浪潮就是我们希望说短视频这个浪潮，我是不是可以有幸直接参与这个进程。所以第三段就是我直接出来做了这家公司，然后我们现在聚焦的是通过AI的技术和大批量的人工的这些优化，解决多渠道的爆款视频的生成和分发的问题。
Thank you. Hello everyone, I’m Will. My name is quite easy to remember—since my surname is Fu, "Will" sounds like "will be rich," so I chose this name hoping for a bright future. My background is relatively straightforward and can be divided into three phases. The first phase was in domestic marketing. With the rise of Baidu, I worked at Baidu, Xiaohongshu, and Zhihu, focusing on the ecosystem of domestic content traffic. The second phase involved applying domestic traffic strategies—since China’s operational tactics in the mobile internet era are quite advanced—to overseas markets. So, I joined Mobvista, where I witnessed the explosive growth of TikTok. The third wave is about whether I can directly participate in the short video trend. Hence, the third phase led me to start this company, which now focuses on leveraging AI technology and large-scale manual optimization to solve the challenges of generating and distributing viral videos across multiple channels.

发言人3 06:13
Speaker 3 06:13

我们先来聊一聊，可能听众看到标题以后最关心的话题，就是爆款视频的打造。前面我们也提到，视频在某种程度上已经占据了用户最多的注意力。而爆款视频无疑是几乎每个商家或者说品牌都梦寐以求的，因为那意味着巨大的流量和影响力。市面上其实也流传着无数版本的所谓爆款视频的打法。比如说什么黄金3秒开头，加2到5个富有争议的观点，加互动式的结尾等等。从标题到拍摄每一个细节都有所谓的操盘手总结过很多公式。甚至还有很多人靠麦克就赚的盆满钵满。
Let's start by discussing the topic that listeners are probably most curious about after seeing the title: creating viral videos. As we mentioned earlier, videos have, to some extent, captured the most user attention. Viral videos are undoubtedly something almost every merchant or brand dreams of, as they represent massive traffic and influence. There are countless so-called "formulas" for creating viral videos circulating in the market. For example, the golden 3-second opening, adding 2 to 5 controversial points, an interactive ending, and so on. From the title to every detail of filming, there are so-called "operators" who have summarized many formulas. Some people have even made a fortune just by selling these formulas.

发言人3 06:52
Speaker 3 06:52

但事实上爆款视频依旧稀缺，尤其是商业产品方向的视频。而leap不仅在海外媒体上打造过我之前提到的1.05亿播放的ins短视频，6140万播放的TK短视频，还能持续的制造爆款。上次咱们聊我记得是6‰的爆款率。也有数据表明在流量暴涨之后，有搜索的溢出，还有交易量的增长。也想请问有具体聊一聊就是怎么在海外社媒持续打造爆款的营销视频，有没有什么底层的逻辑或者说什么能力？
But in reality, viral videos remain scarce, especially those related to commercial products. Leap has not only created the 105-million-view Instagram short video and the 61.4-million-view TikTok short video I mentioned earlier on overseas platforms but has also consistently produced viral hits. Last time we spoke, I remember the viral rate was 6‰. Data also shows that after a surge in traffic, there is spillover in search volume and growth in transaction volume. I’d like to ask specifically: how do you consistently create viral marketing videos on overseas social media? Is there any underlying logic or capability behind it?

发言人1 07:25
Speaker 1 07:25

OK首先我觉得这个爆款本身它是一个稀缺的，因为不是所有的这个视频都可以走向爆款。然后也不是说营销的部分，我们为爆款润，这个部分也很简单，因为我们自己做了很多的case，我们发现有的视频它其实播放量并不是很高，比如说只有几千，但是反而却有比较多的订单。所以我们拆解来看，其实爆款它会分成两大矩阵，一个是偏曝光的矩阵，一个是偏转化的矩阵。如果是为了破圈，那这种曝光其实是越多越好的。如果是为了转化，比如说我们所说的核心TA其实反而这个播放量并不是越大越好。
OK, first of all, I think viral videos themselves are scarce because not every video can go viral. Also, it’s not just about the marketing aspect—we don’t aim solely for virality. This is quite simple because we’ve worked on many cases and found that some videos don’t actually have very high view counts, say only a few thousand, but surprisingly generate a lot of orders. So, when we break it down, viral videos can be divided into two main matrices: one leans toward exposure, and the other leans toward conversion. If the goal is to break through the circle, then the more exposure, the better. But if the goal is conversion, such as targeting our core TA (target audience), then higher view counts aren’t necessarily better.

发言人1 08:02
Speaker 1 08:02

因为真正击穿这个消费者心智的一些CTA就是cota action的视频。首先它不具备这种viral的属性，同时它可以基于一些核心的消费者击穿拿到结果。所以还是目标不一样，他对这个爆款的定义不一样。然后如果我们聚焦在偏这个average，也就是说打认知阶段破圈的这样的一个视频，的确爆款是有一定的方法论。然后我通常把爆款的这个方法论，我在很多场合说这个其实是有点像波浪形的理论，就是我们叫脉冲。
Because the videos that truly penetrate consumer psychology are those with clear CTAs (call-to-action). First, they don't inherently possess viral attributes, yet they can achieve results by targeting core consumer insights. So the objectives differ, and so does the definition of what makes a "hit." If we focus more on the average—meaning videos that break through during the awareness stage—there is indeed a certain methodology behind creating hits. I often describe this methodology, which I’ve mentioned in many contexts, as resembling a wave-like theory, what we call "pulse marketing."

发言人1 08:34
Speaker 1 08:34

一般来讲的话，我们通过基于这个消费洞察，比如说了解客户的这个产品品牌，还有竞品的一些已有的爆款视频。包括这个KO的爆款视频，我们还是可以总结出一些规律。比如说哪样的这个功能，比如说我们以电视为例，防水电视我防水电视可以放在户外，可以放在室内，也可以放在淋浴间。我们可以做非常多元的尝试，然后找到在哪一个场景，哪一个卖点和哪一个故事线，它有可能是可以报的。所以对我们来说就是追求不同的人需求，还有功能。包括这个破圈的，我们叫爽点的组合，它会形成多元的脚本。我们就会用相对高效率和相对偏低的成本去做一些赛马的测试。一旦测试成功，我们就再通过这个裂变的方式去放大这种的方法论。
Generally speaking, by leveraging consumer insights—such as understanding the client's product, brand, and existing hit videos from competitors, including KO's successful videos—we can identify certain patterns. For example, which features resonate? Take TVs as an example: a waterproof TV can be placed outdoors, indoors, or even in a shower. We can experiment with diverse scenarios to determine which setting, selling point, or storyline has the potential to go viral. For us, it’s about addressing different consumer needs and functionalities, including what we call the "wow factor" combinations that break through. These elements form diverse scripts. We then conduct efficient, low-cost A/B tests. Once a test succeeds, we amplify the methodology through a fission-like expansion strategy.

发言人1 09:27
Speaker 1 09:27

SOP其实是可以有效的控制一个我们叫小爆款的发生率。但是我们没有办法决定大爆款，因为我们自己对推荐算法的理解是通常大规模的小爆管。如果我们多加了一些hash tag去教育这个推荐算法。有点像是以前我们是做搜索引擎优化，是教育google，百度的搜索引擎。现在其实我们是通过大量的短视频，包括图文帖子的内容加上hashtag去教育推荐算法。所以当我们有在短时间内有很多的小脉冲的时候，然后他都带上同样的hashtag。那这个时候其实教育这个大脉冲的机会就会变得越来越大。整个流量的逻辑在这个社媒上就像是一个波浪形的脉冲。小脉冲我们是可以预见的。因为当赛马足够多的时候，赛马的频次，赛马的多元度，包括赛马的裂变的次数，这个小脉冲是一定可以预见的。
SOP can effectively control the occurrence rate of what we call small viral hits. However, we cannot determine major viral hits because our understanding of recommendation algorithms suggests that large-scale small viral hits are more common. If we add more hashtags to train the recommendation algorithm, it’s somewhat similar to how we used to do search engine optimization (SEO) to train Google or Baidu’s search engines. Now, we essentially train the recommendation algorithm through a large volume of short videos and graphic posts, supplemented with hashtags. So, when we have many small pulses in a short period, all tagged with the same hashtag, the likelihood of triggering a major pulse increases significantly. The logic of traffic flow on social media resembles a wave-like pulse. Small pulses are predictable because when there are enough "horse races" (content variations), the frequency, diversity, and fission rate of these races make small pulses almost certain to occur.

发言人1 10:20
Speaker 1 10:20

大脉冲我觉得更多时候是一个有点像时间匹配，就比如说刚好这段时间这个场景下的这个内容的卷度没有那么高。比如说我们搜防水电视，三星LG并没有进场的时候，那如果我们经营的这个品牌，它刚好在这个词下的内容小脉冲很多，就容易把小脉冲推向大脉冲，也就是看到大爆款。所以我们在内部的时候，我要求团队其实不太会追求所谓的大爆款。我们更多是把可工程化的小爆款变得sustainable。大爆款的过程中你会发现只要把小爆款，也就是小脉冲，我们叫百万级的小脉冲做好千万级的。通常来说目前的规律大概是10到20条左右的小爆款，其中会有一条冲向过千万，甚至于说过亿。
Major pulses, I think, are more about timing alignment. For example, it could be that during a certain period, the competition for a specific type of content isn’t as intense. Take "waterproof TVs" as an example—if Samsung and LG haven’t entered the market yet, and our brand happens to have many small pulses around this keyword, it becomes easier to push these small pulses into a major pulse, resulting in a viral hit. That’s why internally, I don’t encourage the team to chase so-called major viral hits. Instead, we focus on making small, scalable viral hits sustainable. In the process of achieving major viral hits, you’ll notice that as long as small viral hits—what we call "million-level pulses"—are optimized to reach "ten-million-level," the current pattern suggests that out of 10 to 20 small viral hits, one will break through to tens of millions or even hundreds of millions of views.

发言人1 11:09
Speaker 1 11:09

这里面展开来说，其实不同的行业的大爆款的上限是不一样。我们做了一个很知名的品牌，像美的集团。你就会发现它不同的这个事业线，它其实大脉冲是不一样的。每次我们拿到的需求说，为什么我的产品不能像另外的一个事业线的产品有更大的脉冲。但其实我们分析来看，就是如果我们把不同行业的KO的数据，比如说像一些20万粉丝以下的KL数据的播放拿下来。
To elaborate further, the upper limits for viral hits actually vary across different industries. We've worked with well-known brands like Midea Group, and you'll notice that different product lines have distinct viral potential. Often, clients ask why their products can't achieve the same level of virality as other product lines. However, our analysis shows that when we examine KO data across industries - for instance, viewership data from KL accounts with under 200,000 followers -

发言人1 11:36
Speaker 1 11:36

每一个不同产品线它所在的这个竞争内容的卷度，包括这个产品本身的这个痒点、爽点带来的大脉冲的上限就是不一样。比如说我们做烤箱，它的上限可能就在500万。我们做扫地机它的上限可能就是在5000万。所以这里面其实不是一个横向对比。
Each product line operates in content ecosystems with different competitive intensities, and the upper limits of virality determined by the product's pain points and pleasure points naturally differ. For example, ovens might cap at 5 million views, while robotic vacuums could reach 50 million. This fundamentally isn't an apples-to-apples comparison.

发言人1 11:54
Speaker 1 11:54

我们更希望是说把我们的小脉冲，包括我们换来大脉冲，去跟20万以下的这个KL的脉冲去做对比，这里面为什么是20万以下的粉丝呢？是因为这个低于20万以上的，它其实它的脉冲有可能是来源于它的原始粉丝的积累，还有他的人设和号召力。20万以下其实通常就是比较拼内容，因为我们是用这种矩阵的旗号的方式帮品牌做自营。所以这个时候如果同样的去弱化人设带来的脉冲影响力，纯PK内容的话，其实我们更大的程度是找到我们的相对比较优势。就是说我们的这个小脉冲和大脉冲的频率，包括大脉冲的上限是能够远胜于腰部以下的这个KL就是20万以下的那这样的话其实这个过程中是可以被方法润滑或者工程化。也就是说随着我们做这个短视频营销做的越久，其实品牌开始会对自己的增长有一定的预期。
We prefer to compare our small pulses, including our large pulses, with those of KOLs (Key Opinion Leaders) who have fewer than 200,000 followers. Why focus on those with under 200,000 followers? Because for those with fewer than 200,000, their pulse might stem from their accumulated original followers, as well as their persona and influence. Below 200,000 followers, it’s usually more about competing on content quality. Since we use a matrix-based approach to help brands build their own operations, by minimizing the influence of personal branding and purely competing on content, we can better identify our relative advantages. That is, the frequency of our small and large pulses, as well as the upper limit of our large pulses, can far surpass those of mid-tier and lower-tier KOLs (i.e., those with under 200,000 followers). This process can be methodically refined or engineered. In other words, the longer we engage in short-video marketing, the more brands can develop certain expectations for their growth.

发言人1 12:47
Speaker 1 12:47

比如说我们每周会有1到2个小麦冲过百万，然后每个月大概会有一条超过小麦冲上限的。但是一个亿还是5000万，甚至两个亿，至少我目前也不确定的。而且这个竞争的卷度随着时间的变化也是不确定的。但是这个小脉冲像这个一浪接一浪，这个相对还是有据可循的对。
For example, we typically have 1 to 2 small pulses exceeding a million views per week, and roughly one pulse per month that surpasses the upper limit of small pulses. However, whether it’s 100 million, 50 million, or even 200 million views, even I can’t be certain at this point. Moreover, the intensity of competition is unpredictable over time. But these small pulses, like waves rolling in one after another, are relatively more traceable. Right.

发言人4 13:11
Speaker 4 13:11

就是先踏踏实实做一些基础的工作，把基础工作都做到位了，就跟当年公众号的10万加一样，这个是运气。然后基础工作拼的是实力。
First, focus on laying solid groundwork by completing all fundamental tasks diligently. Just like achieving 100,000+ reads on public accounts back in the day, luck plays a role. But the foundation relies on capability.

发言人1 13:22
Speaker 1 13:22

其实我们经常做内容的时候，我们会发现就有点像是谈恋爱。我经常用谈恋爱说一个例子，如果你每一次的跟你的女生的接触都是奔着结婚去的，也就是说奔着售卖产品，往往这次的这个约会其实对女生的感受是不自然。那反而是没有办法激起他心中的这个涟漪，也就是浪花，也就是这个脉冲。所以我们追求小脉冲的时候，是当我们不太一味的把大脉冲当成目标的时候，我们只是做不断的不同场景的约会，讲着不同的话术，送着不同的礼物。我们只是在做这样的一个我们叫有迹可循的尝试的时候。女生总归在一场电影和一个城市，比如说是巴黎，还有一个我们穿着西装的这样的一个风格会被打动。那这个moment是我们要放大的，因为我们已经找到这个女生喜欢的风格。品牌和消费者的接触其实和这个原理是相对类似的。
Actually, when creating content, we often find it somewhat akin to dating. I frequently use dating as an analogy: if every interaction with a woman is solely aimed at marriage—or in our case, selling a product—the experience feels unnatural for her. This approach fails to create ripples, or what we call "pulses," in her heart. When we pursue small pulses, we're not fixated on achieving big pulses as the ultimate goal. Instead, we focus on arranging various dates in different settings, using diverse conversational approaches, and giving thoughtful gifts. We're simply making traceable attempts. Eventually, there will be a moment—perhaps during a movie, in a city like Paris, or when we're dressed in suits—that truly resonates with her. That's the moment we need to amplify, because we've discovered her preferred style. The interaction between brands and consumers follows a similar principle.

发言人1 14:18
Speaker 1 14:18

如果我们每一次约会都奔着像相亲一样说，我们什么时候可以结婚，你什么时候可以带上我给你买戒指。往往这样的一个效果，其实是没有我们之前说的这种多脚本测试的效果来的。好OK顺着。
If every time we date, we approach it like a blind date, asking when we can get married or when you can wear the ring I bought you, the outcome often isn't as effective as the multi-script testing approach we discussed earlier. Okay, moving on.

发言人3 14:32
Speaker 3 14:32

刚才的话题我们延伸一点，聊一聊出海数字营销。上次咱们沟通时你也聊到在跑通视频营销之前，lip其实在AI声纹、AI声图等方面都有过尝试，也都取得过了一些结果。同时因为你的背景刚才也介绍了，比如说会亮就是专门做出海移动营销的，所以你们团队在海外社媒营销SEO广告投放的数字营销方面，也都有长期的经验。所以视频营销其实只是你们业务中的一环。虽然可能是最重要的一环，但实际上立还有中后端的各个业务能力是可以提供相对完整解决方案的，或者说是可以打通整个营销链路的。又因为海内外主流的电商平台媒介消费习惯和文化等方面的不同，显然这一套链路与国内市场也是不同的那也想请你结合过去的经验和大家分享一下，海外市场一套完整的数字营销链路是怎样的，有没有什么具体的案例？
Let’s expand a bit on the previous topic and talk about overseas digital marketing. Last time we spoke, you mentioned that before refining video marketing, Leap had already experimented with AI voiceprints, AI voice-to-image, and other areas, achieving some results. Given your background—as you just mentioned, for instance, Will specializes in overseas mobile marketing—your team also has extensive experience in overseas social media marketing, SEO, and ad placements. So, video marketing is just one part of your business. While it might be the most critical part, Leap actually offers a relatively comprehensive solution by integrating mid- and back-end capabilities, or even connecting the entire marketing funnel. Due to differences in mainstream e-commerce platforms, media consumption habits, and cultural factors between domestic and international markets, this funnel is clearly distinct from the domestic one. Could you share with us, based on your past experience, what a complete overseas digital marketing funnel looks like? Are there any specific case studies?

发言人1 15:29
Speaker 1 15:29

OK, 首先我讲一下我们公司的目前的聚焦，其实现在已经升级到了说我们聚焦在视频车。这原因很简单，因为在我去年出来创业的时候，其实做的所有的尝试，因为创业是九死一生的，所以我们去年做了很多尝试，其实失败的。我可以解释一下，为什么在AI做图的部分其实没有拿到很好的结果。但是他现在却因为你做短视频营销的时候，有些时候你会发图文帖，所以它会变成我们现在做矩阵图文帖的一部分。但是图单独的这个业务线，我们是经历了一次去年的失败。
OK, first, let me talk about our company's current focus, which has now evolved to concentrate on video content. The reason is quite simple. When I started my entrepreneurial journey last year, we tried many approaches because starting a business is a high-risk endeavor. Most of our attempts last year actually failed. I can explain why we didn’t achieve great results in the AI-generated image segment. However, when it comes to short video marketing, sometimes you still need to post image-text content, so it has become part of our current matrix for image-text posts. But as a standalone business line for images, we experienced a failure last year.

发言人1 16:04
Speaker 1 16:04

然后主要是因为我们当时的应用场景是聚焦在了我们叫亚马逊，或者说是独立站的落地详情页的优化。因为这里面有很多图和文，然后这个时候我们就发现，不管是用SD还是用MJ的这个组合，去年的这个AI的水平在验收的过程中，它的成本是非常高的。就是当我给了非常多的这个版本的内容之后，给到客户，客户他会用一些feeling的这个验收的标准。比如说我的审美感觉，这样的话带来重复的沟通，包括说这个制作的工作，他没有找到一个验收的供应点。
The main reason was that our application scenario at the time was focused on optimizing landing pages for Amazon or independent e-commerce sites. These pages involve a lot of images and text, and we realized that whether using Stable Diffusion (SD) or MidJourney (MJ), the AI’s capabilities last year resulted in very high costs during the approval process. Even after providing clients with numerous versions of content, they would evaluate it based on subjective feelings, such as personal aesthetic preferences. This led to repeated communication and production efforts without a clear benchmark for approval.

发言人1 16:37
Speaker 1 16:37

然后我们为什么做视频的时候，我们相对来讲找到一定的这法规，是因为首先视频它的审美价值他并没有像图文那么的高。也就是说这个视频30秒或者说15秒，它只要相对的有趣好玩。也就是从消费者ugc角度，这个视频本身它的趣味性够的话，其实它不会太追求审美的下限。所以它的审美下显示比图要低的。然后第二个是我们的验收方式，从原来的靠感觉的共识验收，到现在进入到了通过这个我们叫流量的共识验收。比如说我们的验收并不是看我们的视频的数量，而是看我们完成了多少的声量、流量甚至于是销量。所以我们的结算方式一般是CPM、CPC.
Then why do we find certain regulations when making videos? It's primarily because the aesthetic value of videos isn't as high as that of images and text. In other words, a 30-second or 15-second video only needs to be relatively fun and engaging. From the perspective of consumer-generated content (UGC), as long as the video itself is entertaining enough, people won’t be too demanding about its aesthetic threshold. So its aesthetic standard is lower than that of images. Secondly, our evaluation method has shifted from relying on subjective consensus to what we call "traffic consensus evaluation." For example, our assessment isn’t based on the quantity of videos we produce but rather on how much buzz, traffic, or even sales we generate. That’s why our settlement methods are typically CPM or CPC.

发言人1 17:22
Speaker 1 17:22

甚至于说如果这个客户和我们合作的比较深度了，我们可以直接按照RY对赌的方式去做结算。所以当我们有增量流量销量的共识作为结算依据的时候，我们就会规避掉很多时候在感觉共识上面花的这个精力和时间。这也是为什么我们现在其实是聚焦在短视频这个部分。然后再回到你说的这个问题，就是说整个海外营销行业切分的板块。首先海外营销整个线上市场也只占了不到20%。
Moreover, if a client collaborates with us more deeply, we can even settle accounts directly based on a return-on-investment (ROI) betting model. So when we use incremental traffic and sales as the basis for settlement, we avoid much of the effort and time spent on subjective consensus. This is also why we now focus primarily on short videos. Returning to your question about the segmentation of the overseas marketing industry, first of all, the entire online market for overseas marketing accounts for less than 20%.

