跳至主要内容

Tencent Q1 2023 Earnings Call Transcript

  • Advertising on Weixin properties, which contribute over half of our total advertising revenue, outgrew the overall ad business.
    微信广告占据腾讯广告的半壁江山,万万没想到的

  • So, on the game business, not to be pedantic, but we feel very strongly that what you referred to as old games are, in fact, evergreen games. And if you look at our game portfolio, many of the biggest, most successful titles, including Honour of Kings, League of Legends, CrossFire, CrossFire Mobile, Peacekeeper Elite, PUBG Mobile, Arena Breakout, are competitive games, often team-based competitive games that are designed to be balanced, fair, and playable for years or for decades.
    培育常青游戏

  • You know, players responded to that appealing content by investing, in some cases, more time and, in other cases, more money or both. And the more money translates into more gross receipts and, therefore, more revenue. In terms of whether there were special promotional campaigns, you know, we always promote the virtual outfits. I think, over time, we've become progressively better at promoting the right virtual outfits at the right time of the calendar year.
    So, the business has become more akin to luxury goods. And that there are certain virtual outfits we promote to tie into the Lunar New Year. There are others we promote tie into the May Day holidays, into the summer period, and so on and so forth. But that's an ongoing trend as opposed to a single-quarter phenomenon.
    洞见,游戏皮肤,仔细想想确实是像奢侈品一样的生意,但差异点在于奢侈品在低成本+高单价,而游戏皮肤是低成本+低单价(但同时有极多的COPY走量)

  • As you know, cloud only represents a mid-single-digit percentage of our total revenue. You know, within that, the price cuts only apply to infrastructure as a service, which is a subset of that mid-single digits. And then even within infrastructure as a service, if you look at the price cuts that have been widely announced in the industry in recent weeks, they apply to long-term prepaid contracts, and those long-term prepaid contracts are generally entered into by small- and medium-sized businesses, but not by larger enterprises. So, you know, this is an industry where, over time, the input costs of servers and bandwidth and so forth are generally falling and it makes sense that the participants in the industry pass on those falling input costs into falling pricing to their customers.
    云的这一轮降价,是基于近几年服务器与带宽综合成本下降,带来的定价端的下降,是行业正常现象,不是亏钱争夺市占率

  • And then in terms of applications, I think there will be a lot of different applications. Overall, the user-to-machine interactions, in particular, search and information generation, and those type of applications would actually be more disruptive. And at the same time, there will be more of a rush of applications appearing in those use cases. So, general search, vertical search, in terms of content aggregation, a lot of these kind of applications would see new disruptors coming into play.
    人──>机器交互领域,AI大模型的冲击会更大

  • So, thank you for the question, Charlene. The competitive landscape for the acquisitions of game companies outside China has been fairly stable for the past few years. And while there are a number of companies that are acquisitive in this space, often their objectives do not overlap with our objectives, meaning they may be companies with a big console installed base and they're trying to acquire game studios to supercharge that console installed base, or they may be companies which are valued on near-term earnings and they're trying to bolt on additional earnings, versus in our case, we're not married to any single hardware device. And so, we've been acquisitive with mobile game studios, acquisitive with PC game studios, periodically active with console game studios as well.
    And in addition