发言人1 17:48
Speaker 1 17:48

我们以美国为例，所以海外营销其实它更大的程度是线下。我们先就数字营销来拆出来看，基本上是三个板块。第一个板块当然是这个消费洞察，就是我需要先很了解我的消费者，所以我可以通过这种social类似于的方式，也就是线上的聆听的方式，或者说线下调研的方式。现在主流的做法当然是通过线上的VOC social listening的方式。因为它的成本还有它的数据量够大，以线下作为补位。这个解决的问题就是说我对消费者说什么，也就是说消费者到底是讲什么画像。
Take the U.S. as an example—overseas marketing is largely offline. If we break it down in terms of digital marketing, there are essentially three segments. The first segment is consumer insights, meaning I need to deeply understand my consumers. This can be achieved through methods like social listening—essentially online monitoring—or offline surveys. Currently, the mainstream approach is online VOC (Voice of Customer) social listening because of its cost-effectiveness and the sheer volume of data available, supplemented by offline methods. This addresses the question of what to say to consumers and what their profiles look like.

发言人1 18:19
Speaker 1 18:19

然后第二个流程就是内容生产。然后内容生产包括像图文生产、短视频生产、长视频生产，包括甚至于说像一些VRTVC，甚至于说直播，这都属于叫内容生产环节。然后第三个环节也是最重要的环节是内容的分发。也就是说我有了内容，其实我是需要发出去的。但是分发其实是有成本的。比如说你买广告，其实是有对应的流量的CBC的成本或者CPM的成本。然后你通过KOLKL可能会需要坑位费。如果你通过一些线下的banner或者线下的户外大牌，它都是有对应的这个价格。
The second process is content production. This includes graphic and text production, short video production, long video production, and even formats like VR/TVC, as well as live streaming—all of these fall under the content production phase. The third and most critical step is content distribution. In other words, once we have the content, it needs to be disseminated. However, distribution comes with costs. For example, purchasing ads involves corresponding traffic costs like CBC or CPM. Collaborating with KOLs/KOCs may require placement fees. Even offline channels like banners or outdoor billboards have their associated costs.

发言人1 18:53
Speaker 1 18:53

当然如果这些内容你都觉得说分发成本比如说相对过于高的时候，还有一种方式就是通过自营的触点。比如说我们说私域的触点，what's up faced group, 或者说自己的官网，包括我们现在经常在做的这个举证号edm这些触点相对来讲是偏低成本。也就是说我们做好的内容可以实现一个分发。所以基本上是消费洞察，也就是洞察内容生产到内容分发。
Of course, if you find the distribution costs too high, there’s another approach: leveraging owned touchpoints. For instance, private domain touchpoints like WhatsApp groups, Facebook groups, or your own official website, as well as the matrix accounts and EDM campaigns we frequently work on—these touchpoints are relatively low-cost. Essentially, well-produced content can achieve distribution through these channels. So the core workflow consists of consumer insights, content production, and content distribution.

发言人1 19:17
Speaker 1 19:17

然后最后一个环节，就是看这个企业的生意的规模。如果说这个规模比较大的话，其实需要建立一个循环的飞轮。也就是说这个数据的反馈机制。就是当我经历了一轮洞察到生产到分发，我的效果如何？我要迭代我的这个洞察的方式，我要迭代我内容生产的这个内容方向甚至内容形式。包括迭代我的这个全域分发的触点，然后拿到更好的效果。
Then the final step is to examine the scale of the business. If the scale is relatively large, it actually requires establishing a cyclical flywheel—essentially, a data feedback mechanism. That is, after going through a cycle of insights, production, and distribution, how effective was it? I need to iterate on my approach to insights, iterate on the content direction or even the format of my content production, and iterate on the touchpoints of my omnichannel distribution to achieve better results.

发言人1 19:44
Speaker 1 19:44

然后这个效果也是综合来看，其实不应该只是看说一个内容是不是马上带来订单，应该是要分短期价值和长期价值。比如说有哪些触点和哪些内容适合即时消费的，就是短期内容很高的RI然后哪些内容和哪些触点是适合长期主义。比如像家瑞的句子，专注的其实偏长期主义。
Moreover, this effectiveness should be evaluated comprehensively. It shouldn’t just focus on whether a piece of content immediately drives orders; short-term and long-term value should be differentiated. For example, some touchpoints and content are suited for immediate consumption—content with high short-term ROI—while others are more aligned with long-term strategies. Take Jiarui’s sentences, for instance, which focus more on long-termism.

发言人1 20:05
Speaker 1 20:05

包括像一些高科技的产品，其实它的很大的销售场景不是看到一个爆款视频马上下单。我们作为一个猫砂盆的案例，我们就发现我们花了基本上一个月可以做接近两个亿的播放，但是他的订单却没我们想象中有几百个消费行为非常有意思。他是看完我们的视频进入到独立站，然后我们跟踪他的行为。我们通过SEM rush这种第三方工具发现他进完独立站的下一步是去了FB。然后我们再深入看FB的时候，发现当有爆款视频的时候，很多人是跑到FB找到这个猫砂盆有没有对应的FB group。也就是说当我买一个几百美金的猫砂盆，其实我不光是看到你的内容觉得有意思，看你的观望解决了信任。其实我还非常想知道有没有其他的已购的宠物的这些用户，他用你的产品的体验如何？他打消了最后疑虑。
Including some high-tech products, their major sales scenario isn't making an immediate purchase after seeing a viral video. Take our cat litter box case as an example—we found that spending about a month could generate nearly 200 million views, but the actual orders didn't match our expectation of hundreds of conversions. The consumer behavior here is quite interesting. After watching our video, they'd visit the independent website, and we tracked their actions. Using third-party tools like SEMrush, we noticed their next step after the website was heading to Facebook. Upon deeper analysis of Facebook, we discovered that when a video went viral, many users would search Facebook to see if there was a corresponding group for that cat litter box. In other words, when purchasing a cat litter box costing several hundred dollars, it's not just about finding the content interesting or having trust built through the website. They also really want to know if there are other pet owners who've bought the product and what their experiences are. This helps eliminate their final doubts.

发言人1 20:55
Speaker 1 20:55

所以从last click，也就是说最后一次触点，结婚的最后一个触点来说的话。其实往往是这个私欲环节。所以我们说全域触点它有点像是消费旅程，每个旅程都挺重要的。不能说单位的说我们做数字营销只关注我们这个触点的内容有没有RY。我觉得这个相对来讲，短期低客单的是可以以偏概全的。如果是长周期的高客单的，还是要拉到这个全链条来看这样的一个整体。
So, from the perspective of the last click—the final touchpoint before conversion—it often comes down to this private domain interaction. That's why we say omnichannel touchpoints are somewhat like a consumer journey, where each step is important. We can't narrowly focus on just whether the content at our touchpoint has a good return on investment (ROI). This approach might work for short-term, low-ticket items, but for high-ticket, long-cycle products, we need to evaluate the entire chain holistically.

发言人4 21:23
Speaker 4 21:23

所以从你的视角来讲，从洞察到内容生成再到分发，然后再归因又循环走到下一个循环。
So from your perspective, it's about moving from insights to content generation, then to distribution, followed by attribution, and cycling back to the next iteration.

发言人1 21:31
Speaker 1 21:31

因为我会非常希望整个工作流可以被数字化。所以我通常跟团队沟通的时候说，这三个流程有没有办法被足够的工程化。就是说第一步是我们把所有的洞察到生产到分发先数字化。数字化的主要过程就是标签化。比如说我有没有一些人的标签，内容的标签、需求的标签、货品的标签。
Because I strongly desire the entire workflow to be digitized. So when I communicate with the team, I often ask whether these three processes can be sufficiently engineered. That is, the first step is to digitize everything from insights to production to distribution. The main process of digitization is tagging. For example, do we have tags for people, content, needs, and products?

发言人1 21:55
Speaker 1 21:55

然后第二步，其实是我基于这样的一个标签去做更好的美每个环节的生成。所以如果我们拆解一下，你就会发现，首先洞察里面已经有很多AI公司可以基于一些大数据的方式快速提炼一些消费的方向。包括产品的这个优劣势的卖点。其实有助于我们去找到和消费者沟通的故事线，这是洞察部分，其实数字化是可以解决的。然后沉淀了一些内容的标签和人群的标签。
The second step is actually about leveraging these tags to enhance generation at each stage. If we break it down, you'll notice that many AI companies can already quickly identify consumer trends through big data analytics during the insight phase. This includes pinpointing product strengths and selling points, which helps us craft communication narratives with consumers. The insight phase can indeed be addressed through digitization, resulting in accumulated content tags and audience segmentation tags.

发言人1 22:22
Speaker 1 22:22

内容生产的部分，我们今天看到很多人说，sorry出来了，这个视频行业要被颠覆。真实的情况是这样的，就是我们现在通过外籍留学生，包括这个多脚本的赛马，其实带来的播放量，你会发现我们可以争取到的是千万，甚至5000万，甚至上亿。我看了一下所有的官号，不管是p OpenAI、runway他们全渠道的设备，他们的播放量少有能过百万的。哪怕OpenAI的这个instagram的官方账号，其实很多的视频不是那种宣传片的视频，其实播放量也就是十几万、20万。
Regarding content production, many people today claim that "sorry" (AI) has emerged and will disrupt the video industry. The reality is this: through collaborations with international students and multi-script competitions, we've achieved view counts in the tens of millions, even 50 million or over 100 million. I've examined all official accounts, whether it's OpenAI, Runway, or others across all platforms—their view counts rarely exceed a million. Even OpenAI's official Instagram account, where many videos aren't promotional clips, typically garners just 100,000 to 200,000 views.

发言人1 22:55
Speaker 1 22:55

所以这个时候做内容生产这部分，纵然可以用一些AI的技术，我们还是组合的来看。比如说一个30秒的视频，如果前5秒的引流开场是sorry runway这样的方式，后面的15到20秒还是一个真实的用户。其实它是兼具了AI生成的效率，加上这个视频的可看性。如果整个30秒全是AI5秒之间的组成，其实这样的视频咱们看多了，它很难推向新的流量池。所以我觉得这个人机交互在整个内容生产层面其实是可以发挥很大的作用。
So when it comes to content production, even though we can utilize some AI technologies, we still approach it from a combined perspective. For example, in a 30-second video, if the first 5 seconds use an attention-grabbing opener like "sorry runway," the remaining 15 to 20 seconds should still feature a real user. This approach combines the efficiency of AI generation with the watchability of the video. If the entire 30 seconds were composed of AI-generated 5-second clips, such videos—after seeing too many of them—would struggle to break into new traffic pools. Therefore, I believe human-AI interaction can play a significant role in the entire content production process.

发言人1 23:28
Speaker 1 23:28

然后再回到内容的这个分发。也就是说我们现在看到有很多的工具说，我可以做智能投放。比如说你瞄定一个rush，那AI就可以告诉你在不同的触点应该花多少钱。有点像是智能的归因评估和决策。市场上有很多这样的一个叫智能投放的AI的公司。它其实就基于你所有全流程的数据去分配你的预算。然后锚定一个rush，自动去调整你的这个预算的分配。
Now, let’s circle back to content distribution. Currently, we see many tools claiming to offer intelligent ad placement. For instance, if you target a rush, AI can tell you how much to spend across different touchpoints. It’s somewhat like intelligent attribution evaluation and decision-making. There are many companies in the market offering this so-called "intelligent ad placement" AI. Essentially, they allocate your budget based on your full-funnel data, anchor to a rush, and automatically adjust your budget distribution.

发言人1 23:55
Speaker 1 23:55

最后一环，其实就是比如说像长周期，因为有大量的用户，大量的评论。然后这个时候，其实靠人的方式是远远做不过来。因为我们每个月会帮我们的客户做2到3亿的播放。其实这样的评论通常是有过几十万甚至上百万的。这个时候靠人的方式去把这些评论变成私域用户的难度是非常大的，也不是成本可以解决的，因为我们招太多的人也不可以很好的解决这个问题。这就需要把所有拿到的评论，包括整个行业的一些对话的话术，包括一方的数据放到小模型里面去做一些训练。
The final step involves handling long cycles with massive user bases and comment volumes. At this scale, manual approaches become completely inadequate. We currently help our clients achieve 200-300 million monthly views, which typically generates hundreds of thousands or even millions of comments. Manually converting these comments into private domain users is extremely challenging and cost-prohibitive—hiring more staff wouldn't effectively solve this. The solution requires feeding all collected comments, industry-specific dialogue templates, and first-party data into small models for training.

发言人1 24:30
Speaker 1 24:30

这个时候也有一些第三方的专有的AI的agent，在整个客服环节，私域环节发力的很好。我们现在主要合作的是佳瑞这边在TK，包括未来在ins上的合作，目前效果还是非常不错。所以这是一个全流程。我们其实希望说整个flow洞察生产分发再到私域整个环节，AI agent也就形成一个非常高效的flow。人机的效率可以更强对。
There are also specialized third-party AI agents performing exceptionally well in customer service and private domain operations. Our current primary partnership is with Jiarui for TikTok collaborations, with future plans for Instagram—the results have been very promising. This forms a complete workflow. Our vision is to establish an efficient end-to-end flow from insights and content creation to distribution and private domain conversion, where AI agents create a highly effective workflow that significantly enhances human-machine efficiency.

发言人4 24:59
Speaker 4 24:59

感谢我非常好的合作伙伴，我跟leo也是认识好多年，然后有各种合作。这一次除了我们一起来录这次播客，然后will他的整个TK也是巨子新推出的tiktok产品的最初期的天使用户，以及我们从产品的打磨到最早期的天使用户的试用。因为原来熟悉我们的同学可能知道，我们原来都是在国内做。今年我们陆续推出了很多我们海外的产品。
Thank you to my excellent partner. I've known Leo for many years, and we've collaborated in various ways. This time, besides recording this podcast together, Will was also one of the earliest angel users of our newly launched TikTok product. From product refinement to early-stage user testing by angel users—those familiar with us might know we previously operated mainly in China. This year, we've gradually rolled out many of our overseas products.

发言人4 25:28
Speaker 4 25:28

其实从刚才没有聊到的整个的消费者的整个的数字营销，我们认为其实以GPT为代表，大模型其实开启了一个全新的生产力。而人类历史上从来没有一种能力可以让企业和每一个终端用户之间富有场景，富有温度的进行沟通。以前这种事就是RI算不过来，即使招人也算不过来，因为人是不可能去做这样的事儿的。但是最终的销售的转化可能就是需要这种信任。原来我们说各种的成交是概率成交，我们说是促我不停的推推推。接下来成交方式其实更多是从概率成交到信任成交。
Actually, regarding the broader digital marketing landscape for consumers that we haven't discussed yet, we believe that GPT and large language models represent a completely new productivity paradigm. In human history, there has never been a capability that allows businesses to engage with end users in such contextual and warm ways. Previously, this was simply unfeasible—even hiring more people wouldn't solve it, as humans couldn't possibly handle such tasks at scale. Yet, ultimately, sales conversions often hinge on this kind of trust. In the past, we talked about probability-based conversions—constantly pushing and pushing. Moving forward, the conversion model will shift more from probability-based to trust-based transactions.

发言人4 26:06
Speaker 4 26:06

企业和消费者之间建立了一个长期的信任的链接，建立了情感的链接。而这些东西是可以有机会能够被用agent去实现。从销售服务转化整个的全链路，通过AI更智能化的去实现它。
A long-term trust and emotional connection is established between businesses and consumers. These elements can potentially be achieved through agents. The entire sales and service conversion process can be more intelligently realized through AI.

发言人1 26:22
Speaker 1 26:22

基于佳瑞这个讲的过程中，其实我我一直有个远景，就是说咱们刚刚说四个流程环节不是都被数字化、标签化和A化。就是句子在做的这个客服环节的一个数据。如果把它标签化，它的是非常有价值，甚至可以说是最有价值的一段数据金矿。也就是说如果我们把这一段的叫私域的数据梳理好的话。第一可以指导我的整个人群画像的标签增补，让我整个在投放环节，也就是内容分发环节的效率可以更高。第二通过跟思域的最后，我能找到消费者真实的内容的需求和痛点。所以可以指导我们在做内容创作，像我们这样的团队做更精准的能够贴合用户需求的这个内容引爆。举个例子，标题上说我们可以产生in一点多亿的播放，然后TK是六千多万，YT目前是四千多万，其实就是来源于私域。比如我们不是一上来第一条就达到了一个多亿的。
Building on what Jiarui just shared, I’ve always had a vision. As we mentioned, the four process stages are all being digitized, tagged, and AI-optimized. The data from the customer service segment, for instance, is incredibly valuable—arguably the most valuable data goldmine. If we properly organize this private domain data, first, it can enhance our audience profiling tags, making our content distribution more efficient. Second, by analyzing this private domain data, we can identify consumers’ genuine content needs and pain points, guiding our content creation. For example, teams like ours can produce more precise, user-centric content that resonates. Take the title’s claim of over 105 million views—60+ million on TikTok and 40+ million on YouTube—that all stems from private domain insights. We didn’t hit 100+ million views with the very first video.

发言人1 27:20
Speaker 1 27:20

其实我们做脚本的时候，防水电视我们最开始放到了这个客厅。因为防水比如说像户外，美国有很多大house，我们经常在户外去浇些水，就是狗狗这样跑也不会影响这个电视。然后我们就会发现评论里面就是来自于评论很多的评论说这个防水电视能不能放在这个浴室，就是我在洗澡的时候也可以看。然后这个需求其实是我们在做创作的时候，你是根本就想不到的。然后我们就回应他有大量这样的用户，那个时候其实我们就定制化的回应他。然后找到这部分的需求之后，我们开始做一些拍摄。发现这样的视频回应了消费者私欲的洞察，同时用新的视频传递的时候，引爆了和消费者的颅内高潮。其实这个就是一个非常好的私域反向指导公寓的例子。
Actually, when we were creating the script, we initially placed the waterproof TV in the living room. Because waterproofing is useful outdoors—many Americans have large houses, and we often water the yard—dogs running around wouldn’t affect the TV. Then we noticed a lot of comments asking if the waterproof TV could be placed in the bathroom, so they could watch it while showering. This was a demand we hadn’t even considered during the creative process. We responded to these users in a customized way. After identifying this demand, we started shooting some videos. We found that these videos resonated with consumers' hidden desires, and when we released new content, it triggered a strong emotional response. This is a great example of how private insights can guide public content.

发言人1 28:11
Speaker 1 28:11

那一样again公寓这边我们拍了大量的视频，其实这个视频的评论回复什么内容，比如说我把所有公寓的内容分成不同的多元的内容脚本。我就可以提前告诉我们家的这个句子，这个产品不同的内容对应应该是怎样的category的评论的组合。也就是说从消费旅程来看，不同的消费被分类了，每一个旅程都是personalized的experience。
Similarly, for public content, we shot a lot of videos. The comments and responses to these videos—for instance, I categorized all public content into different, diverse script types—allowed me to inform our team in advance about which product content should correspond to which category of comment combinations. In other words, from the consumer journey perspective, different consumers were segmented, and each journey became a personalized experience.

发言人1 28:39
Speaker 1 28:39

所以我们公寓的数据反哺了私域，私域的数据反哺了公寓。而整体全流程的数据进入到客户的CDP，也就是未来可能整个系统对于企业来讲需要两层架构。一层是基于他的大数据中台，上面可能是它的内容体系或者说它的标签体系在上面。再是全流程的应用，包含了洞察、生产、分发、私域，甚至于说未来的这个产品预测、供应链、库存。整个企业被有点像是agent化。其实就有点像现在这个sales force，其实把自己都改名字叫agent force。就是所有以前的这种CM套件全变成A型化。也是未来的一个趋势。所以我们很期待未来不光是用好句子，也是把句子的所有的数据和我们的内容标签的数据形成更大的集成和整合，然后反向赋能。
So the data from our apartments feeds back into the private domain, and the private domain data feeds back into the apartments. The entire end-to-end data flows into the customer's CDP, meaning the future system for enterprises might require a two-tier architecture. One layer is based on their big data platform, on top of which lies their content system or tagging system. Then there's the full-process application, encompassing insights, production, distribution, private domains, and even future product forecasting, supply chain, and inventory. The entire enterprise becomes somewhat "agentized." It’s similar to how Salesforce has rebranded itself as "Agent Force"—all the traditional CRM suites are turning into agent-based systems. This is a future trend. So we’re excited not just about leveraging good sentences but also integrating all the sentence data with our content tagging data to create greater synergy and then empower the system in reverse.

发言人4 29:32
Speaker 4 29:32

对，其实我们之前一直在聊到一件事情，包括我们和我们的客户聊，都是说公域种草私域成交，公域种草私域转化。其实最终可能大家希望把下面那个环补上。私域转化的同时，私域还能拿到洞察。然后我让公域更好的去投放，这样就会成为一个很好的飞轮。而这件事是需要通过数字化和智能化的手段去完成的那下一个话题，我们就聊一聊这个KA大客户lip在寻找PMF的过程中一头连的是产品和服务以及团队的迭代，而另一头连接的就是客户。但是大家不知道的可能是，real除了是一个营销专家，他也是一个top。最早我认识的时候他就是一个sales，然后也是我一直希望能把我拉到我的团队，我们一起来做一些更大的事情。无论之前是在国双还是在会亮，其实reo都服务了非常多头部的大客户，reo本人也发挥了非常显著的价值。
Yes, actually, we’ve been discussing this topic all along, including with our clients—the idea of "planting seeds in public domains and harvesting conversions in private domains." Ultimately, everyone hopes to close the loop. While conversions happen in private domains, insights can also be extracted from private domains. This allows public domains to optimize their targeting, creating a powerful flywheel effect. Achieving this requires digital and intelligent solutions. The next topic we’ll discuss is how Leap, a key account client, navigates PMF (Product-Market Fit). On one end, it connects to product and service iterations and team evolution, while on the other, it links to customers. What many might not know is that Will isn’t just a marketing expert—he’s also a top-tier salesperson. When I first met him, he was in sales, and I’ve always hoped to bring him onto my team to tackle bigger challenges together. Whether at Gridsum or Huiliang, Will has served many top-tier clients and delivered exceptional value.

发言人4 30:29
Speaker 4 30:29

我跟W一起去聊过，客户没有是一个能量场非常强的人。我见过很多的top sales风格都不太一样。没有其实是一种独具一格的风格，它有非常强的感染力，它能感染。那你要感染客户，包括感染和他一起去客户现场的人，包括当时也会感染我。虽然有些产品是我做的，但可能从他嘴里说出来可能会更富有想象力，激动人心。
I've had conversations with W before. The client is someone with an incredibly strong energy field. I've met many top salespeople, and their styles vary greatly. W has a unique style—it's highly infectious and can captivate others. This ability to captivate extends to clients, colleagues visiting clients together, and even me at the time. Although some products were developed by me, when W describes them, they sound more imaginative and exciting.

发言人4 30:53
Speaker 4 30:53

而且我觉得will的很多叙事方式其实很美式的那包括我们其实之前也聊过说to B2C的逻辑是不太一样的。很直接的体现就是to c不需要销售团队。那to b尤其是大的这种to大B，没有销售团队是很难实现规模化的营收的。因为to b的决策流程更复杂，客单价更高。包括你需要说服他，考验你的综合能力也会更强。
Moreover, I think Will's narrative style is very American. We've previously discussed how the logic of B2B and B2C differs significantly. A direct manifestation is that B2C doesn't require a sales team, whereas B2B—especially large-scale B2B—struggles to achieve scalable revenue without one. This is because B2B decision-making processes are more complex, with higher average order values. It also tests your comprehensive abilities more rigorously, as you need to persuade clients effectively.

发言人4 31:17
Speaker 4 31:17

我们知道lip其实更多服务的是全球化的消费品牌。但是哪怕你刚刚开始创业的时候，立刻就会有头部的大客户去跟你们合作。比如说名创优品，比如美的，他们愿意和你们共同探索合作模式和最佳实践。我相信除了所谓的客户资源，其实背后有更多更底层的东西。其实很多人聊到说大客户销售就是搞资源。其实从我的视角看到，其实不只是一个资源型的销售，其实更多还有一些东西需要去赢得大客户的信任。这也是我非常欣赏view的一点。所以也想请你在这里给大家分享一下。那你是如何在早期就能找到大客户，并且他们愿意跟你共创K的决策的流程是怎么样？
We know that Leap primarily serves global consumer brands. But even when you were just starting your business, top-tier clients immediately came to collaborate with you. For example, Miniso and Midea were willing to explore cooperation models and best practices with you. I believe that beyond so-called client resources, there are deeper underlying factors at play. Many people talk about enterprise sales as being all about leveraging resources. From my perspective, it’s not just resource-driven sales—there’s more to it, like earning the trust of major clients. This is something I truly admire about your approach. So, I’d like you to share with everyone here: How were you able to secure major clients early on, and what was the decision-making process like for them to co-create with you?

发言人1 32:02
Speaker 1 32:02

首先我觉得我早期找到大客户，比如说去年我找到大客户的时候，其实我并没有办法把大客户达成成交。甚至于说我非常好的大客户，最后是变成了我只是创业的一个adviser，但是确实没有发生交易，直到现在。所以我也有很多大客户去失败的时候，但是我喜欢看终局。其实我太太经常跟我说，说你怎么每次失败好像你都无所谓。
First, I think when I initially approached major clients—for example, last year when I reached out to them—I wasn’t actually able to close deals with them. In fact, some of the very promising clients ended up becoming just advisers for my startup, but no transactions materialized, even to this day. So, I’ve had my share of failures with major clients, but I like to focus on the endgame. My wife often tells me, "Why do you seem so unfazed every time you fail?"

发言人1 32:28
Speaker 1 32:28

因为我觉得我总会成功，或者说我们的team总会成功。包括嘉瑞刚有提到说希望我加入他的团队。现在在此时此刻还没有，但我觉得从中局的角度来讲，加入团队是一种合作形式。其实我们现在也是一种合作形式，我更倾向于的状态就是像水一样be water，我们是一种共赢的状态。所以通常在接触一个客户的时候，如果这次输掉了，我不会认为我们真的输掉了。我会认为是在此时此刻我们没有找到一个共赢的基础。但是在不远的未来我们可以找到，可能这样比较形而上学。
Because I believe I will always succeed, or rather, our team will always succeed. Just as Jiarui mentioned earlier about hoping I would join his team—at this moment, it hasn’t happened yet. But from a long-term perspective, joining a team is a form of collaboration. In fact, what we’re doing now is also a form of collaboration. The state I prefer is to be like water—fluid and mutually beneficial. So when engaging with a client, if we lose this time, I don’t consider it a real loss. I see it as a moment where we haven’t yet found a foundation for mutual benefit. But in the near future, we will find it. Maybe this sounds a bit metaphysical.

发言人1 33:02
Speaker 1 33:02

我说的比较直接一点，比如说我们做客户的时候，我会去思考他为什么需要我这样的一个新的服务商，那我就需要找到参照物。通常这个时候把KL的这个预算拿出来，发现头部客户在这个搜索上的花费是非常多的。我们就在思考，如果说我们想拿到一部分的KO的份额，我们的比较优势在哪里？所以整个行业你发现K6基本上是不太按照效果付费的。基本上就是说我的坑位多少，你寄个样品多少RY我是不保证的。
To put it more bluntly, when working with clients, I think about why they would need a new service provider like us. That’s when I need to find a reference point. Usually, we look at KL’s budget and realize top clients spend a lot on search. Then we ask ourselves: if we want to capture a share of KO’s market, where is our comparative advantage? Across the industry, you’ll notice K6 rarely pays based on performance. It’s mostly about how many slots you have, how many samples you send—RY is never guaranteed.

发言人1 33:35
Speaker 1 33:35

这个时候我们就第一性原理思考，有没有一种可能性是说我们提供一项搜索的服务，同时我们按效果结算，而不是像KO一样按照所谓的坑位结算。最早我们做举账号的时候是卖账号的，就是我帮你起一个账号，一个账号多少钱。但这样的话就跟KL没有办法形成差异化的优势。所以最后我们去定标一些大客户的时候，大客户就反过来问我们说整个搜索市场都没有一个按效果结算的方式。为什么你们按效果结算？然后这个时候，在投标结束之后，因为他找到我们的对手都找不到，他找了18家都没找到我们的对手。然后这个时候说我们也没有办法，因为我们只能按效果结算，我才有机会服务你。所以如果有容易的钱赚的话，我也想按照账号结算，我也想按照发帖量结算。
At this point, we thought from first principles: Is there a possibility that we could offer a search service while settling payments based on performance, rather than like KO, which settles based on so-called "slot positions"? When we first started building accounts, we sold accounts—meaning we’d help you set up an account for a fixed price. But this approach didn’t give us a differentiated advantage over KL. So when we later targeted major clients, they actually asked us why we were settling based on performance when no one else in the entire search market was doing so. After the bidding process ended, they couldn’t even find our competitors—they looked at 18 companies and still couldn’t find a match. At that point, we had no choice but to say, "We can only serve you if we settle based on performance. If there were easy money to be made, we’d also prefer to settle by account or by post volume."

发言人1 34:22
Speaker 1 34:22

但是因为我们作为一个新的玩家，我要进入这个市场，我如果follow原来的规则，我很难拿到份额。所以这个时候我们才跟团队说，如果我们要持续的拿到更大的social的budget的份额，可能我们的上限是30%。但我们能不能通过一种差异化的结算方式先拿到2%。
But as a new player trying to break into this market, if we followed the old rules, it would be very hard for us to gain market share. That’s why we told the team: If we want to consistently capture a larger share of the social budget—even if our ceiling is 30%—could we first secure 2% by adopting a differentiated payment model?

发言人1 34:40
Speaker 1 34:40

5%是因为这样的原因。从现在倒推回来看，最开始按CPMCPC甚至按照ross结算。Social是一个非常痛苦的过程。因为不是所有的产品拍个短视频就有订单，所以我们有些项目是亏钱的。但是这个过程中，我们没有改变这种结算逻辑。就是我团队经常跟我说，他说老板咱们能不能不要自己卷自己？我说我们必须卷自己，因为我们不卷自己，我们就会被卷死。
5% is due to this reason. Looking back now, initially settling by CPM, CPC, or even by Ross was a very painful process. Because not all products generate orders just by shooting a short video, some of our projects lost money. But throughout this process, we didn’t change this settlement logic. My team often tells me, "Boss, can we stop pushing ourselves so hard?" I say we must push ourselves, because if we don’t, we’ll be crushed by the competition.

发言人1 35:05
Speaker 1 35:05

所以当我们坚持说我就是按照效果结算的时候，我们开始发现了一个复利。就是账号的粉丝是越做越多的，爆款效率是越来越高的。也就是说虽然对赌CPM、CPC甚至CPS，我们的完成实现难度是越来越低的。反观KL你今年跟他合作的价格，比如说是一千或者两千美金。但是明年他可能会涨粉，找他人更多一条帖子就不是2000美金了，也就是确定性的。长周期来看，KL的CPM是越来越高，而QC或者说自营KOC或者自营的KOS，它的对应的CPMCPC是越来越低的。同时我们还帮品牌做资产。
So when we insisted on settling based on performance, we began to discover compounding benefits. The followers of the accounts kept growing, and the efficiency of creating hits kept improving. In other words, although betting on CPM, CPC, or even CPS, our ability to achieve these goals became increasingly easier. In contrast, KL’s collaboration price with you this year might be, say, $1,000 or $2,000. But next year, they might gain more followers, and the cost per post won’t be $2,000 anymore—it’s a certainty. Over the long term, KL’s CPM keeps rising, while QC, or self-operated KOCs or KOS, sees their corresponding CPM and CPC decreasing. At the same time, we also help brands build assets.

发言人1 35:47
Speaker 1 35:47

所以回到你说的大客户逻辑怎么切，我倒不觉得大客户完全是靠关系，其实还是看你提供的服务有没有找到一个USP。就是我在整个生态位为什么需要我？我创业的时候我经常会问自己，整个市场生态和供需关系里面究竟需不需要我们这一家新的公司，我们有没有存在的意义？我们存在在这个生态位上的差异化是什么？把这个问题想清楚，我觉得始终在大客户的过程中会找到一些小的机会。然后小的机会构建好了，完成好了，这也是我创业和我以前在大企业最大的差异。
So returning to your question about how to approach the logic of major clients, I don't believe major clients are entirely relationship-driven. It really comes down to whether the service you provide has identified a USP (Unique Selling Proposition). Why am I needed in this entire ecosystem? When I was starting my business, I often asked myself: Does the market ecosystem and supply-demand relationship actually need another new company like ours? Do we have a reason to exist? What is our differentiation in this niche? Once you clarify these questions, I believe you'll always find small opportunities in the process of engaging major clients. Then, by building and executing well on these small opportunities—this is precisely the biggest difference between entrepreneurship and my previous experience in large corporations.

发言人1 36:19
Speaker 1 36:19

在大企业我其实通常不用太过分的担忧，我的后盾就是我的交付，我只要把生意拿进来，在开会的时候，开管理会的时候，就是各种谈笑风生。你们这个交付的怎么样？资源不到位到底是什么情况？就是我们只要把责任扯清楚就可以了。
In a large corporation, I typically didn't have to worry excessively because I had the backing of delivery support. My job was simply to bring in the business, and during management meetings, it was all about smooth talk. "How's the delivery going?" "What's the issue with resource allocation?"—we just needed to clarify responsibilities and that was it.

发言人1 36:36
Speaker 1 36:36

但是在创业的时候，比如说你的大客户的需求过于大，你的团队是承接不住的。所以这里面就会变成除了拿客户的过程，还需要多一步是消化的系统的打造。就是我应该一口一口的吃。比如说我的终极目标是收入预算，我们要稳定吃掉30%甚至更多。但是现在团队的交付能力力和对于爆款的拆分理解能力，和对于跟像生态伙伴，包括具体在内的整个全流程的效率能力，还不支撑我们马上吃掉30%。所以我通常会跟客户画一个roadman，比如说q one我们预计的目标是我们做好你的5%。
But when starting a business, for example, if the demands of your major clients are too large, your team might not be able to handle them. So, in addition to acquiring clients, you also need to focus on building a system to digest these demands—meaning you should take it step by step. For instance, my ultimate goal might be a revenue target where we aim to steadily capture 30% or even more. However, the current team's delivery capacity, ability to break down viral hits, and efficiency in collaborating with ecosystem partners—including the entire end-to-end process—aren’t yet strong enough to immediately handle 30%. So, I usually outline a roadmap for clients, saying something like, "In Q1, our target is to deliver 5% for you."

发言人1 37:14
Speaker 1 37:14

这个时候客户说我可以给你更多的预算，我说我不要，我真的不要。因为我觉得如果你给我10%，我未必能达到超出一定预期的结果。我们能不能稳扎稳打？这样的话，我站在客户的视角，他在那内部也会有更大的reputation，也可以争取更多的资源。而且从终局来看，可能这个客户他会从这个大客客户跳到另外一个大客户，他还是我的客户。那我为什么要急于在这样一个项目上，一定要把我的赛做的足够大呢？所以这里面就是从大客户逻辑，这里面还是回到一个叫长期价值和短期价值。
At this point, the client might say, "I can give you more budget," but I’d respond, "No, I really don’t want it." Because I feel that if you give me 10%, I might not be able to deliver results that exceed expectations. Can we take it slow and steady? From the client’s perspective, this approach helps them build stronger internal reputation and secure more resources. And in the long run, even if this client moves from one major account to another, they’ll still be my client. So why rush to scale up too quickly on a single project? This brings us back to the logic of major clients—it’s about balancing long-term value versus short-term gains.

发言人1 37:47
Speaker 1 37:47

比如说佳瑞你是我的客户，但我一开始注定要跟佳瑞你在未来的三年，甚至未来的五年持续共赢的话。我其实不用急于说，我一定是立刻马上加入你的团队去实现共赢。它的供应方式非常多，甚至未来有跟我们公司层面发生更深的合作，这个是非常常见的一个逻辑。那你再回来看每一个单点赢和输，其实在长远的目标来看，它都是一个积累的过程。所以收回来就是说这个大客路径上来讲，传统的这个需要一定的关系，需要一定基础。
For example, Jiarui, if you're my client, and I'm determined to achieve mutual success with you over the next three or even five years, I don’t necessarily need to rush into immediately joining your team to realize that win-win. There are many ways to collaborate—even deeper cooperation at the company level in the future is a very common approach. When you look back at each individual win or loss, from a long-term perspective, it’s all part of an accumulation process. So, to summarize, when it comes to the path of securing major clients, the traditional approach requires certain relationships and foundational groundwork.

发言人1 38:20
Speaker 1 38:20

包括所谓的大家说创业公司第一天就要拿大客户。我觉得创业是一个我不能说是一个巨坑，是一个很大的坑。谨慎谨慎再谨慎。就是我不太建议说在大厂或者说大的企业里面做到比较好的位置，有一个稳健的收入。包括现在整个经济环境也是在慢慢恢复的过程。
Including the so-called notion that startups should land major clients from day one. I think entrepreneurship is—I wouldn’t call it a giant pitfall, but it’s certainly a significant one. Proceed with caution, caution, and more caution. I wouldn’t recommend leaving a stable, well-paying position at a large corporation or enterprise, especially given that the overall economic environment is still in a gradual recovery phase.

发言人1 38:38
Speaker 1 38:38

所以如果是对自身能力过于自信，我说的是过于自信，认为这些大客户都是你本人的关系的话，可能第一年会被啪啪打脸。我其实亲身就是这样的经历，我经历过一个从自信自负到自卑的一样的过程，但现在进入到一个我自己的心态。也比较像是一个脉冲，就相对比较稳定的一个股票在往上扬。就是它有上浮下浮，但那个上浮的阀值和下浮阀值慢慢变得有点control。所以大客户的获取其实也是这样的。
So if you're overly confident in your own abilities—I mean excessively confident—thinking these major clients are all due to your personal connections, you might get a harsh reality check in the first year. I’ve personally experienced this myself, going through a journey from confidence and arrogance to self-doubt, but now I’ve reached a mindset where it’s more like a pulse—a relatively stable stock trending upward. There are ups and downs, but the thresholds for those fluctuations are gradually becoming more controlled. Acquiring major clients is actually the same way.

发言人1 39:09
Speaker 1 39:09

我不太认为说大厂给到你的资源，一定等于你出来创业能拿到大客资源。然后不变的地方是找到你的核心大客户在整个生态位置为什么要选择你的原因。比如说我相信佳睿做巨子家公司，也是有非常多多形形色色的友商。你不会跟你的友商功能完全一样，你一定能找到一个点和一个解决方案是你独有的。我觉得找到那一个点非常小的点，哪怕牺牲利润率，甚至说只要在不亏损，甚至可以亏损一部分情况下，因为长线你是可以赢回来。我觉得如果从这个基础来看的话，整个大客户获取的话，整个状态会更轻松。如果一上来第一次的大客户的接触，就是既要又要还要既要拿到这个logo，又要赚利润，同时还需要。迅速规模化，有可能这个预期和现实会有非常大的改。
I don’t necessarily believe that the resources a big company provides you will automatically translate into landing major clients when you start your own business. What remains unchanged is identifying why your core clients, within the broader ecosystem, would choose you. For example, I believe Jia Rui, who works at Juzi Company, also faces a wide array of competitors. You won’t have exactly the same features as your competitors; there must be at least one unique point or solution that only you offer. I think finding that one small, distinctive point—even if it means sacrificing profit margins, or even operating at a slight loss—is worth it because, in the long run, you can recover. From this perspective, the process of acquiring major clients becomes much more manageable. If your very first interaction with a major client is about wanting everything—landing their logo, making a profit, and scaling rapidly—the gap between expectations and reality could be enormous.

发言人4 40:05
Speaker 4 40:05

对我觉得我也非常的幸运，见证了没有从大企业的高管出来创业。经历了自信、自负、自卑又重新站起来的过程。他各种现在还是跪着。
I think I'm also very fortunate to have witnessed someone transitioning from a corporate executive to entrepreneurship. He went through the process of confidence, arrogance, self-doubt, and then standing back up again. Even now, he's still on his knees in various ways.

发言人1 40:18
Speaker 1 40:18

现在还没有完全站起来。
He hasn't fully stood up yet.

发言人4 40:19
Speaker 4 40:19

内心已经完全站起来了。我觉得内心站起来很重要，我也非常欣赏。其实他有一种很强的创业者精神，就是我们说创业者打死的小强。不管我经历过什么，但是最后我还是会重新站起来，包括今年我也有一个特别深的感触。我觉得做人就做公司，就是在做人，做公司就是看你这个创业者怎么做人。其实很多人愿意帮助你这家公司，他是因为想帮助你这个的人。很多人愿意相信你的公司，相信你的公司哪怕今天出了一些问题，依然相信你，陪着你往下走，他也是相信你这个人。就像没有刚刚讲到的，最终其实大家要共赢，也不要着急，和大客户也是一样的。
The inner self has completely stood up. I think it's very important for the inner self to stand up, and I truly admire that. In fact, there's a strong entrepreneurial spirit here—what we call the "indomitable cockroach" of entrepreneurs. No matter what I've been through, I will always rise again. This year, I’ve had a particularly profound realization: running a company is essentially about being a person. How you run a company reflects how you conduct yourself as an entrepreneur. Many people are willing to help your company because they want to help you as a person. Many are willing to believe in your company, even if it hits a rough patch, because they believe in you and are willing to stand by you. As mentioned earlier, the ultimate goal is mutual success—there’s no need to rush. The same applies to dealing with major clients.

发言人4 40:55
Speaker 4 40:55

我今年的认知也是一个重新很大变化的一个过程。今年我经常跟我的销售团队，我们的销售同学，我说你先跟他交朋友，你跟这帮人先交朋友其实很重要的。这个交朋友不是说所谓的真的做关系，而是我们真正的用心去想他需要什么。今天你不需要我不代表着未来你不需要我，我持续是一个有价值的公司，我持续是一个有价值的人。未来在某一个时间点，你需要我的时候，咱们再在一起做业务，在一起做生意。因为本质上人生也是一个蛮长的一个互相协作。具体到的说好像我们公司一个业务做生意，但其实还有很多其他的比较泛化的一些生意，我们都需要一起去做的。
This year has been a significant shift in my perspective. I often tell our sales team: "First, make friends with them." Building genuine connections with people is crucial. This isn’t about superficial networking but truly understanding what they need. Even if they don’t need you today, it doesn’t mean they won’t in the future. As long as I remain a valuable company and a valuable individual, there will come a time when they need me, and we can collaborate on business again. Life, at its core, is a long journey of mutual cooperation. While it may seem like we’re just doing business, there are many broader opportunities we can explore together.

发言人4 41:35
Speaker 4 41:35

所以其实这种场主义，这种共赢的思维才是能做大客户很核心的一点。其实刚才我们听到没有讲到的有一个小细节也可以跟听众们分享一下。就没有一直在讲我们作为创业公司，我们一定要拿一些创新的点去和这大客户聊，才能挖掘出的价值。
So actually, this kind of win-win mindset is truly the core factor in handling major clients. There's a small detail we didn't mention earlier that I'd like to share with the audience. As a startup, we must always bring innovative points to the table when discussing with these major clients to uncover their true value.

发言人4 41:52
Speaker 4 41:52

其实view是我身边为数不多的，我认为一个非技术背景，把AI钻研到极致的人我刚才说的小细节这是什么呢？就是他之前给我发过，他跟complexity一聊聊好几天的聊天记录，他会不停的跟AI聊，包括到现在也是就是我身边有很多另外一拨人，他们不懂AI经常说我们要拥抱AI然后怎么去用报AI呢？就做不了实际，非常想拥抱AI但实际上就在口头上。但六是真的是所有这些工具他就不停的玩，并且真的让这些AI重塑自己的工作流的。很多公司思考都在和AI非常深度的去沟通交流。
Actually, Will is one of the few people I know who, without a technical background, has mastered AI to an extreme level. What was that small detail I mentioned earlier? He once sent me chat logs of him conversing with AI for days—he constantly engages with AI, even now. There's another group of people around me who don't understand AI but often say, "We must embrace AI." Yet when it comes to actually using AI, they struggle. They claim to embrace AI but only pay lip service. Will, however, truly experiments with all these tools and reshapes his workflow with AI. Many companies are now deeply integrating AI into their communication and operations.

发言人4 42:30
Speaker 4 42:30

其实因为我身边都是AI的朋友，我看到的很多都是很技术的同学去做这样的事儿。咱们真的是非常深度的去把创新，把新技术，包括在很多年前没有AI的时候，leo也一直在看最新的技术是什么。因为这些大企业，这些K他想要的新的价值，其实就是一些创新的技术。因为如果是个传统的东西，其实我们新人是进不去的。不管有什么样的关系，除非非常硬的，非常硬的可能就不来做这块了，其实是很难进去的。所以还是找到新的东西，对新的东西的价值的理解回归过来。我认为其实sas业务的核心就是价值创造，包括刚才我们聊的所有东西都是价值创造。
Actually, since most of my friends are in AI, I see many technical people doing this kind of work. We’ve really gone deep into innovation and new technologies, including Leo, who has been exploring the latest tech for years, even before AI became mainstream. These large enterprises and key clients are looking for new value, which often comes from innovative technologies. Traditional approaches are hard for newcomers to break into—unless you have extremely strong connections, and even then, it’s tough. So, the key is to focus on new things and understand their value. I believe the core of SaaS business is value creation, and everything we’ve discussed so far revolves around that.

发言人4 43:08
Speaker 4 43:08

顺着刚才的话题，我们就一起聊一聊这个客户价值。企业因为创造价值而存在。从to b的角度来说，企业核心创造的就是客户价值，只是创造的方式各有不同。目前市面上to b方向的AI加电商的企业，可以简单的来划分为降本和增效两种。聚焦在内容生成方向的公司更多的偏降本，因为改变了内容生产的方式，大幅降低了人力、时间、物料和沟通成本等等。而聚焦在流量和交易侧的公司可能更多的偏增收，通过重诉信息流、商品流和交易流，可以量化的为客户持续的增加收入。这样子咱们也聊到相关的话题。你提到区分市面上很多产品和技术导向的AI加电商的视频公司内部核心是抓住了消费者决策视频化这一趋势，是在内容生产端叠加AI的能力，同时通过爆款视频加精准投放，可以直接为零售品牌客户带来巨大的流量，并在一定程度上达到为客户增收的目的。
Continuing from the previous topic, let’s talk about customer value. Enterprises exist to create value. From a B2B perspective, the core value they create is for their customers, though the methods vary. Currently, AI-powered e-commerce companies in the B2B space can be broadly categorized into two types: cost reduction and revenue growth. Companies focused on content generation lean more toward cost reduction, as they change how content is produced, significantly cutting labor, time, material, and communication costs. On the other hand, companies focused on traffic and transactions tend to prioritize revenue growth by reshaping information flows, product flows, and transaction flows, quantifiably increasing income for clients. This ties into what we’ve been discussing. You mentioned that many product- and tech-driven AI-powered video companies in e-commerce have tapped into the trend of consumer decision-making shifting to video. They enhance content production with AI while leveraging viral videos and precise targeting to drive massive traffic for retail brands, ultimately helping them grow revenue.

发言人4 44:12
Speaker 4 44:12

全球著名的管理咨询大师拉姆查兰也在他的著作客户说如何真正为客户创造价值中说。其中他有一个观点说，为客户创造价值不能只停留在降低成本上，而且还要能够帮助客户增加收入。包括我也一直在思考，我们怎么去把业务做成GMB分成我创造价值，咱们一起分增长收益。那结合多年你与大客户接触的经验，能否详细的分享一下怎么看为客户降本和为客户创收的价值差异？
The globally renowned management consulting guru Ram Charan also stated in his book "What the Customer Wants You to Know: How Everybody Needs to Think Differently About Sales" that creating value for customers cannot be limited to just reducing costs but must also help them increase revenue. I've been pondering how to structure our business into a GMB (Gross Merchandise Business) model where we create value and share the growth benefits together. Based on your extensive experience working with major clients, could you elaborate on the value differences between cost reduction and revenue generation for customers?

发言人1 44:44
Speaker 1 44:44

OK我觉得这个降本和创收本身的价值是不矛盾的。因为我理想中的商业模式，当然是它既大批量的或者说最高效的降本，又实现了大规模的增。因为我的偶像是musk，所以我经常会分析tesla的这个商业模式。其实tesla最开始做model s的时候，它就是一个非常高昂的成本。直到他在上海建这个超级工厂，不然他就有可能破产，所以它就是极致的降本。然后极致的降本，这个车产生出来这个model 3的利润和规模化，又带来了极致的增收。所以我想通过这样一个简单例子，就是说降本和征收它其实是一个我觉得是一个相辅相成的一个点。回到这个AI这个赛道，或者说回到软件这个赛道，我觉得如果是还不具备征收能力的前提，下降本的规模化是一个很好的生存位。
OK, I believe cost reduction and revenue generation are not mutually exclusive in terms of value. In my ideal business model, it would achieve both large-scale or highly efficient cost reduction and substantial revenue growth simultaneously. My idol is Musk, so I often analyze Tesla's business model. When Tesla first launched the Model S, it came with a very high cost. It wasn't until they built the Shanghai Gigafactory that they avoided potential bankruptcy—this exemplifies extreme cost reduction. The extreme cost reduction enabled the Model 3 to achieve profitability and scale, which in turn drove extreme revenue growth. Through this simple example, I want to illustrate that cost reduction and revenue generation are complementary. Returning to the AI sector or the software industry, I think if a company isn’t yet capable of revenue generation, scaling cost reduction is a strong survival strategy.

发言人1 45:35
Speaker 1 45:35

举个例子，比如说降本，我们通常是会把价值链拆解到每一个细分的端。举例单独去替代人工去做客服，替代人工去做简单的一些视频的编辑，就是所谓的混剪。简单的帮助人工去做一些消费的洞察的提炼简单的洞察提炼，这些都是属于降本，还没有做到征收。因为当你把这些内容做完了之后，这个内容并不直接带来转化。但是其实我一直有个设想，如果把多个降本的解决方案或者把多个降本的公司形成一个征收的有机体，这个解决方案就是征收的那现在的市场上，大量的AI公司，它是在整个价值链在征收价值链的其中一环。举个例子，AI的客服软件它是很难征收。你很难从客服里面挖到说有你可以挖到一些潜在的机会。但是它的投产比在以征收角度来看，其实是并不划算的。
For example, when we talk about cost reduction, we typically break down the value chain into every detailed segment. Take replacing human labor for customer service or simple video editing—what we call "remixing"—as an example. Or assisting humans in extracting basic consumer insights—these all fall under cost reduction, not yet revenue generation. Because after completing these tasks, the content doesn’t directly drive conversions. But I’ve always had this idea: if multiple cost-reduction solutions or multiple cost-reduction companies could form an organic entity for revenue generation, that would be the solution. Right now, most AI companies in the market operate within just one segment of the revenue-generating value chain. For instance, AI customer service software struggles to generate revenue. It’s hard to mine potential opportunities from customer service alone. From a revenue-generation perspective, the return on investment isn’t really worthwhile.

发言人1 46:30
Speaker 1 46:30

但我们想象一下，如果把leap的解决方案加上嘉瑞的解决方案，具体的解决方案变成一个整体total solution。我就是征收的。然后这个征收的过程中，句子应该拿到这部分收益的百分比。这也是我今年非常早期的时候，跟佳瑞讨论的诡异模式。
But imagine if we combined Leap’s solution with Jiarui’s solution into a total solution. That would be revenue-generating. And in this process, the sentence should capture a percentage of the revenue. This is the unconventional model I discussed with Jiarui very early this year.

发言人1 46:47
Speaker 1 46:47

但是最后的发起方是我们这样的一个叫流量方，因为我流量方我是更容易拿到征收。假设我没有了具体的这个环节，我能不能征收？我可以，我征收也很累。那把句子单独拎出来去切征收可不可以？也可以，商业模型也很累，所以我们两个降本，你想我的CPN比K要便宜，所以我也在降本，对不对？但你看看我们两个小的点结合成一个有机一个total solution，一个我们闭环的agent force，我们就变成一个征收的solution。那会不会我们把整个价值链里面更多的句子，更多的力放在这个偷偷舍入选，它是一个更加的征收收入税。
But ultimately, the initiator is what we call the traffic side, because as the traffic side, it's easier for us to collect taxes. Suppose I don't have this specific link—can I still collect taxes? Yes, but it would be very laborious. Can we extract individual sentences to collect taxes separately? That's also possible, but the business model would be cumbersome. So both of us are reducing costs—my CPN is cheaper than K's, so I'm also cutting costs, right? But if we combine these two small points into an organic, total solution, a closed-loop agent force, we become a tax collection solution. Could we then allocate more sentences and effort within the entire value chain to this "stealth selection"? It would be a more efficient tax collection revenue model.

发言人1 47:27
Speaker 1 47:27

然后中国的客户和美国的客户也非常有意思。因为我服务过中国的品牌，也服务过美国本土品牌。当然在这里也感谢一些早期帮助的贵人，不管是投资圈朋友还是一些生态伙伴，也介绍一些美国本土包括欧洲本土的客户给我。我发现欧美的客户，就是我们的老外的朋友，通常是非常关注过程指标。就是他知道你会有结果指标，但是他不会去追求苛刻。我一定要完成那个征收的RY，他会说这个过程你做好了，我是愿意付费的，因为别的过程我也能做好。
Chinese and American clients are also very interesting. I’ve served both Chinese brands and local American brands. Of course, I’d like to thank some early supporters—whether they were friends in the investment circle or ecosystem partners—who introduced me to local clients in the U.S. and even Europe. I’ve noticed that our Western clients are usually very focused on process metrics. They know you’ll deliver outcome metrics, but they won’t be overly demanding. They won’t insist on hitting a specific tax collection RY. Instead, they’ll say, "If you handle the process well, I’m willing to pay, because I can manage the rest of the process myself."

发言人1 47:58
Speaker 1 47:58

中国的这个跨境电商，有相当一大部分它其实是工贸的转型。所以工厂思维是我要出货，我已经花了那么多的钱开模打造产品，现在我还要听单点解决方案。你说你解决我的客服，你是我更好的种草，我就是要出货。所以这个时候就不是说这些老板我没有预算，是他希望你把链条打完，告诉他他给你多少钱，他可以相对有一个出货的确定性。
A significant portion of China's cross-border e-commerce actually stems from the transformation of industrial and trade businesses. Factory owners operate with a production mindset—they want to ship goods. Having already invested heavily in mold development and product creation, they now face piecemeal solutions. Whether it's improving customer service or better product seeding, their priority remains shipping products. So, it's not that these owners lack budgets; rather, they want a complete solution that guarantees sales in exchange for their investment.

发言人1 48:23
Speaker 1 48:23

他有没有预算呢？有，但他的预算能不能猜到具体的客我？这中间有一点gap。所以把句子的客服加上live的短视频，甚至再加上比如说某家公司做洞察AI的方式，变成一个有机体。工厂老板就会在一个确定性的征收情况下，把预算拿出来。然后我们几家价值的共创方再去分配对应这个共赢的利益的百分比，那就是一个很好的征收方案了。
Do they have budgets? Yes. But can they predict specific customer acquisition costs? There's a gap here. By integrating customer service with live short videos, and adding, say, an AI-driven insights solution from a certain company, we create an organic system. Factory owners will allocate budgets when they see a clear path to sales. Then, the collaborating value partners can distribute the shared benefits proportionally—that’s an ideal solution.

发言人1 48:52
Speaker 1 48:52

所以我想通过这样的一个例子说明，就是说降本和征收它其实并不是一个割裂的。只是说中国的目前的跨境这个形态，因为用跨境这个词其实本身就已经有单维思维。我们在美国我经常会问美国，我说你们做品牌，你们会做跨境电商吗？他说我们不知道跨境是什么意思。因为在美国创立的公司，通常第一天day one就是他会先用英语系。
So I want to use this example to illustrate that cost reduction and revenue growth are not actually separate concepts. It's just that China's current cross-border model—using the term "cross-border" itself already reflects a one-dimensional mindset. When I'm in the U.S., I often ask Americans, "Do you do branding? Do you engage in cross-border e-commerce?" They respond, "We don't know what 'cross-border' means." Because companies founded in the U.S. typically operate in English-speaking markets from day one.

发言人1 49:17
Speaker 1 49:17

比如说五眼联盟，包括欧洲，它本身就是在全球化，它不存在跨的这个过程。然后我们中国自古以来，因为我喜欢研究历史，我们自古以来是一个中文体系。所以我觉得去东南亚，去亚洲市场，当然我们拿掉可能日韩会相对有一点干。去东亚这个市场，其实我也觉得不叫跨境，它其实是一个大东亚共荣圈，不能这么说，就是它是一个整体。但是跨文化系的话，其实我们是有一些barrier的。所以我觉得就是站在宏观层面上来看出海这件事情，我觉得这一开始就给自己打标签了。
For example, the Five Eyes alliance, including Europe, is inherently globalized—there's no "crossing" process involved. Historically, China has always been part of the Chinese-language system—I enjoy studying history. So when we expand into Southeast Asia or other Asian markets (excluding Japan and South Korea, which might feel slightly disconnected), I don't even consider it "cross-border." It's more like a Greater East Asia Co-Prosperity Sphere—no, that's not the right term—it's more of an integrated whole. However, there are indeed some cultural barriers when crossing language systems. From a macro perspective, I think labeling it as "going global" from the outset is somewhat self-limiting.

发言人1 49:50
Speaker 1 49:50

不如把这件事情变成是说我只是有一个中国非常强的商业效率，然后scale到全球化。比如说抖音就是这样的，抖音中国做的很好，所以海外也做的很好。拼多多中国做的很好，所以题目在海外做的也很好。所以拉回来看这个增收和降本，我不觉得它是一个割裂的形态。就像我说出海和全球化，我不认为它是两件事情，它只是思维形式的一个再次升级。我们为什么要预设自己一定是跨呢？
Why not frame it as leveraging China's strong commercial efficiency and scaling it globally? Take TikTok for example—it performed exceptionally well in China, so it also succeeded overseas. The same goes for Pinduoduo; its success in China translated well internationally. Refocusing on revenue growth and cost reduction, I don't see them as separate issues. Just like how I view overseas expansion and globalization—they aren’t two distinct things but rather an evolution in mindset. Why should we assume we must "cross" anything in the first place?

发言人4 50:18
Speaker 4 50:18

对不对？我觉得刚才有一段说的特别好，其实相当于是我们把很多的样本的链条连起来，其实可能就会重新生成一个增收的新业务。
Right? I think what was just said is particularly insightful. Essentially, by connecting many sample chains, we might actually generate a new revenue stream.

发言人3 50:28
Speaker 3 50:28

对OK下面我们聊一聊出海营销市场，数据显示2023年A股市场具备出海业务的公司超过3500家，占比超过66%。海外收入总额超8.5万亿，占总营收的比重超过11.5%。说明了越来越多的中国企业头部的开始加速的布局海外的业务。另外也有数据表明，2023年中国珠海营销市场的规模约为344亿美元，年复合增长率为23%点。
Okay, let's now discuss the overseas marketing market. Data shows that in 2023, there were over 3,500 companies listed on the A-share market with overseas operations, accounting for more than 66% of the total. Their total overseas revenue exceeded 8.5 trillion yuan, representing over 11.5% of their total revenue. This indicates that an increasing number of leading Chinese companies are accelerating their overseas business expansion. Additionally, data reveals that the size of China's overseas marketing market in 2023 was approximately $34.4 billion, with an annual compound growth rate of 23%.

发言人3 51:00
Speaker 3 51:00

本土企业出海规模的增加引发出海营销市场的发展，同时本土营销服务的企业也在伴随客户出海的过程中，实现了自身出海全球化的进阶。上次我们也聊到就是荔浦在为客户进行海外数字营销服务的过程中，发现自身所具备的从中国市场卷出来的成本优势和综合能力，几乎是可以秒杀海外。当然我们说主要是欧美市场的一些agency。那能否展开聊一聊，就是企业出海在营销方面的痛点，你发现的是什么？以及为什么说中国背景的agency可以秒杀海外？
The growing scale of domestic companies going global has spurred the development of the overseas marketing market. At the same time, local marketing service providers have also advanced their own globalization by accompanying their clients overseas. Last time, we also mentioned that Leap, while providing overseas digital marketing services for clients, discovered that the cost advantages and comprehensive capabilities honed in the Chinese market could easily outperform those abroad—primarily referring to agencies in markets like Europe and the U.S. Could you elaborate on the pain points you've observed for companies going global in terms of marketing? And why do you say that agencies with a Chinese background can outperform their overseas counterparts?

发言人1 51:36
Speaker 1 51:36

两个维度出发。首先第一就是拼多多在海外其实挺成功的，包括虚印在海外也挺成功的。其实他们都发挥了中国制造的这个大规模的优势。所以当我们放到营销圈，其实我觉得整个价值链的制造，商品制造和我们说内容生产，我认为它其实是一个整体，甚至它的这个原理也是相似的。
Two dimensions to consider. First, Pinduoduo has been quite successful overseas, including SHEIN's notable achievements abroad. Essentially, they've leveraged China's manufacturing advantage at scale. When we apply this to the marketing sphere, I believe the entire value chain—product manufacturing and what we call content production—is actually an integrated whole, even sharing similar underlying principles.

发言人1 51:58
Speaker 1 51:58

我们举个例子来说，比如说都在这个产品的环节，生意的本质的价值链就是产品的供应链和这个内容供应链。比如说产品无非就是采购、生产、仓储，然后运输、物流，然后再到消费者到家。但如果我们放到营销的角度，其实内容也是一样的。从内容的策略到生产、到管理，到分发，到被消费者看到。所以既然虚拟和team，它可以发挥中国的商品制造的成本优势，那是不是我们可以发挥中国内容制造的成本优势？
Let me give an example. Take the product cycle: the core value chain in business consists of product supply chain and content supply chain. Products undergo procurement, production, warehousing, transportation/logistics, and finally reach consumers' homes. From a marketing perspective, content follows the same pattern—from content strategy to production, management, distribution, and ultimately consumer engagement. If virtual platforms and teams can capitalize on China's cost advantages in goods manufacturing, shouldn't we similarly leverage China's cost advantages in content creation?

发言人1 52:28
Speaker 1 52:28

回到这个点，为什么可以秒杀？我觉得是在局部的业务场景秒杀，不是全局秒杀。就是当中国的留学生，我说的是外籍留学生，因为中国有非常多一带一路国家，在不管在北京还是广州，广州是全中国大学生最多的地方。因为我们团队在广州，然后我们的公司放在大学城附近，所以它支撑了我们大批量拍摄外模的能力。因为外模它其实单位价格是挺高的，1个小时要几千块。但是如果是外籍留学生实习的方式，通常这个价格是相对可控的。
Returning to this point, why can it outperform? I think it outperforms in specific business scenarios, not universally. Take Chinese international students—I'm referring to foreign students—because China has many from Belt and Road countries, whether in Beijing or Guangzhou. Guangzhou has the highest number of university students in China. Since our team is based in Guangzhou and our company is located near the university town, it supports our ability to shoot large volumes of foreign models. Foreign models are actually quite expensive, costing thousands per hour. But if we use foreign students as interns, the price is relatively more manageable.

发言人1 53:00
Speaker 1 53:00

所以因为中国在一带一路国家整体的吸引，所以外籍留学生的这个我们叫大规模的储备。全中国范围，他其实为我们具备了这很好的优势。如果是在东南亚国家或者在欧美发达国家，首先没有这样的一个相对大规模的人才储备。第二也没有大规模的我们说成本的一个优势。所以这个第一部分。然后第二部分就是的确中国的engineer非常多，所以在AI技术的部分的软件的支持，包括软件之间的内卷的这个卷度的竞争，也会让在应用侧的内容，不管是生产分发还是洞察，会更早的看到一些可落地的结果。
So, due to China's overall appeal to Belt and Road countries, we have what we call a large-scale reserve of foreign students. Nationwide, this gives us a significant advantage. In Southeast Asian countries or developed Western nations, first, there isn’t such a relatively large talent pool. Second, there’s no large-scale cost advantage either. That’s the first part. The second part is that China indeed has a vast number of engineers, so the support for AI technology in software, including the intense competition among software companies, allows us to see practical results earlier in content applications—whether in production, distribution, or insights.

发言人1 53:39
Speaker 1 53:39

相比欧美的话，我看到欧美更多的希望讲一些大趋势。但你说真的落地到说提高了多少LI降低了多少CBC的实际落地case，我认为是落后于中国的在这个点上。所以AI技术到偏落地的应用车，我说的不是偏大模型和这个基建车落地的应用车，我认为中国是走在前面的。因为我们有太多的比较有挑战的场景需要被落地，而且老板们都挺着急的那第三个部分就是我们说除了这个视频的大规模的外籍学生的拍摄，包括AI的技术，其实就是我们说整个链条的数据的的分析能力。
Compared to Europe and the US, I notice that Western markets tend to focus more on discussing macro trends. But when it comes to practical implementation—like quantifying how much LI (Lead Increase) was improved or CBC (Cost per Business Conversion) was reduced—I believe China is ahead in this aspect. Regarding AI technology applications that lean toward practical implementation (I'm not referring to large models or infrastructure-level applications), I think China is leading. This is because we have numerous challenging scenarios that require implementation, and business leaders are quite eager to act. The third aspect is what we call the large-scale production involving foreign students for videos, along with AI technology, which essentially refers to the analytical capabilities across the entire data chain.

发言人1 54:15
Speaker 1 54:15

比如说我在国内做这件事情，我能找到我的上下游，我能找到句子，我能找到这个洞察，甚至我能找到这个舆情，找到更多维度的公司，成为一个价值链。但可能在美国做这件事情，你会发现它的门槛其实挺高的。因为你要整合多套系统，首先系统本身就不便宜。中国的这些创业企业，它其实可以接受play by performance，就是整个大的结果。你跟各个创始人聊一下，其实大家是能够谈成一个有点像区块链式的这种分配模式。也就是说我们的拆解重组的优化的方式会非常高小美国是更希望是说我把一个产品做的比较完整了，然后我以产品POG的这样的一个方式，它很难被重构。所以由这三个的环节，就是我认为中国是具备优势，就是大规模的外籍学生的生产，大规模应用偏应用侧的AI技术。
For example, when doing this domestically in China, I can identify my upstream and downstream partners, access sentence-level data, uncover insights, and even monitor public sentiment—connecting with multi-dimensional companies to form a value chain. However, doing the same in the U.S., you'd find the barriers are quite high. Integrating multiple systems is required, and the systems themselves aren't cheap. Chinese startups are more willing to adopt a performance-based approach—focusing on the overall outcome. If you talk to various founders, you'll find they can agree on a distribution model somewhat akin to blockchain. This means our methods for deconstruction, reorganization, and optimization are highly flexible. In contrast, the U.S. prefers to develop a more complete product upfront and market it as a Product-Oriented Growth (POG) solution, which is harder to reconfigure. These three factors—large-scale foreign student production, widespread application-focused AI technology, and flexible optimization approaches—are where I believe China holds an advantage.

发言人1 55:07
Speaker 1 55:07

比如说我可以说一个非常实际的例子，我们用国内的维度这样的去生成视频的效率，其实是比皮卡。皮卡是硅谷的一家公司，我们都用过是效率高很多的。还有包括成本，其实核心就是成本。然后再到我们说价值链的这个解决方案式的组合，这个效率我觉得都是非常高的那落地到这个应用场景上，你就会发现，基本上在做大规模批量的中等质量的视频上面，我们的价格和我们的CPMCPC的效率的优势大概会是美国的10到20倍。但是在这个高质量的，比如说偏TVC的，偏品宣的，偏要非常深刻的了解消费洞察的。还是要去比如说洛杉矶，去纽约，去巴黎当去拍摄。所以我们说在局部上，就在中等质量大规模的内容生产分发和效果营销方式的这样的一个细分赛道。
For example, I can give a very practical example. From a domestic perspective, our efficiency in generating videos is actually much higher than Pika. Pika is a Silicon Valley company, and we've all used it. This includes cost, which is essentially the core factor. Then when we talk about the value chain's solution-oriented combination, the efficiency is remarkably high. When applied to real-world scenarios, you'll find that for large-scale, medium-quality video production, our pricing and our CPM/CPC efficiency advantages are roughly 10 to 20 times better than those in the U.S. However, for high-quality content—such as TVC-style, brand-focused, or deeply consumer-insight-driven productions—you still need to go to places like Los Angeles, New York, or Paris for filming. So, in a specific niche—medium-quality, large-scale content production, distribution, and performance marketing—we have a clear edge.

发言人1 56:03
Speaker 1 56:03

我觉得中国乃至整个亚洲，它其实是具备非常大的优势。所以从医院、警察来讲，我也希望说这套大规模的优势可以先从中国出发，慢慢释放整个东亚的红利。因为除了中国具备大规模优势，其实还有两个地方，我觉得大规模优势是挺明显的。一个是印度，当然中国创业者很难进去。然后还有一个，其实是像东南亚的印尼，它的人口也是非常多的。印尼加菲律宾，菲律宾的英语很好。所以其实东南亚可能会是中国的创业者如果想全球化的下下一个供给侧，然后再往后可能是东欧、南美，当然可能会有时间文化的代差。所以我觉得整体东亚这件事情在做这个大规模中等质量的内容分发和效果营销，是具备绝对的降维打击的优势。
I believe China, and indeed all of Asia, holds a significant advantage. From a strategic standpoint, I hope this large-scale advantage can start in China and gradually extend to the broader East Asian region. Beyond China, there are two other regions where this large-scale advantage is quite evident. One is India, though it's challenging for Chinese entrepreneurs to enter. The other is Southeast Asia, particularly Indonesia, which has a massive population. Indonesia plus the Philippines—where English proficiency is high—could be the next supply-side opportunity for Chinese entrepreneurs looking to globalize. After that, Eastern Europe and South America might follow, though there could be cultural and temporal gaps. Overall, East Asia has an absolute advantage in large-scale, medium-quality content distribution and performance marketing, creating a "dimensional reduction" competitive edge.

发言人3 56:52
Speaker 3 56:52

第一个问题，就是企业出海在营销方面的痛点。
The first issue is the pain points enterprises face in marketing when going global.

发言人1 56:56
Speaker 1 56:56

这个我刚刚也说了，其实现在的企业在出海的营销，某种程度上宏观痛点都是要出货，都是要追求更好的权威和RN这是他们的显性痛点。就是他们只会用一句话去概括他们所有的痛点，隐性痛点就是过程痛点。比如说有没有更高的流量获取的效率，有没有更快的敏捷的帮助产品创新和开发的方式，有没有更快的去回馈，包括进行私域管理，管好所有的这个消费流程的这样的一个方式。所以就是企业的痛点。
As I mentioned earlier, currently for enterprises expanding overseas, their macro-level pain points in marketing largely revolve around sales volume - they all seek better authority and ROI (this being their explicit pain point). They tend to summarize all their pain points in one sentence. The implicit pain points relate to process challenges: whether there's higher efficiency in traffic acquisition, whether there are more agile methods to assist product innovation and development, whether there are faster feedback mechanisms, including private domain management - effectively managing the entire consumer journey. These constitute the enterprises' pain points.

发言人1 57:31
Speaker 1 57:31

如果从宏观拆解，无非就是我要生意的大的增长。老板一号位关注的是大的增长，下面的职能部门就是围绕他的过程的痛点去出发，有这个生产端，有消费，有这个组织端。所以如果从整体的除外来看，如果咱们以偏概全只看一号位的话，我觉得终归还是回到了生意的增长。也就是说在中美这个博弈的大背景下，包括这个关税的不确定性，其实大家还是有一种强烈的焦虑感，还是希望说拿到更大的出货的这种确定性。包括说我做更好的内容，我做更好的品宣，其实最终的目标也是为了更好的有生意的增长。当然也有一说品牌长期资产的部分，但是如果非要排一个优先级，我相信相当大一部分还是用两个比较简单的词汇，就是出货。
If we break it down from a macro perspective, it ultimately comes down to achieving significant business growth. The top priority for the CEO is substantial growth, while functional departments focus on addressing pain points in the process—whether in production, consumption, or organizational aspects. So if we look at the overall picture but narrow our focus solely to the CEO's perspective, I believe it still circles back to business growth. That is to say, against the backdrop of U.S.-China tensions, including tariff uncertainties, there remains a strong sense of anxiety, with everyone hoping for greater certainty in shipments. Even efforts like creating better content or enhancing brand communication ultimately aim to drive better business growth. Of course, there’s also the aspect of long-term brand equity, but if we had to prioritize, I believe a significant portion would boil down to two simple words: shipments.

发言人3 58:23
Speaker 3 58:23

OK. 我们接下来就聊一聊AI，上次咱们聊的你给lip的定位是AI驱动的全球数字导购。运用AI的总结推理生成个性化的能力，结合不同客户的需求场景，以爆款视频矩阵账号为切入点，达成为客户提升品牌影响力和销售额的目标。
OK. Let’s now shift our discussion to AI. Last time we spoke, you positioned Leap as an AI-driven global digital shopping guide. Leveraging AI’s capabilities in summarization, reasoning, and generating personalized content, combined with different customer needs and scenarios, Leap aims to use viral video matrix accounts as an entry point to help clients enhance brand influence and sales.

发言人3 58:43
Speaker 3 58:43

其实AI导购的说法，在这轮AI浪潮中也有很多其他玩家在尝试。首先就是电商平台，因为场景非常吻合，电商平台拥有庞大的用户基数、消费数据、商品内容和商品供给叠加AI后，可以提供给消费者更加智能的购物体验，甚至重塑电商购物的流程。比如说亚马逊在今年的10月9号，我们这个博客发布大概会在25年，但是我们录制时间还是在24年的年底。亚马逊在今年的10月9号就在AWS一年一度的reinvent技术大会上推出了全新的AI导购通过。通过文献详尽的商品信息，总结商品特点和买家反馈，记录顾客购物偏好以及美化UI界面等方式，旨在为消费者提供更加精准高效的购物体验。同时借助先进的AI技术，AI导购能够针对消费者的细致的特定的需求，快速筛选出合适的产品，让购物流程更加便捷。
Actually, the concept of AI shopping assistants has been explored by many other players during this wave of AI advancements. First are e-commerce platforms, as the scenario fits perfectly. With their massive user base, consumer data, product content, and inventory, e-commerce platforms can leverage AI to provide consumers with smarter shopping experiences and even reshape the e-commerce purchasing process. For example, Amazon—though our blog will likely be published in 2025, we're recording this at the end of 2024—launched a new AI shopping assistant at its annual AWS re:Invent conference on October 9 this year. By leveraging detailed product information, summarizing product features and buyer feedback, tracking customer preferences, and enhancing UI design, it aims to deliver a more precise and efficient shopping experience. Additionally, powered by advanced AI, the assistant can quickly filter suitable products based on consumers' specific and nuanced needs, making the shopping process more convenient.

发言人3 59:42
Speaker 3 59:42

类似国内的抖音电商近期也正式上线了AI导购的服务，叫智能购物。基于豆包大模型算法，入口就放在抖音商城APP的个人中心里面，点开就可以享受免费的AI导购服务。根据官方的介绍，目前智能购物主要有两大功能，一是能够提供商品咨询以及商品对比的服务，第二是可以帮用户选礼物。
Similarly, Douyin E-commerce in China recently officially rolled out its AI shopping assistant service, called Smart Shopping. Based on the Doubao large model algorithm, the feature is accessible via the personal center in the Douyin Mall app, where users can tap to enjoy free AI shopping assistance. According to official introductions, Smart Shopping currently offers two main functions: providing product inquiries and comparison services, and helping users select gifts.

发言人3 01:00:06
Speaker 3 01:00:06

除了电商平台，有些大模型厂商为了整合生态资源，切入更多的场景，也有类似方向尝试。比如说kimi它就有一个kimi加什么值得买。Kimi官方定义为超强导购，当然这种方式很轻了，但也从侧面证明AI导购是一个值得探索的方向。说回leap，想请你分享一下为什么定位leap是AI数字导购。那AI在其中的核心价值是什么？
Beyond e-commerce platforms, some major model manufacturers have also made similar attempts to integrate ecosystem resources and tap into more scenarios. For example, Kimi has a feature called "Kimi + What's Worth Buying." Officially defined as a super shopping guide, this approach is relatively lightweight, but it also demonstrates from another angle that AI shopping guidance is a direction worth exploring. Coming back to Leap, could you share why you positioned Leap as an AI digital shopping guide? What is the core value of AI in this context?

发言人1 01:00:32
Speaker 1 01:00:32

导购应该是我们今年早期的一个定位。其实现在来看，我觉得导购把我们的做的事情概括了一部分，但不全部。首先说导购的部分是这样的，就是当我们做起了大量的举证账号的时候，这些举证账号就像一个活生生的一个导购账号。就是我们经常说国内做家居卫浴的，其实都会有很多导购去开通小导出账号。我们帮这个出海企业做的其实有点像是他的海外的导购。因为他们通常没有线下的门店。所以我们在想说，当他们没有线下门店的时候，而美国的消费者也想向中国的消费者找到导购的时候，他们在社交媒体上是不是能找到这样的账号。所以我们起号的过程中通常会有三步。
The shopping guide concept was our early positioning this year. Looking back now, I feel that "shopping guide" captures part of what we do, but not all of it. First, regarding the shopping guide aspect: when we built a large number of matrix accounts, these accounts functioned like lively shopping guide accounts. For instance, in China, many home and bathroom brands have shopping guides who operate small accounts. What we’re doing for overseas enterprises is somewhat akin to being their international shopping guides, as they usually don’t have physical stores overseas. So, we wondered: when these brands lack physical stores and American consumers, like Chinese consumers, seek shopping guidance, could they find such accounts on social media? Therefore, our account-building process typically involves three steps.

发言人1 01:01:13
Speaker 1 01:01:13

第一步是把这个账号人设像是一个jack mary Peter这样的一个品牌的导购号，所以我们说他是导购。当粉丝量做的比较大的时候，我们有一部分的导购会升级变成官号的区域号。比如说这个品牌的germany品牌的，就是因为我已经孵化成功了，没有那么孵化成功的一般继续维持这个导购的身份，就是他在社媒上的定位，更像是一个导购。
The first step is to position this account as a brand shopping guide, like Jack, Mary, or Peter, which is why we refer to it as a shopping guide. When the follower count grows significantly, some of these shopping guides will be upgraded to become regional accounts for the official brand. For example, the Germany account for this brand has already been successfully incubated. Those that haven’t achieved the same level of success will continue to maintain their shopping guide identity, meaning their social media positioning remains more like a shopping guide.

发言人1 01:01:35
Speaker 1 01:01:35

但是为什么说我们导购我觉得没有完全囊括呢？是因为当我在做导购的过程中，我拿到了爆款所谓的爆款公式，或者说爆款的大量的优质的内容，它可以被全处理的放大。包括我们通过爆款视频直接去做投放，包括把爆款视频放到落地页，提升落地的转化率。这两个动作包括把爆款实力放在举个例子，像句子做的这个私域对话里面，是不是可以把我们的爆款视频变成一个对话的一个知识库的一部分。像这些动作其实它不能完全用导购来概括。也就是我的内核是我的爆款视频。我的出发点是一堆社媒账号，像导购一样。但是我的这个未来的场景是全触点的爆款视频。
But why do I feel the term "shopping guide" doesn’t fully encompass it? Because during the process of being a shopping guide, I’ve discovered what we call the "viral formula" or a large volume of high-quality content that can be amplified across all channels. This includes directly using viral videos for ad placements and incorporating them into landing pages to boost conversion rates. These actions—such as integrating viral content into private domain conversations, like what Juzi is doing—can turn our viral videos into part of a conversational knowledge base. Such moves can’t be entirely summarized as just "shopping guide." In essence, my core is the viral video. My starting point is a cluster of social media accounts that function like shopping guides, but my future vision is viral videos that engage across all touchpoints.

发言人1 01:02:20
Speaker 1 01:02:20

也就是说其实我们蛮希望我们公司围绕着消费的生态，我经常跟团队开个玩笑说，如果海外的消费者，我们以美国为例，美国消费者打开手机或者打开这个电脑，他每天看短视频加长视频的时间假设有5个小时。我们希望我们的爆款视频涵盖在5个小时里面的2个小时。也就是说他在任何一个moments看到的视频，他在广告看到的视频是我们的爆款视频。他刷短视频看到的举证号是我们的视频。他进入到官网落地看到了什么视频，他在实域对话的时候看到了什么视频，他甚至未来去线下户外大牌看到的还是我们的爆款视频。所以我们其实希望他看到的整个链条都是我们的优质内容。
In other words, we genuinely hope our company can build an ecosystem around consumption. I often joke with my team that if we take American consumers as an example, assuming they spend about 5 hours a day watching short and long videos on their phones or computers, we want our hit videos to occupy 2 of those 5 hours. This means that whenever they watch videos—whether it's ads, short videos from our matrix accounts, landing page videos on official websites, conversational videos in private domains, or even future offline outdoor billboards—they’ll see our hit content. Essentially, we want every touchpoint in their entire journey to be filled with our high-quality content.

发言人1 01:03:03
Speaker 1 01:03:03

只是我们从导购这样的一个环节先出发，这样更好的解释。因为也回到我刚刚说的一个问题，就是中国的这个出海的企业，它的核心数据还是出货。所以用导购这个词更利于双方达成一个共识点。
We’re just starting with the shopping guide aspect to make it easier to explain. This ties back to what I mentioned earlier—the core metric for Chinese companies going global is still shipments. Using the term "shopping guide" helps both sides reach a common understanding more easily.

发言人3 01:03:18
Speaker 3 01:03:18

所以就是在各个环节都运用AI的能力。
So it's about leveraging AI capabilities at every stage.

发言人1 01:03:22
Speaker 1 01:03:22

对，你可以理解成内容生产的部分自然是运用了AI的能力。不管是用AI去做一些脚本，还有脚本的裂变分发，全触点的分发，包括分发多少预算，用了AI的能力未必是生成式，也可以是这个决策型ai再到思域的部分的对话的交互，当然自然是这个大模型的能力。然后再到甚至未来通过所有的消费洞察反向指导客户的新的产品的开发，他都是运用到了AI的能力。所以用数字导购这个词去概括，它更像是整件这个事情的一个入口，但它未必是一个收口，收口其实是把AI能力贯穿整个消费链条，从内容的生产到内容的分发，再到数据的回流迭代。
Exactly. You can understand that the content production part naturally utilizes AI capabilities. Whether it's using AI to create scripts, script fragmentation and distribution, omnichannel distribution, including budget allocation for distribution, the AI capabilities employed aren't necessarily generative—they can also be decision-making AI. Then there's the interactive dialogue in private domains, which of course relies on large model capabilities. Looking further ahead, even using all consumption insights to guide new product development for clients involves applying AI capabilities. So using the term "digital shopping guide" to summarize this is more like describing the entry point of the whole process, but it's not necessarily the endpoint. The endpoint is actually about integrating AI capabilities throughout the entire consumption chain—from content production to content distribution, and then to data feedback iteration.

发言人4 01:04:06
Speaker 4 01:04:06

其实回到年初的时候，刚才我们说这个还是2024年录的博客。年初的时候一年前了差不多我跟leo我们当时聊了蛮多的，我非常激动的，我说句子。接下来聚焦做RPA加AI打造大模型驱动的数字员工，释放万倍生产力，定义未来工作方式，用AI重塑整个的工作流。然后vivo后来基于这个又做了一个非常高级的升级，我觉得特别好。后来定位成数字导购。其实我最近也在思考这件事，数字员工又有点过于抽象化，怎么能做的更聚焦落地，还是从sales的角度，怎么能够让我们的目标客户听懂这到底是一个什么东西，我这个是很重要的。所以既然我们刚才聊到了所有的agent到AI应用，继续聊AI应用。
Actually, going back to the beginning of the year—just now we mentioned this was a podcast recorded in 2024. It was about a year ago at the start of the year when Leo and I had quite a few discussions. I was really excited and said something like: "Next, we’ll focus on combining RPA with AI to create large-model-driven digital employees, unleashing ten thousand times the productivity, defining the future of work, and reshaping entire workflows with AI." Later, Vivo took this a step further with a very advanced upgrade, which I thought was fantastic. Eventually, it was positioned as a digital shopping assistant. Recently, I’ve been reflecting on this—the term "digital employee" feels a bit too abstract. How can we make it more focused and practical? From a sales perspective, how can we ensure our target customers understand what this actually is? That’s very important. So, since we’ve just been talking about agents and AI applications, let’s continue the discussion on AI applications.

发言人4 01:04:52
Speaker 4 01:04:52

上次我们聊到一个很有意思的话题，就是目前AI生成的数字人视频。比如黑镇那种具有真人形象的视频，其实可能并不适合带货或者做这种效果导向的场景，因为和真人的感受还是不太一样的。你认为那种视频可能会比较适合类似培训这种对个性化和情感性要求比较低的场景。
Last time, we touched on a very interesting topic: AI-generated digital human videos. For example, videos like those from Hei Zhen, which feature lifelike human appearances, might not actually be suitable for live-streaming sales or performance-driven scenarios because the experience still feels different from interacting with a real person. You mentioned that such videos might be more appropriate for scenarios like training, where the requirements for personalization and emotional engagement are relatively low.

发言人4 01:05:15
Speaker 4 01:05:15

在营销场景，比如说基本上做的我们都是结合这个right检索增强生成或者说外挂知识库去做的。这些agent其实已经覆盖了很多的线上的客服销售的场景。包括文本语音等多模态场景的应用，其实核心就是线上服务的能力。还有一些经过像fine的图文、视频等模态的内容编辑器，以及处理一些简单的流程化重复性工作的agent，也已经找到了PMF，并且也产生了一定量的收入。因为你一直在接触最一线的客户和市场，而且在各方面应用AI提升营销效率，降低营销成本，也拿到了标好好的结果。所以就想请你就几个具体的案例来聊一聊在出海营销的整个链路上，不同场景需要什么样的AI的能力，有什么样好的use case。
In marketing scenarios, for example, what we primarily do is combine retrieval-augmented generation or external knowledge bases. These agents already cover many online customer service and sales scenarios. This includes applications across multiple modalities like text and voice, with the core being online service capabilities. Additionally, content editors for modalities such as fine images, text, and videos, as well as agents handling simple, repetitive workflow tasks, have already found product-market fit (PMF) and generated a certain amount of revenue. Since you’ve been consistently engaging with frontline customers and markets, and leveraging AI to enhance marketing efficiency and reduce costs, achieving solid results, I’d like to ask you to discuss a few specific cases. What AI capabilities are needed across different scenarios in the entire overseas marketing chain, and what are some good use cases?

发言人1 01:06:08
Speaker 1 01:06:08

那我就用一个电动单车的一个case做一个完整的阐述。比如说我们拿到一个自行车的电动自行车，它其实是挺重的。我们就在想这样一个自行车户外的场地那么多，那我们难道每个方向脚本都拍一遍吗？其实这个成本是很高的，它不像做小家电。小家电我们其实租几个欧美范儿的公寓，我们的成本是可控的。但是如果移到户外，我们去公园，那就是几个编导的一天的时间，甚至几天的时间可能都没有爆款。
Let me use an electric bike case as a complete example. For instance, we had an electric bicycle that was quite heavy. We wondered: with so many outdoor locations for such a bike, would we have to film scripts for every possible direction? The cost would be very high—it’s not like small appliances. For small appliances, we could rent a few apartments with a Western aesthetic, and the cost would be manageable. But for outdoor shoots, like in parks, it would take a whole day or even several days for a few directors, and we still might not get a hit.

发言人1 01:06:40
Speaker 1 01:06:40

所以我们就要借助AI其实第一步就是我们通过把所有的爆款这个行业的视频，已有的视频全部放在这个AI里面去做两轮。一个是OCR先识别它的这个故事线，现在我们主要用的是这个german I因为这确实2.0之前是1.5，现在2.0它的整个视频的识别，哪怕没有旁白识别图片也是非常强的。所以我们一般会把爆款视频先放入到江南去做整体大规模的视频识别。
So we leverage AI, and the first step is actually to input all the viral videos from this industry into the AI system for two rounds of processing. First, we use OCR to identify the storyline. Currently, we mainly use German I because, truth be told, before version 2.0, it was 1.5, and now with 2.0, its video recognition capabilities—even for images without narration—are exceptionally strong. So, we typically start by feeding the viral videos into Jiangnan for large-scale video recognition.

发言人1 01:07:08
Speaker 1 01:07:08

识别完了之后，我们会归纳总结，从已有的视频里面几条热门的故事线，然后再做具体的尝试。比如说这个故事线告诉我们说，我们一定要强调避震。就是这个避震的功能其实是很多美国消费者买电动单车最关注的一个点。因为他们经常会爬山地，如果太正的话，其实骑的感觉是非常差的。所以我们就强调如果是避震这个点，我们是不是真的要走向那么多公园呢？会不会其实我们在我们这个小区找一个相对有欧美感觉的庭院就可以这样做的尝试。然后我们就真的这样拍了拍完之后我们在youtube上的视频就拿到了1700万的播放，一个非常好的结果。
After the recognition is complete, we summarize and extract several popular storylines from the existing videos, and then we proceed with specific tests. For example, one storyline tells us that we must emphasize shock absorption. This feature is actually a key concern for many American consumers when purchasing electric bikes, as they often ride on mountainous terrain. If the bike lacks proper shock absorption, the riding experience can be very poor. So, we asked ourselves: If shock absorption is the focus, do we really need to film in so many parks? Couldn’t we just find a courtyard in our neighborhood that has a European or American vibe and try filming there? And that’s exactly what we did. After shooting, our video on YouTube garnered 17 million views—a fantastic result.

发言人1 01:07:44
Speaker 1 01:07:44

然后紧接着我们就进一步，我们就在想这样的视频是不是可以裂变呢？我们就把这样的一个脚本，我们1700万的脚本又放到AI里面，让AI基于我们已有的脚本去裂变成20条到30条。然后我们把所有裂变的视频也拿过来去做拍摄，发现裂变的这个视频拍出来的播放量还是有几百万。然后紧接着我们又突发奇想，我是不是可以把这个裂变的视频找老外来拍摄呢？也是同样的脚本，结果我们又做了一些脚本的微创新。然后我们没有干涉老外，让老外基于我们的访拍，老外有一些自己的想法。但是所有的视频强调的核心点都是逼真，所以每一个视频都基本上在500万以上。然后我们也取得了比较好的销售的效果。
Then immediately afterward, we took it a step further and wondered if such videos could be replicated. We took that same script—our 17-million-view script—and fed it back into AI, asking the AI to generate 20 to 30 variations based on our existing script. We then filmed all these replicated videos and found that they still garnered millions of views each. Next, we had a sudden idea: what if we had foreigners film these replicated videos? Using the same script, we made some minor creative tweaks. We didn’t interfere with the foreigners’ approach; we let them bring their own ideas based on our reference footage. But the core emphasis across all videos was realism, so each one consistently achieved over 5 million views. This also led to significantly improved sales results.

发言人1 01:08:23
Speaker 1 01:08:23

所以这个环节其实是我们用AI快速的找到洞察，再到AI去生成多元的脚本，再到AI去做裂变。其实它发挥了我们在视频的图像识别。包括故事性的总结到这个内容的裂变。
So this process essentially involved using AI to quickly identify insights, then leveraging AI to generate diverse scripts, and finally employing AI for replication. It showcased our capabilities in video image recognition, narrative summarization, and content replication.

发言人1 01:08:36
Speaker 1 01:08:36

这个过程完成了之后，因为我们的短视频有段流量是到独立站成交，因为它的客单价很高。那这个时候我们就要去分析我们做的大量的内容，尤其是受欢迎的内容，它对于投放有没有价值。我们就开始把我们的所有的视频提炼出内容标签，把这些内容标签放到google或者meta的投放系统。
After this process was completed, since a portion of our short video traffic converts on independent websites—given their high average order value—we then needed to analyze the large volume of content we produced, especially the popular content, to determine its value for ad placements. We began extracting content tags from all our videos and inputting these tags into Google or Meta's ad placement systems.

发言人1 01:08:55
Speaker 1 01:08:55

然后投放系统拿到了这些非常精准的内容方向的标签之后，它的AI投放的方式，因为现在大厂不管是tiktok、meta、google都有智能投放。就是你只要圈选你的目标，他就会基于你的LI ros的要求去帮你圈定人群。但你需要给他一些一方数据的输入。所以我们把爆款视频的内容标签输入给到google meta，然后再给google matter定一个目标。这个时候我们明显发现我们可以接受的ross，比如说3或者4可以消耗更大的人群包。因为平台从AI计算来讲，他更懂你，因为你给了他的输入，所以投放过程中我们就拿到了更好的RY表现。
Once the ad systems received these highly precise content-oriented tags, their AI-driven placement methods—since major platforms like TikTok, Meta, and Google now all offer smart placements—allowed us to simply define our targets. The system would then identify audiences based on our ROAS (Return on Ad Spend) requirements. However, it needed some first-party data inputs. So, we fed the content tags from our viral videos to Google and Meta and set a target for them. At this point, we clearly observed that we could achieve acceptable ROAS, say 3 or 4, while reaching larger audience pools. From an AI computation perspective, the platform understands you better because you've provided the input, resulting in better performance metrics during the ad placements.

发言人1 01:09:33
Speaker 1 01:09:33

这个环节中间当更多人投放，那就会有更多人的评论。评论一旦上来的时候，我们又希望进一步去把很多的人能够拉到FB group或者拉到what sap。所以这个时候，大量的评论，其实如果不再及时的对话，其实这些人就流失了很多消费机会。因为你不及时对话，他会觉得你的客服，尤其对于高客单产品，如果不及时去对话，很容易被流失。因为美国人他会觉得是不是你这企业不够大，你的售后服务可能跟不上。所以这个评论的技术维护其实也是要用到AI无论是说老客户的召回，我们说remarketing还是说retargeting还是这个潜客的及时第一时间做承接。
During this phase, as more people engage in advertising, there will naturally be more comments. Once the comments start pouring in, we aim to further guide many of these users to join our Facebook group or WhatsApp. At this point, if we don’t respond to the large volume of comments promptly, we risk losing many potential sales opportunities. Failing to engage in timely conversations can make customers—especially those considering high-ticket products—feel that your customer service is lacking. Americans, in particular, might perceive your business as too small or worry that after-sales support is inadequate. Therefore, managing these comments effectively also requires AI, whether it’s for remarketing to existing customers, retargeting, or promptly engaging potential leads to secure their interest.

发言人1 01:10:15
Speaker 1 01:10:15

来自客服邮件营销、短信营销这里面这个千人千面的内容的定制，其实都是依赖于AI的。因为如果我们一天都有个几万个评论，让每个人去甄选和筛查，其实这个工作量是很大的。最好是人机结合，比如说AI先筛一遍，有个初始的回复，然后AI筛查掉这个最高意向的客户，高意向的由人去跟进。
The customization of personalized content for customer service emails and SMS marketing relies heavily on AI. If we receive tens of thousands of comments daily, manually screening and selecting each one would be an enormous task. The optimal approach is a human-AI collaboration: AI can first filter the comments and provide initial responses, then identify the most high-intent customers for human follow-up.

发言人1 01:10:37
Speaker 1 01:10:37

中等第一印象就是我们叫养鱼。养鱼的这个环节可以交给AI的这个过程中去这样形成一个协同关系。这个协同其实AI也在被培训，因为每次养鱼和这个潜客的成交的所有的数据再反馈给AI那AI的养的鱼就会越来越多，甚至AI可以直接去承接一些高潜的用户了。
The first impression is what we call "fish farming." The fish farming process can be handed over to AI, creating a collaborative relationship. This collaboration essentially trains the AI, because all the data from fish farming and closing deals with potential customers is fed back to the AI, allowing it to "farm" more and more fish. Eventually, the AI can even directly handle some high-potential users.

发言人1 01:10:59
Speaker 1 01:10:59

人力可以被进一步释放，释放出来的人可以进一步去想产品要怎么去创新，内容要怎么去创新。包括说这个整体的围绕消费者互动的大的故事线要怎么去创新。仍旧从很多机械性的工作回到了偏原创性的工作，这也能很释放每个人的生产的效率。
Human resources can be further liberated, freeing up people to focus more on how to innovate products and content. This includes thinking about how to innovate the overarching narrative around consumer interactions. It shifts many mechanical tasks back to more creative work, which can significantly enhance everyone's productivity.

发言人4 01:11:19
Speaker 4 01:11:19

我也顺便给我的公司桔子互动带个货。刚才聊到了很多比较好的这些场景。然后如果大家对于tiktok的评论的及时回复，用agent去在评论区去做转化，以及怎么在tap里面去做私域成交，怎么在国产里边做一个agent做一个数字销售，都可以直接联系我们，联系主播。
Let me also take this opportunity to promote my company, Orange Interactive. We've just discussed many excellent scenarios. If you're interested in timely replies to TikTok comments, using agents to drive conversions in the comment section, or how to facilitate private transactions within the app, or even creating a digital sales agent within domestic products, feel free to reach out to us directly—contact the host.

发言人1 01:11:42
Speaker 1 01:11:42

可以联系我，我这边也可以帮助去承接这部分的服务。
You can also contact me—I can help provide these services as well.

发言人4 01:11:46
Speaker 4 01:11:46

对，可能联系reo会更好。因为vio可能他会有更强的工具之外的一整套的解决方案，更多的怎么从流量侧完整成交的解决方案。
Yes, contacting REO might be better. Because VIO may have not just stronger tools but a complete set of solutions, more comprehensive solutions for achieving conversions from the traffic side.

发言人3 01:11:56
Speaker 3 01:11:56

接下来我们聊的宏观一点，看一看全球市场。全球市场是一个很大的话题，其实刚才will也有涉及到北美、中国、欧洲、日韩、东南亚、拉美、中东、非洲等等。每个市场都不一样，开放和封闭并存，全球化和区域化交织，政治经济、文化产业结构、消费习惯等等都同时存在差异和共性，而且还在不断的变化。但是目前有一个客观的现实摆在几乎所有中国，尤其是有规模的企业面前，就是不出海就出局。这也是我们这个博客诞生的一个原因，就希望能够传递出海全球化一线的声音。
Next, let’s discuss on a broader level and look at the global market. The global market is a vast topic, and Will just touched on regions like North America, China, Europe, Japan and Korea, Southeast Asia, Latin America, the Middle East, Africa, etc. Each market is different, with openness and closedness coexisting, globalization and regionalization intertwined, and variations and commonalities in political economy, cultural-industrial structures, consumption habits, and more—all constantly evolving. However, there’s an objective reality facing almost all Chinese companies, especially those of scale: go global or be left behind. This is also one reason our podcast was born—to amplify firsthand voices on globalization and going global.

发言人3 01:12:32
Speaker 3 01:12:32

上次咱们聊，我觉得你的观点很有意思。比如说刚才你也涉及到欧美市场，包括美国、欧洲、澳洲还有日本。这些市场的规则，这其实是由美国公司来制定的。而东南亚其实是中国市场的延伸，规则是由中国的公司来制定的那再比如说拉美、中东你认为是新兴市场，而东欧、非洲、印度这些更像是孤岛。具体到商品，比如你认为带电的都适合做美国市场等等。因为你过去的经历很大程度上都是在做中国企业的出海，不管是硬件还是软件，不管是美国市场还是东南亚市场。所以也想请你继续谈一谈怎么看当前全球主要商业市场的格局，以及各自存在的渤海全球化的机会。
Last time we talked, I found your perspective quite interesting. For example, you mentioned markets like the US, Europe, Australia, and Japan—where the rules are essentially set by American companies. Meanwhile, Southeast Asia is more like an extension of the Chinese market, with rules shaped by Chinese companies. Then there are emerging markets like Latin America and the Middle East, while regions like Eastern Europe, Africa, and India seem more like isolated islands. When it comes to specific products, you suggested that anything involving electronics is well-suited for the US market, among other insights. Given your extensive experience in helping Chinese companies expand overseas—whether in hardware or software, targeting the US or Southeast Asia—I’d love to hear more about how you view the current landscape of major global commercial markets and the opportunities they present for globalization.

发言人1 01:13:16
Speaker 1 01:13:16

因为出海的类目一般是三个类目，一个是消费品，一个是本地生活就是像茶饮，奈雪的茶这样这个lucky coffee，还有一个是这个APP，就比如说像佳瑞的出海对吧？像这个AI软件的出海。所以这三个行业出海，因为时间关系，我重点讲消费品出海。
The categories for overseas expansion generally fall into three types: consumer goods, local lifestyle (such as tea beverages like Nayuki Tea or Luckin Coffee), and apps (for example, Jia Rui's overseas expansion, right? Like AI software going global). Due to time constraints, I’ll focus primarily on the overseas expansion of consumer goods among these three industries.

发言人1 01:13:32
Speaker 1 01:13:32

消费品出海，其实回到消品自身，如果这个商品本身它的竞争的壁垒是比较高。比如说它涉及到高研发强度和比较强的技术专利，那建议的这个出海路线还是以美国或者说北美统一大市场为主。北美做好了之后，从英国出发，其实欧洲英德法意西这五个发达国家其实就进去了。然后欧洲进去之后，基本上这个时候，北美的延伸就会延伸到墨西哥，然后再到南美，因为他们是一个体系。然后这个英德法意西，尤其是像德国覆盖了之后，其实像中欧的一些地方。然后法国覆盖了之后，其实北非的一些地方，其实这两个角就已经撕开了。那剩下的这个部分通常如果还要专注一些高净值的用户，就是日韩包括澳新，其实五眼联盟里面已经cover了。把这些发达国家全做完了之后，对于这种高价值的产品，或者说有一定壁垒的产品，可以选择做一些廉价产品，或者说低客单价的产品进去。
For consumer goods going global, it really comes down to the product itself. If the product has high competitive barriers—such as involving intensive R&D and strong technological patents—then the recommended route for global expansion should primarily focus on the United States or the unified North American market. Once North America is established, starting from the UK, you can essentially enter the five major developed markets in Europe: the UK, Germany, France, Italy, and Spain. After penetrating Europe, the natural extension from North America would lead to Mexico and then South America, as they share a similar system. For instance, Germany’s coverage effectively opens up parts of Central Europe, while France’s reach extends to parts of North Africa—these two corners alone create significant inroads. The remaining focus, if targeting high-net-worth users, would be on Japan, South Korea, and Australia-New Zealand, which are already covered within the Five Eyes alliance. After securing these developed markets, for high-value or barrier-heavy products, you can then consider introducing more affordable or lower-priced products.

发言人1 01:14:30
Speaker 1 01:14:30

发展中国家，比如说像中东，包括东南亚，还有南美、非洲，非洲可能是最后一个赛道了。基本上就是以这种高端的产品去打造这个品牌力，然后由中低端产品去进入发展中国家这样的一个曲线。如果是一般的跨境电商卖家，我把品牌出海和卖家除外，其实分成两类。卖家除外。简单来说就是说一上来可能没有办法承担用大几百万、上千万这样的一个研发的成本。
For developing markets—such as the Middle East, Southeast Asia, South America, and Africa (likely the last frontier)—the strategy is to build brand strength with premium products first, then enter these regions with mid-to-low-end products. For general cross-border e-commerce sellers (excluding brand globalization and certain seller categories), they can be broadly divided into two types. Excluding certain sellers, the simpler scenario is that many may not initially afford the R&D costs running into millions or tens of millions.

发言人1 01:14:59
Speaker 1 01:14:59

想说能够尽快拿到一些正反馈，能够实现尽快的营收平衡，或者说实现尽快的回报。那就是我们把它定义成卖家这类商品通常它的客单价不会特别高，一般是低于200美金。它的商品的迭代周期要非常快。因为通常价格不高的产品，它的卷度也是比较高的。就是你可以做，你看到这个机会，别人也可以做。这样的一个产品，通常就要找到一个核心的市场了。
The goal is to obtain positive feedback as quickly as possible, achieve revenue balance promptly, or realize returns swiftly. We define this as targeting seller-type products, which typically don't have particularly high price points—usually under $200. The product iteration cycle needs to be very fast because generally, lower-priced products face intense competition. If you spot an opportunity, others can too. For such products, it's crucial to identify a core market.

发言人1 01:15:23
Speaker 1 01:15:23

举个例子，可以借这个平台的东风，不管是出海，是小龙徐印、提莫，包括速卖通或者当下最热门这个title shop，然后找出一个细分的市场做一个切入点。但这里面最好是回到洞察，不能以偏概论。就是说你要看从选品角度，这四大平台跟着哪一个平台出去，在哪一个单一市场和单一的类目下有一定的利润空间。虽然说现在大家新闻上看到，比如说像提莫他的这个利润率并不高，但这个也不是说以偏概全，有些垂直行业还是有一些利润空间在的。所以对于客单价不高的话，可以借东风，就是借出海的平台。当然也可以说，直接在价格带在比如说50美金到200美金之间，可以直接上美国的平台，比如说amazon或者说德国的auto。
For example, you can leverage the momentum of platforms—whether for global expansion, like Xiaolong Xuyin, Timo, AliExpress, or the currently trending TikTok Shop—and find a niche market as your entry point. However, it's best to base this on insights rather than generalizations. From a product selection perspective, you need to determine which of these four platforms to use for expansion and identify a specific market and category where there's potential for profit. While news reports might suggest, for instance, that Timo's profit margins aren't high, this isn't universally true—some verticals still offer room for profit. For products with lower price points, you can ride the wave of global expansion platforms. Alternatively, for items priced between $50 and $200, you could directly target platforms like Amazon in the U.S. or Auto in Germany.

发言人1 01:16:08
Speaker 1 01:16:08

就是找一些当地的平台，甚至可以考虑进驻线下。比如说像线下的一些小的经销商，不要进沃尔玛大的。因为大的线下商超其实是有很大的门槛要求的。像这些小的商超夫妻店，比如说像一些指甲油这样的一个小产品。其实你要进入到是一些这种小店，反而它的销量甚至会好于你的这个线上。但总体来讲，做卖家其实核心拼的还是选品的效率，就是你确实要找到一个细分的市场。
Consider exploring local platforms and even offline channels. For instance, target small local distributors rather than big retailers like Walmart, as major supermarkets often have high entry barriers. Small mom-and-pop stores—like those selling niche products such as nail polish—can actually outperform online sales in some cases. Ultimately, the key to success as a seller lies in product selection efficiency—you really need to identify a niche market.

发言人1 01:16:37
Speaker 1 01:16:37

当下让我来看整个卖家的角度来讲，一般比较稳妥的路径最好是选择一些新兴的市场。比如说今年应该就是说24年整体来看的话比较热，可能还是南美这个赛道就是竞争没有那么多，但是相对来讲，消费力也在逐步起来。如果24年去杀进东南亚这个市场，有可能他已经过了他最好的一个timing。因为东南亚的时间可能会在22年到23年左右的这个时间。
From a seller's perspective, the safest path right now is to focus on emerging markets. For example, in 2024, Latin America stands out as a hot but less competitive sector with growing purchasing power. If you enter the Southeast Asian market in 2024, you might have missed its prime timing, as the golden period for Southeast Asia was likely around 2022 to 2023.

发言人1 01:17:04
Speaker 1 01:17:04

非洲现在还没有完整的发展起来，中东市场是在投资圈融资圈比较热门。但是真正从消费角度来讲，其实中东市场其实挺复杂的，他对于特定类目有一定的要求。所以我目前看到中东还是偏品牌，做好了欧美日韩，然后它会单独进中东。因为中东海湾国家它其实消费力并不差，你不能把它归为像这种卖家的市场。所以中东会有点像between，就是它在两者之间你可以做品牌往下升，你也可以做这个卖家的方式往上升。但总体来讲如果是说卖家，我个人的建议不应对。南美是一个非常不错的市场，包括在俄罗斯方向。因为俄罗斯跟中国的贸易也是因为特殊的情况比较多，俄罗斯方向的一些产品线也是可以值得做一些进步的。
Africa has not yet fully developed, while the Middle East market is quite popular in investment and financing circles. However, from a consumer perspective, the Middle East market is actually quite complex, with specific requirements for certain categories. So far, what I've observed is that the Middle East still leans toward branding—companies first establish themselves in Europe, the U.S., Japan, and South Korea before entering the Middle East separately. The Gulf countries in the Middle East actually have strong purchasing power; you can't categorize them as a typical seller's market. The Middle East is somewhat in-between—you can either build a brand and scale down or adopt a seller's approach and scale up. But overall, if we're talking about sellers, my personal advice is not to prioritize it. South America is a very promising market, as is Russia. Due to the unique circumstances surrounding trade between Russia and China, certain product lines in the Russian market are also worth exploring for growth.

发言人3 01:17:48
Speaker 3 01:17:48

大概是这样两个维度。你刚才说的是偏卖家，那偏品牌的就是我想要做一个全球化品牌的。
These are roughly the two dimensions. What you mentioned earlier leans more toward sellers, while branding is about wanting to create a global brand.

发言人1 01:17:56
Speaker 1 01:17:56

中学生化品牌的话，第一核心就是你的这个。产品力，因为产品我觉得始终是做品牌的核心。因为一个好的产品，它的生命周期和它的现金流支撑可以让你做很多多元化的事情。但如果你本身产品一没有够好，你就急于扩张去做0，有可能会被反噬，这是第一个。
For building a mid-range chemical brand, the first core element is your product strength. Because I believe products are always the heart of branding. A good product's lifecycle and cash flow support can enable you to pursue many diversified initiatives. But if your product itself isn't strong enough and you rush to expand recklessly, it might backfire—that's the first point.

发言人1 01:18:15
Speaker 1 01:18:15

我觉得产品始终是核心，营销做的再漂亮，就像我们去服务品牌的时候，其实我也是优先去看它的产品力。所以我们通常会免费帮这个产品先拍个三四条，发到短视频平台上面，看一下它的播放量怎么样。我们通过数据去验证。
I believe products remain the core. No matter how impressive the marketing is—like when we serve brands, I always prioritize evaluating their product strength first. So we typically help shoot three or four free promotional videos for the product and post them on short video platforms to observe their viewership performance. We validate through data.

发言人1 01:18:31
Speaker 1 01:18:31

所以其实对于产品第一步假设没有问题了，第二步回到这个全球化扩张。最好是找到自己的当地的竞争的情况。如果说一上来进美国，觉得这个相对的市场会比较难进入，能拿到一个份额的话，可以考虑从欧洲的一些国家，比如说从英国先切入，或者从加拿大先切入，先在一个小的市场去做你产品的PMF，那不至于影响到你的大盘。就像title shop刚推出的时候，其实它先推出英国小店，然后在英国小店有一定的基础之后，他才去探索美国。然后美国打开之后，他才开始把各大国家，全球各个区域在陆续打开。所以做品牌也是可以有一个这样的一个PMF的市场，可以是英国，可以是澳洲一些小众市场。
So actually, regarding the product, once the initial hypothesis is validated, the second step is to return to global expansion. It's best to first understand the local competitive landscape. If entering the U.S. market directly seems relatively difficult to penetrate and gain a share, you might consider starting with some European countries, such as the U.K., or Canada—entering a smaller market first to validate your product's PMF (Product-Market Fit). This way, it won’t negatively impact your broader strategy. For example, when Title Shop first launched, it started with a U.K. store. Only after establishing a foothold there did it explore the U.S. market. After succeeding in the U.S., it gradually expanded to other regions globally. Similarly, building a brand can follow this PMF approach—starting with markets like the U.K. or niche markets in Australia.

发言人1 01:19:12
Speaker 1 01:19:12

先做一个成本可控的一个尝试，然后再进入美国，如果一上来进入美国高举高打的话，如果资金实力相对比较充裕一些，我建议也不要单独去只看电商。线下其实是一个非常关键的赛道。如果可行的话，我建议是空军和陆军并行。空军就是说电商的方式可以拆分成amazon这种平台。电商也有包括DDC这种独立站，还有包括d to shop这样的内容介绍，包括像ins youtube也有这样类似的这种尝试路径的话，虽然没有办法马上进入到像我们他给costco这样的一个核心头部的商超，但是可以考虑找一些当地的这个K就是这种当地的这种经销商的销售，进入一些细分的小众的市场。比如说美国的中部地区的一些市场也会有一些反馈。
Start with a cost-controlled experiment before entering the U.S. If you decide to enter the U.S. aggressively from the outset—and if you have relatively strong financial resources—I’d advise not focusing solely on e-commerce. Offline channels are actually a critical avenue. If feasible, I recommend running both "air force" and "ground force" strategies in parallel. The "air force" refers to e-commerce approaches, which can be broken down into platforms like Amazon, DTC (Direct-to-Consumer) independent sites, and content-driven platforms like TikTok Shop. While it might not be possible to immediately enter core retail giants like Costco, you could consider partnering with local distributors (KOLs or local sales channels) to penetrate niche markets. For instance, some midwestern regions in the U.S. might also yield positive feedback.

发言人1 01:19:57
Speaker 1 01:19:57

然后当然还有一些海军的层面，海军层面就是可以去考虑有些产品相对低客单的，不用一定是品牌除外，有一个现金流。比如说考虑被贴牌，如果你有比较强的制造能力，你可以有两条产线。所以我觉得这个海陆风的这种方式，会让整体这个空军市场走得更顺。如果单独只靠空军，就是只靠你这个电商整体来讲风险会大一些。还是回到我刚刚说那个策略，可以先找一个细分做一个试水，然后再放大。
Then, of course, there’s the naval aspect. On the naval front, you can consider products with relatively low price points—they don’t necessarily have to be branded—to generate cash flow. For example, think about white-labeling. If you have strong manufacturing capabilities, you could operate two production lines. So, I believe this combined land-sea-air approach will make the overall market strategy smoother. Relying solely on the air force—meaning just e-commerce—would generally carry higher risks. It goes back to the strategy I mentioned earlier: start with a niche segment to test the waters, then scale up.

发言人3 01:20:28
Speaker 3 01:20:28

接下来我们把视角放到企业身上，如果你去问中国有一定规模的企业创始人想不想出海呢？那十有八九可能都说想出海。但实际上我个人的一个发现，就是小部分有能力出海的企业已经在行动，甚至已经拿到结果了。而大部分没有出海的企业，除了缺乏勇气，更多的就是不知道怎么出海，也不知道需要什么样的能力。其实这个能力可能并不一定非要是有海外的经历或者有多少的资源，也并不一定非要建立本土化的团队。就是我最近和一个全球化做的很好的一个软件公司的高管交流学到的总之可能并没有想象中那么难。
Now let's shift our perspective to businesses. If you were to ask founders of mid-to-large Chinese companies whether they want to expand overseas, eight or nine out of ten would likely say yes. However, based on my personal observations, only a small portion of companies with the capability to go global are actually taking action—some have even achieved results. For most companies that haven't ventured abroad, beyond lacking courage, the bigger issue is not knowing how to expand internationally or what capabilities they need. In reality, these capabilities don't necessarily require overseas experience or abundant resources, nor do they always demand building a localized team. As I recently learned from an executive at a highly globalized software company, it might not be as difficult as imagined.

发言人3 01:21:03
Speaker 3 01:21:03

另外一方面，我看梁丽颖老师的新书真需求，有一个章节的标题就是中国到该出大牌和大师的时候了。我很赞成这个观点，这可能不是多数企业的机会，但这个历史性的机会是存在的。因为lip主要也是在做这个出海营销的服务，且你已经服务过很多头部的全球化的企业，包括一些国际的巨头。也想请你站在一个普通的准备出海的中国企业的这个视角，从营销这个方面分享一下。就是企业要想做好这种出海或者说全球化的营销核心需要做什么准备，以及需要具备哪些能力呢？
On the other hand, I read Teacher Liang Liying's new book "True Demand," which has a chapter titled "It's Time for China to Produce Big Brands and Masters." I strongly agree with this view. This may not be an opportunity for most companies, but this historic opportunity does exist. Since Leap is primarily engaged in providing overseas marketing services and has already served many leading global enterprises, including some international giants, I’d like to ask you to share, from the perspective of an ordinary Chinese company preparing to go global, what core preparations and capabilities are needed to excel in overseas or global marketing?

发言人1 01:21:38
Speaker 1 01:21:38

我觉得就是两个部分。第一个是文化上的一个适应。以我自己的企业为例，你像我的个人经历，其实很长一段时间是在做国内营销。然后我到海外营销其实也是经历了大概有2到3年的转变。就是我们不能一味的把把国内的成功模式套用在海外上，还是要有一个敬畏。
I think there are two key aspects. The first is cultural adaptation. Taking my own company as an example, my personal experience has largely been in domestic marketing. Transitioning to overseas marketing took me about 2 to 3 years. We can’t simply replicate domestic success models overseas; we must approach it with respect.

发言人1 01:21:56
Speaker 1 01:21:56

所以我觉得第一个是文化和组织能力的适配。如果这个适配中间感觉挑战比较大的话，可以考虑相似度。比如说抖音的人才其实通过一定的调整可以去做tiktok这样的内容电商，天猫的人才可以一定程度上适配到amazon电商。但如果条件允许的话，建议还是报native。就是你的团队从一开始的海外的原生能力和国内它其实是两套班子，可以在数据洞察层面上去做一些共享。但是建议是从一开始的文化和组织结构打造上，就是就是原生化的。
So I think the first consideration is the alignment of culture and organizational capabilities. If the challenges in this alignment feel significant, similarity can be considered. For example, TikTok talent can, with some adjustments, transition to content e-commerce like TikTok, and Tmall talent can to some extent adapt to Amazon e-commerce. However, if conditions permit, it's advisable to go native. That is, your team's overseas native capabilities and domestic capabilities should essentially be two separate units from the outset, though some data insights can be shared. But it's recommended that from the very beginning, the culture and organizational structure should be built natively.

发言人1 01:22:28
Speaker 1 01:22:28

甚至可以考虑包括像产品的设计迭代这样的，可以直接考虑放在海外的本土。比如说做美国市场，可以直接考虑在洛杉矶，在美国的中西部地区建厂建研发。如果有这个能力，所以这个环节就是跟消费洞察和产品研发，产品定位相关的。我觉得他会需要极强的本地化的能力，也就是本土化理解能力。可以在中国找到一些留学背景的人，也可以如果条件允许，直接在美国有这样的当地的团队的一个配置。
You can even consider aspects like product design iteration to be directly handled locally overseas. For instance, if targeting the U.S. market, you could directly consider setting up factories and R&D in Los Angeles or the Midwest. If the capability exists, this step is closely tied to consumer insights, product development, and product positioning. I believe it requires extremely strong localization capabilities, meaning a deep understanding of the local context. You can find people in China with study-abroad backgrounds, or if conditions allow, directly establish a local team configuration in the U.S.

发言人1 01:22:56
Speaker 1 01:22:56

然后第二个部分就是本身的这个资金实力，因为从海外要拿到正反馈的时间周期，其实它不会特别的短。所以如果说资金储备没有办法支撑你的供应链的回款，以及海外的这个仓储物流，包括甚至可能会有一些坏账。如果没有这样的一个支撑，包括产品它也不是说每个产品的生命周期都是足够长，也百分之百成功。所以要有一定的资金的储备，实现这个风控的能力。
The second aspect is the financial strength itself. The time cycle to receive positive feedback from overseas isn't particularly short. If your capital reserves can't support your supply chain payments, overseas warehousing and logistics, and even potential bad debts—without such support—and considering that not every product has a sufficiently long lifecycle or is guaranteed 100% success, it's essential to have adequate capital reserves to achieve risk control capabilities.

发言人1 01:23:25
Speaker 1 01:23:25

然后第三个当然就是我认为是最核心的能力，就是说你所处的这个海外市场，它是不是一个成长性的市场。虽然现在大家都在讲做美国市场，其实是对我们这样的公司，就是我们以服务型为主的公司，我一定是去成熟市场。因为成熟市场就意味着大量的竞争，拼的已经不是产品的进入，拼的是营销的效率。所以我们一定会选择竞争最激烈的这个商品市场，反而我们能吃到最大的份额。但是如果对于产品或者说对于品牌角度，需要找到适合自己的新兴市场，或者做一个增量市场，我倒不一定认为说一定要全部挤进美国这个市场。
The third point, which I believe is the most critical capability, is whether the overseas market you're targeting is a growth market. While everyone is talking about entering the U.S. market, for companies like ours—primarily service-oriented—we must focus on mature markets. A mature market implies intense competition, where the battle is no longer about product entry but marketing efficiency. Therefore, we always choose the most competitive product markets, as that’s where we can capture the largest share. However, for products or brands, it’s more about finding emerging markets that suit them or tapping into incremental markets. I don’t necessarily think it’s mandatory to crowd into the U.S. market entirely.

发言人1 01:24:04
Speaker 1 01:24:04

然后另外一个点就是说我觉得现在的企业，你不能简单用出海这个思维来代表。最好慢慢向从单纯的跨这个思路也升级到我们叫全球化企业的时候。就像美国的企业，欧洲的企业，乃至甚至日本的企业，其实他们从day one就已经是全球化的企业。因为他们不会说我先赢下这里，再赢下这里。比如说像日本企业，它其实没办法。韩国企业你看商业史就发现他其实没办法，他必须赢下全球性的市场，不然的话没有办法养活他这个企业的规模。所以，只是我们中国的单一市场够大，然后中国的历史自古以来都是说中国的市场已经够大了，外面的都是蚂蚁。但是在当下的一个市场，因为国内的增长放缓，恢复也需要一定的时间。
Another point is that I think current enterprises can no longer be simply represented by the mindset of "going global." It's better to gradually upgrade from the mere concept of crossing borders to what we call a globalized enterprise. Just like American companies, European companies, and even Japanese companies—they have been globalized enterprises from day one. They don't say, "Let's win here first, then win there." For example, Japanese companies actually have no choice. If you look at the history of Korean businesses, you'll find they also have no choice—they must win in global markets; otherwise, they can't sustain the scale of their enterprises. It's just that China's single market is large enough, and historically, China has always believed its domestic market is vast enough, treating external markets as insignificant. But in today's market, with domestic growth slowing and recovery taking time...

发言人1 01:24:49
Speaker 1 01:24:49

这里面有被动出海和主动出海。我觉得被动出海未必一定是消极观点去看这件事情，但是被动出海的时候一定要找准适合自己的生态位。就举个例子，如果在国内本身没有拿下头牌，比如说像品牌出海，通常在国内如果没有拿下国内的制高点，国内反而变成不上不下。比如说国内在第五第六位。举个例子，国内的份额还在不断挤压，这个时候可能出海的心态会被影响。所以我建议如果品牌出海，最好是国内涨涨相对可预期的确定性之后，选择潜力大的市场，然后整个心态上的出海的节奏会有的放矢。
There are two types of globalization: passive and proactive. I don’t think passive globalization should necessarily be viewed negatively, but when going global passively, it’s crucial to find the right niche that suits you. For example, if a brand hasn’t secured a leading position domestically—say, it’s stuck in the fifth or sixth spot—and its domestic market share is shrinking, the mindset for going global might be affected. So, I suggest that if a brand is going global, it’s best to do so after achieving relatively predictable certainty domestically, then target high-potential markets. This way, the pace and focus of globalization will be more purposeful.

发言人1 01:25:26
Speaker 1 01:25:26

但如果假设国内这个市场又割舍不了，又一直在节节败退。那这个时候出海的期待就希望快速回流，有可能出来反而做的不好。以我们自身为例，我最开始创业的时候，其实我有很多国内的需求，小红书的种草和抖音的种草。我第一年创业的时候，其实我是既要又要还要，但最后我发现我失败的非常惨。因为当你既要又要还要的时候，你的思考体系和你的生态位的扎根的深度都是不够的。直到我直接彻底，即使我们非常懂国内营销，但是当我彻底拒绝国内这个赛道的时候，我在出海市场才算真正拿到了我的应有的份额。然后在现在我们出海相对来讲已经有一定的生态位的同时，偶尔有一些国内需求，我会在比较稳的方式去做一些。
But if we assume that the domestic market cannot be abandoned and yet continues to suffer repeated setbacks, the expectation for going global would be to achieve quick returns. However, it’s possible that venturing abroad might not yield good results. Take our own experience as an example. When I first started my entrepreneurial journey, I had many domestic demands, such as seeding content on Xiaohongshu and Douyin. In my first year of entrepreneurship, I wanted to have it all—but in the end, I failed miserably. Because when you try to have it all, your thinking framework and the depth of your niche’s roots are insufficient. It wasn’t until I made a clean break—even though we were very familiar with domestic marketing—that I truly secured my rightful share in the global market. Now, as we’ve established a relatively solid niche overseas, I occasionally handle domestic demands in a more cautious and stable manner.

发言人1 01:26:16
Speaker 1 01:26:16

所以我想通过自己的这个企业的实际的例子告诉大家，就是如果要出海的话，建议是聚焦国内转账。千万不要让它变成是你的一个包袱，而让它变成你的一个助力。如果国内还是一场包袱，我建议清洗完国内战场这样的品牌再考虑出海。如果非要并行的话，考虑和中国文化极其相近的东南亚市场，不建议去欧美。这仅个人观点。
So I want to use my own company as a real-life example to show everyone that if you're planning to go global, I recommend focusing on domestic transfers first. Don't let it become a burden—instead, make it your advantage. If the domestic market is still a burden, I suggest cleaning up your home market before considering overseas expansion. If you must operate in parallel, consider Southeast Asian markets that share strong cultural similarities with China, rather than jumping into Europe or America. This is just my personal opinion.

发言人4 01:26:40
Speaker 4 01:26:40

没有也是说到最后，过去几年天天跟我说的都是我国内包袱太重，没有很多观点，我也很同意。第一个一定要聚焦。因为对于创业公司来讲，包括甚至所谓的成功，很多公司走到一定阶段，它依然还是一家创业公司。所以非常核心的就是聚焦。而且尤其重要的就是创始人的经历是非常重要的，这个是非常宝贵的资产。所以到底往哪个地方打这个事儿在各种阶段都需要聚焦。
At the end of the day, over the past few years, people have often told me that my domestic baggage is too heavy and that I lack many perspectives, which I also agree with. The first priority must be focus. Because for startups—and even many so-called successful companies that have reached a certain stage—they are still essentially startups. So the core principle is focus. And what’s especially important is that the founder’s experience is crucial—it’s an invaluable asset. Therefore, no matter the stage, deciding where to direct your efforts requires focus.

发言人4 01:27:09
Speaker 4 01:27:09

我也很同意没有说到的本土化的这件事儿，我其实不太同意。就你刚刚说到的和哪一家全球化高管说的说不需要本土化团队，因为核心你要知道你的客户是谁，你的客户在哪，你就需要在哪。我们今天很多说的跨境的这些公司，其实它本质上没有服务跨境。包括巨子，我们服务的依然是中国品牌出海。我的客户依然是中国的公司，只不过是他在服务海外的公司。所以当明确自己的客户是谁的时候，才知道说我的本土化团队到底大家在哪。
I also agree with what was said about localization, though I actually disagree somewhat. Regarding what you just mentioned—what a certain global executive said about not needing a localized team—the core issue is knowing who your customers are and where they are. That’s where you need to be. Many of the so-called cross-border companies we talk about today aren’t truly serving cross-border markets. Even giants like us are still serving Chinese brands going global. Our clients are still Chinese companies; it’s just that they’re serving overseas companies. So only when you clearly identify who your customers are can you determine where your localized team should be.

发言人4 01:27:43
Speaker 4 01:27:43

就像我们我刚刚讲到的，不管是我们说跨境也好，全球化也好，那客户在哪儿，我们的团队就应该在哪。服务好客户很重要。但虽然我们跨着呢出海着呢全球化的，但是我的客户依然在国内。
As I just mentioned, whether we're talking about cross-border or globalization, our team should be where our customers are. Serving customers well is crucial. But even though we're crossing borders, going global, my clients are still based in China.

发言人1 01:27:57
Speaker 1 01:27:57

其实这个我可以补充佳睿，其实佳芮说的是这个在现阶段的确是服务好中国品牌的出海，也就是chinese brands。其实我相信肖睿，包括我们经常出国，其实你下一步的计划我们是一样的，我们一定是服务global chinese。因为我们如果去直接服务western的这些people，其实我们服务的门槛是很高的，包括这个新人的构建。所以通常可以分成三个阶段。从落地的角度来讲，当然终极目标是全球性的服务。第一步是服务中国企业的出海，第二步服务是海外华人企业，他可能就是在原生的中国面孔、华人面孔，global chinese.
Actually, I can add to Jiarui's point. What Jiarui mentioned is indeed about serving Chinese brands going global at this stage. I believe Xiaorui, including those of us who often travel abroad, share the same next-step plan—we must serve global Chinese. Directly serving Western people presents a high barrier, including the construction of new talent. Therefore, it can typically be divided into three stages. From an implementation perspective, the ultimate goal is global service. The first step is serving Chinese enterprises going global, the second step is serving overseas Chinese businesses—those with native Chinese or ethnic Chinese faces, the global Chinese.

发言人1 01:28:31
Speaker 1 01:28:31

第三步就是global，所以这是三步，我其实今年有服务过global客户。其实挺痛苦的，每天早上五点钟起来跟他痛苦，是一个老外，他对方也挺痛苦的，我也挺痛苦的。所以后面我们就说要不我们这个排名还不太合适，所以我们是可以服务的。但是还是分阶段给到自己找到一个很好的节奏比较重要。
The third step is going global. So these are the three steps. I actually served a global client this year, and it was quite painful. I had to wake up at 5 AM every morning to deal with the pain—it was a foreigner, and both sides were suffering. I was in pain, and so was he. Later, we agreed that our current ranking wasn’t quite suitable, so we can still provide services. But it’s more important to find a good rhythm for yourself in stages.

发言人1 01:28:54
Speaker 1 01:28:54

这就回到刚刚说的品牌，如果国内还是一个包袱的情况下，建议选择市场门槛。所以说市场进入拿到正反馈时间不要那么长的，东南亚也是这样的一个道理。但是终极目标是不变的，终极目标一定是global，不是仅仅的china brand go global，只是china brand go global第一步，我相信嘉睿你很快就会服务global chinese。就是我们说所谓在美国的华人企业家。
This brings us back to what we just discussed about branding. If the domestic market is still a burden, I’d recommend choosing a market with lower barriers to entry. The goal is to enter a market where you can get positive feedback quickly—Southeast Asia follows the same logic. But the ultimate goal remains unchanged: it must be global. Not just "China brand go global," but rather, "China brand go global" is just the first step. I believe, Jiarui, you’ll soon be serving global Chinese—what we call Chinese entrepreneurs in the U.S.

发言人3 01:29:18
Speaker 3 01:29:18

就是你刚刚说的那个，其实也挺惊讶到我的。那是一家独角兽中国企业，然后他们踩了个坑。其实之后也有可能我们会找来聊，就是他们一个海外的市场负责人高管，搭建了很多本土化的团队，验证了海外市场的失败。他们又让本土的高管去做全球化的业务，没有建立本土团队，反而去年做到了十倍增长，所以我觉得特别神奇。所以并不代表我有绝对的观点，就是说一定要有本土化的团队，或者不要有本土化的团队。但我觉得实践是告诉我们有这种现象。
What you just mentioned actually surprised me too. That was a Chinese unicorn company, and they stumbled into a pitfall. We might actually invite them for a discussion later—their overseas market executive, who built many localized teams, only to validate failure in the international market. Then they had local executives handle global operations without establishing local teams, yet achieved tenfold growth last year. I find it particularly fascinating. So it doesn’t mean I hold an absolute view—whether you must have localized teams or not. But practice has shown us that such phenomena exist.

发言人1 01:29:54
Speaker 1 01:29:54

对，其实我觉得未来团队构成可能都是global的。只是说在不同的这个价值链上，哪一个体系或者哪一个国家，哪一个能力上限具备优势，然后最后还是一个价值链组合。我觉得团队的构成，就像我刚刚说的，打造这个降本增效的解决方案的这个构成是有异曲同工的意思。就是谁擅长做哪一块，谁就整合合进来。但最终考虑的是企业一号位他的这个团队文化的适配度，就是你怎么从一个中国文化变成全球文化，甚至各个国家文化的一个适配。这个其实是最大的挑战，反而不是我们说战术层面上怎么影响这个市场的具体的方式。
Yes, I actually think that in the future, team composition will likely be global. It's just a matter of identifying which system, country, or capability has the upper hand in different parts of the value chain, and ultimately forming a combined value chain. The composition of the team, as I mentioned earlier, shares a similar logic with building cost-effective and efficient solutions—it's about integrating whoever excels in a particular area. But the ultimate consideration is how well the team culture aligns with the company's top leadership—how you transition from a Chinese culture to a global one, or even adapt to the cultures of various countries. That is actually the biggest challenge, rather than the tactical specifics of how to influence the market.

发言人3 01:30:28
Speaker 3 01:30:28

对对对，那个创始人他就是一个非常硅谷化的人，就是做软件理解。
Exactly, that founder is a very Silicon Valley-type person, meaning he truly understands software.

发言人4 01:30:33
Speaker 4 01:30:33

好，最后一个话题，回到你个人之前，你在国双汇亮在这些上市公司做高管，薪水也非常高。而且还有非常多的这些平台给你带来的资源。从职业发展的角度来讲，其实也算拿到了一个不小的成就。
Alright, the last topic brings us back to your personal background. Previously, you held executive positions at listed companies like Guoshuang and Huiliang, with very high salaries. Moreover, these platforms provided you with abundant resources. From a career development perspective, you’ve already achieved significant accomplishments.

发言人4 01:30:50
Speaker 4 01:30:50

其实前面在讲说我一直要挖你来句子，而且想挖了好多年，每一个阶段都在做不同的动作。我们刚认识的时候，当时你在国当做消费事业部的总经理。最开始是我把你当成一个渠道，帮我介绍了很多的客户。然后后来发现好像你的句子也更好。甚至先后有两次差点都成功了。我们都谈到了最后的阶段，然后最终你还是先选择创建的这个lip。
Actually, earlier I mentioned that I’ve always wanted to recruit you to Juzi, and I’ve been trying for years, making different moves at each stage. When we first met, you were the General Manager of the Consumer Division at Guo. At first, I saw you as a channel, and you introduced many clients to me. Later, I realized that Juzi might be a better fit for you. There were even two occasions when we almost succeeded. We reached the final stages of negotiation, but in the end, you chose to create Leap first.

发言人4 01:31:15
Speaker 4 01:31:15

到最后一个话题，我们可以聊一聊比较sharp的问题。最后一次差点成功，那天晚上我记得特别清楚，其实已经成功了。因为那天当时你老婆也在，然后你特别好的朋友也在，然后我们一起起再聊，然后庆祝你加入句子，大家喝得很开心，然后各种cheers。然后第二天醒来你就后悔了，你说了很多原因，归到最后其实你说你就是想做老大。我觉得我们看到很多创业者，其实看到创业者背后这个人。
For the last topic, let’s discuss something sharper. I remember the last time we almost succeeded very clearly—it was practically a done deal. That night, your wife was there, and so was your close friend. We were all chatting and celebrating your joining Juzi, cheering and having a great time. But the next morning, you regretted it. You gave many reasons, but ultimately, you said you just wanted to be the boss. I think what we see in many entrepreneurs is the person behind the entrepreneur.

发言人4 01:31:44
Speaker 4 01:31:44

回到人的话题，我也是创业者。虽然我的创业初衷不是想做老大，我是误打误撞赶上了一些机会。即使大家最开始创业的初衷或者目的不一样，但是有一点其实都是很相似的，包括我跟will在很多地方是很相似的人，就是不太安分的人，骨子里不太安分，想做一些不一样的事情，想创造一些新的东西，想实现自己的价值，做自己感兴趣的事情，想要极大的自由度，想看自己能成长到什么程度，你的边界在哪里。虽然我最后发你好多次失败了，现在我们用另外一种方式去合作。当时特别逗，当时view的朋友们私信跟我说，渣男接下了我的感情。
Returning to the topic of people, I'm also an entrepreneur. Although my initial intention wasn't to become the boss—I stumbled into some opportunities by chance. Even if everyone's original purpose or goal for starting a business is different, there's actually one thing that's quite similar, including how Will and I are alike in many ways—we're restless souls, inherently restless, wanting to do something different, to create something new, to realize our own value, to pursue what interests us, to have immense freedom, and to see how far we can grow and where our limits lie. Although I ultimately failed many times in reaching out to you, now we're collaborating in another way. It was quite funny at the time—friends from View privately messaged me saying, "The playboy took over my feelings."

发言人4 01:32:24
Speaker 4 01:32:24

不过站在创业者的角度来讲还是非常可以理解你为什么选择自己去创业。哪怕经历了很多刚才你说的各种循环，从低谷到高峰又到低谷，非常痛苦。每个创业者都经历的过程，但依然说我要自己去做一件事情。我自己也从学生时代开始创业，也去过大厂，其实没有待多久。所以从职业经理人的角度，其实你的视角其实更多维，完整的上市公司高管和完整的创业者。所以想听你展开聊一聊，作为一个上市公司的高管和创业者之间有什么区别，然后你创业后的感受是什么样的？
However, from an entrepreneur's perspective, it's completely understandable why you chose to start your own business. Even after going through all the cycles you mentioned earlier—from lows to highs and back to lows—it's incredibly painful. It's a process every entrepreneur goes through, yet you still say, "I want to do this myself." I also started my entrepreneurial journey during my student days and have worked at big companies, though not for long. So from a professional manager's standpoint, your perspective is actually more multidimensional—having been both a full-fledged executive at a listed company and a full-fledged entrepreneur. I'd love to hear you elaborate on the differences between being an executive at a listed company and being an entrepreneur, and what your feelings have been since starting your own business.

发言人1 01:33:01
Speaker 1 01:33:01

我觉得区别还是挺大的。首先说一下为什么创业，包括刚刚佳瑞说到我们历史的这些缘分，我觉得还是那句话，从长周期来讲，其实我们会有更深的缘分。从短周期来讲，为什么选择创业？是因为在职场上我经历了从我应该是siri加入的我前公司，然后一直到中再到退市，我经历了完整的这个周期。但是我没有经历过从种子到C其实有点像是拼图的意思。就是我想我既然经历过从C到IPU到退市，其实退市的过程中也是挺难受的，即使在职场里面。所以我自己想把这个拼图拼完。
I think the difference is quite significant. First, let me talk about why I chose to start a business, including the historical connections Jiarui just mentioned. I still believe that, from a long-term perspective, we actually share deeper bonds. From a short-term perspective, why did I choose entrepreneurship? It's because, in my career, I experienced the entire cycle—from joining my previous company when Siri was acquired, all the way through to its delisting. However, I hadn't experienced the journey from seed stage to Series C—it felt like completing a puzzle. Since I’d already gone through the process from Series C to IPO to delisting—which was quite painful, even within a corporate setting—I wanted to complete the puzzle myself.

发言人1 01:33:40
Speaker 1 01:33:40

然后第二个就是回到边界感。因为我不知道如果自己去参与做这件事情，他会经历什么。所以我当时在思考一个下限，就比如说我们花了三年的时间，最后我们荒废了青春，然后也没有任何的收益，我能不能接受？如果我能接受，那这件事情我还是想试一下，它会是我人生的一个平台。因为如果这个时候我再不跳出来，我选择再继续在大厂里面，我可能第一会错过这波AI带来的一些线上级的机会。第二个可能会错过我创业最佳的这个时间段。因为可能再过几年家庭上面的考量，还有包括可能在大厂再经营几年的人脉的积累，会让我的跳出这个既定的循环的门槛会更高。所以这个是我觉得已经到这个时候了。
The second point relates to boundaries. Because I didn’t know what I would experience if I took the leap into entrepreneurship, I was thinking about the worst-case scenario. For example, what if we spent three years with no results, wasted our youth, and gained nothing? Could I accept that? If the answer was yes, then I still wanted to give it a try—it would become a platform for my life. If I didn’t take the leap now and chose to stay in a big company, I might, first, miss the massive opportunities brought by AI, and second, miss the optimal window for entrepreneurship. A few more years might bring family considerations and deeper professional networks, making it even harder to break out of the established cycle. So, I felt the time was right.

发言人1 01:34:26
Speaker 1 01:34:26

然后回到不是选择参与创业，而是主导创业这个过程。还是回到我刚刚说的那件事情。从两个维度来看，就是我觉得在大厂经历这个上市，包括也管了一些事业部，这个过程这个我已经经历过了。其实我想经历过的是一个打造一个我不知道会是什么东西的一个过程。所以这个体验其实是参与创业未必能带来的。
Then returning to the process—not just choosing to participate in entrepreneurship, but leading it. It still comes back to what I mentioned earlier. Looking at it from two dimensions: I feel that going through an IPO at a major company, including managing some business units, is something I've already experienced. What I really want to experience is the process of building something whose outcome I can't predict. That’s an experience that merely participating in a startup might not necessarily bring.

发言人1 01:34:48
Speaker 1 01:34:48

但现在我可能观点会不太一样，现在我已经不太想把那句话变成说我一定要当老大。我觉得那是两年前非常不成熟的一个话。我现在其实是我比较喜欢道德经，就是像水一样。我核心是我在生态位上能不能帮助客户拿到更好的效率。所以这个拿到效率的方式可以是我参与创业，可以是我自己创业，也可以是我在大厂。但终极目的还是更疯狂的创造价值的这个过程。如果创业我当老大的形式，我没有在疯狂创造价值，我也是会被淘汰的。然后最终我淘汰之后，我可能还是会参与创业，或者还是会回到大厂。
But now my perspective might be a bit different. I no longer feel the need to insist on being the boss. I think that was a very immature thing to say two years ago. Nowadays, I’m quite drawn to the Tao Te Ching—the idea of being like water. My core focus is whether, in my role, I can help clients achieve greater efficiency. The way to achieve that efficiency could be through participating in a startup, founding my own, or working at a major company. But the ultimate goal is still the process of creating value more intensely. If, as the boss of a startup, I’m not creating value intensely, I’ll still be 淘汰 ed. And after 淘汰 ion, I might still end up participating in a startup or returning to a major company.

发言人1 01:35:22
Speaker 1 01:35:22

所以如果你回归本质的话，其实市场需要的是更高的效率，而不是我自己需要的。然后再回到心态上的变化，落差非常大。曾经有太多次想放弃，以及说是后悔说是不是现在如果还在这个大厂的话，享受着这个比较好的酒店，享受的比较好的带薪出差，享受的我们不能说颐指气使，但是享受着这种指指点点，指点江山，谈笑风生。当然那个感觉是不一样的。
So if you return to the essence, what the market truly needs is higher efficiency, not what I personally need. Then, coming back to the shift in mindset, the contrast is enormous. There were so many times I wanted to give up, even regretted it, wondering if I’d still be at that big company now, enjoying the nicer hotels, the paid business trips, and—well, I wouldn’t say bossing people around, but enjoying the sense of authority, pointing things out, discussing strategies, and chatting freely. Of course, the feeling was entirely different.

发言人1 01:35:50
Speaker 1 01:35:50

但现在回来看这件事情，就是我觉得创业是没有终极的游戏，然后职场更像是一个有限游戏。这是第一个不一样的地方。就是终局游戏就一意味着永远没有结束的时候，你每天起来，比如说今天我们的对话就是一场游戏。因为这场对话既是关乎于我们的对业务的讨论，也是关乎于我们怎么样看待我们自己的企业。其实佳瑞你也会有自己新的感受，我也会有我新的感受。它是一个中游戏。
But looking back now, I think entrepreneurship is an infinite game, while the corporate world is more like a finite game. That’s the first key difference. An infinite game means it never truly ends. Every day you wake up—for example, today’s conversation is a game in itself. Because this dialogue is both about our business discussions and how we view our own company. In fact, Jiari, you’ll have your own new insights, and I’ll have mine. It’s a medium game.

发言人1 01:36:14
Speaker 1 01:36:14

然后第二，游戏的规则变了。职场的时候游戏的规则是已经被定义了。你只要拿到满分，就像你考试一样，就会有人激励你奖励你。但是现在的规则是自己去定义的，甚至这个规则是你跟全球的各个价值创造者碰撞出来的。
Secondly, the rules of the game have changed. In the workplace, the rules were already defined. You just needed to score full marks, like in an exam, and someone would motivate and reward you. But now, the rules are defined by yourself—or rather, they emerge from collisions with value creators worldwide.

发言人1 01:36:32
Speaker 1 01:36:32

就像佳瑞，我跟你的对话可以碰撞出一个新的解决方案，它的可塑性太强。如果是全球的家瑞，全球像你这么优秀的创业者。当这个事情变成全球的时候，你会发现它的可能性会变得非常大我举个很简单的例子是我创业带来前所未有的一种兴奋感。比如说我们在做美的的时候，我们会涉及到很多场景，要打造成卡哇伊，然后卡哇伊你是要买一些毛绒玩具的。而刚好名创给我们派了几千个毛绒玩具，我们找了一个小熊，我们用小熊的毛绒玩具放在美的的烤箱，结果这个视频超过了300万的播放。
Take Jiarui, for example. Our conversation could spark a brand-new solution—the plasticity is immense. Imagine Jiaruis globally, entrepreneurs as outstanding as you worldwide. When this scales globally, the possibilities become enormous. Let me give you a simple example of the unprecedented excitement I’ve felt while building my business. For instance, when we worked on Midea, we had to create many scenarios, aiming for a kawaii aesthetic—which meant buying plush toys. Coincidentally, Miniso sent us thousands of plush toys. We picked a little bear and placed it inside a Midea oven. The resulting video surpassed 3 million views.

发言人1 01:37:06
Speaker 1 01:37:06

这就是有点像是不同的要素之间组合在一起的时候，它的迸发的魅力是你想象不到的。因为它的规则已经变了。然后这个再释放到我所做的事情，为什么会做跨境电商？包括做全球性的这个事情。因为我非常享受东西方的这个要素不确定性的组合，就是我很享受无限游戏和无限规则的无限循环。
It's somewhat like when different elements combine, the explosive charm is beyond your imagination. Because the rules have already changed. Then applying this to what I do—why engage in cross-border e-commerce? Including this global endeavor. Because I truly relish the unpredictable combination of Eastern and Western elements—I deeply enjoy the infinite loop of infinite games and infinite rules.

发言人1 01:37:29
Speaker 1 01:37:29

所以这个过程中我也不知道我会经历什么。可能它会带来财富，可能它会带来很大的失意。但是最有意思的地方是我有一个非常确定性的观点是，即使最后我们亏的这个裤衩都掉了，我们拿到手上的认知，还有我们的能力的锻炼，是在任何一个大厂乃至参与创业的稀缺精神、稀缺能力。所以当这个不确定的点摆在中局前面的时候，你就无所畏惧了。
So in this process, I don’t even know what I’ll experience. It might bring wealth, or it might bring great disappointment. But the most interesting part is that I hold a very certain belief: even if we end up losing everything, the knowledge we gain and the skills we hone are rare qualities—whether in any major corporation or in the entrepreneurial spirit. So when this uncertainty stands before the mid-game, you become fearless.

发言人1 01:37:58
Speaker 1 01:37:58

所以回到现在来看，创业这件事情当然是如履薄冰。从很骄傲觉得这个还搞不定。第一年我觉得这个还搞不定，就是会非常王之蔑视，就是觉得什么事情都能搞定。再到极其的谦卑，极其的自卑。再到现在进入到一个在自己的专属领域非常的自信，然后在自己不了解的领域非常的敬畏，就形成了一个相对的平衡。我觉这是这个创业这段历程教给我的。
Looking back now, entrepreneurship is indeed like walking on thin ice. From initially feeling overly proud and thinking "this can't be that hard" - in the first year I had this supreme confidence, feeling like I could handle anything. Then shifting to extreme humility and even self-doubt. Now I've reached a point where I'm very confident in my specialized domain, while maintaining deep respect for areas I don't understand, achieving a relative balance. I think this is what the entrepreneurial journey has taught me.

发言人1 01:38:24
Speaker 1 01:38:24

然后另外就是一个永远是空杯心态。就是你接触的人，包括今天跟两位这么有心的对话，这种新的传播形式，播客的形式，其实他都会给我新的一个启发。所以我很享受一种原子级的元素的不确定性的这个随机组合迸发出的新的创意带来的可能性。但是因为我的这个MBTI是INTP，就是天马行空。但是我又想我现在做企业的时候，希望这种无序的随机性架在一个有限的商业价值规则下面。也就是说我们可以保证创意的无限迸发和商业的有限速赢中间找到一种平衡。
Another lesson is maintaining an empty-cup mindset. Every person you meet, including today's thoughtful conversation with you two, and new communication formats like podcasts - they all bring fresh inspiration. I truly enjoy the possibilities that emerge from the random combination of atomic-level elements of uncertainty sparking new creativity. But since my MBTI is INTP - the visionary type - while running a business, I want to channel this chaotic randomness within the framework of limited commercial value rules. That is to say, we need to find a balance between unlimited creative bursts and achieving limited quick wins in business.

发言人4 01:39:05
Speaker 4 01:39:05

我觉得回到刚刚我说我们就想当老大这件事，其实它只是一个玩笑。为什么这么说？因为其实我们看到底层逻辑，其实还是回归到价值创造。刚才VO也在讲说到底他能不能控制这个价值创造，他能不能决定说出海就要怎么去做。因为其实之前确实我一直在想做出海，但是我自己的出海也不够有决心，包括当时我们聊下来的时候也一直在说我国内的包袱太重。所以其实是没有办法控制说，到底有多少的精力能够all in到出海这件事情上。
I think going back to what I just said about wanting to be the leader, it was actually just a joke. Why do I say that? Because fundamentally, what we see is that it all comes down to value creation. Just now, VO was also talking about whether he can control this value creation, whether he can dictate how to approach going global. The truth is, I’ve been thinking about going global for a while, but I haven’t been fully committed to it myself. Even during our discussions, we kept mentioning how burdened I am by domestic responsibilities. So, realistically, it’s impossible to control how much energy I can fully dedicate to going global.

发言人4 01:39:38
Speaker 4 01:39:38

直到今天依然这在我们看来也是一个问号，我们也在持续的思考努力中，所以回归到其实作为一个创业者，核心还是在于思考我到底能够创造什么样的价值。然后在创造价值里边，我能够主导决定性的是什么，我自己独特性的价值创造的优势在哪里？另外就我刚才聊到的，其实包括我自己去创业，今年是我创业的第十年。
Even today, this remains a question mark in our eyes, and we’re still continuously reflecting and working on it. So, coming back to the essence of being an entrepreneur, the core lies in thinking about what kind of value I can actually create. Within that value creation, what can I decisively lead? Where does my unique advantage in value creation lie? Additionally, as I mentioned earlier, this includes my own entrepreneurial journey—this year marks my tenth year in entrepreneurship.

发言人4 01:40:03
Speaker 4 01:40:03

回归到个体来看，其实创业给我们带来最终的可能不是财富的积累，但是个人成长上一个巨大的积累，而这个积累是远胜于财富的。为什么刚刚讲到我说跟leo也是同一种人，其实我们最终经历过这些，可能大家希望追求的还是个人成长。在个人成长之上，可能会带来一些财富的，或者是其他的一些附属品，它自然而然都会过来。但是可能初心还是蛮重要的一件事情。当然非常好的一点是我们最近又开始新的一个合作，并且在出海上也有了一个蛮大的一个进展。一直在路上。
Looking at it from an individual perspective, entrepreneurship ultimately may not bring us wealth accumulation, but it does bring tremendous personal growth—an accumulation far more valuable than wealth. That’s why I mentioned earlier that Leo and I are the same kind of people. In the end, after going through all this, what we truly seek is personal growth. On the foundation of personal growth, wealth or other byproducts may naturally follow. But staying true to our original intentions remains crucial. Of course, it’s also great that we’ve recently started a new collaboration and made significant progress in our overseas ventures. The journey continues.

发言人3 01:40:41
Speaker 3 01:40:41

好，感谢。唯有今天约了可能有一年的时间，一年多的时间，然后我们终于对上话了。而且其实每次对话都有一些不一样的地方，我觉得其实也都是大家在成长的一个过程中，在重新塑造自己，也在重新塑造自己做的这个事情的这个过程中。但我始终我觉得是一个特别有能量，能够给别人带来能量，带来激情，也能够从中学习到很多的这么一个角色。我们也期待之后能有机会更多的跟我交流，包括嘉睿和vivo这边的合作，也希望更上一层楼。如果大家觉得这期播客有打动你的地方，或者觉得有意思的地方，欢迎分享给你的朋友。也欢迎大家在苹果podcast、小宇宙、spotify、喜马拉雅等平台订阅和收听我们的播客。感谢大家这期的收听，我们就到这边，大家拜拜。
Alright, thank you. It’s been about a year—over a year, actually—since we first planned this, and we’ve finally had this conversation. And every time we talk, there’s something new. I think it’s a reflection of everyone’s growth, of reshaping ourselves and what we’re working on. But throughout it all, I’ve always felt that this is a role full of energy, one that inspires others, brings passion, and offers so much to learn from. I look forward to more opportunities to exchange ideas in the future, including further collaboration with Jiarui and Vivo, and hope we can take things to the next level. If this episode resonated with you or you found it interesting, please share it with your friends. You’re also welcome to subscribe and listen to our podcast on platforms like Apple Podcasts, Xiaoyuzhou, Spotify, and Himalaya. Thanks for tuning in this time—that’s all for now. Goodbye, everyone.

发言人4 01:41:31
Speaker 4 01:41:31

我们再见，拜拜。
See you later, bye-bye.

发言人2 01:41:38
Speaker 2 01:41:38

Happy undergo. Everybody here want. We and. It's been waited for york.

发言人2 01:41:57
Speaker 2 01:41:57

Welcome to do you has been waited for you to. I could tell this be. I like so bright, but they never blind me.

发言人2 01:42:32
Speaker 2 01:42:32

We we first struct our bags on apartment floor, put. Everyone, do you want boys and boys and girls and girls?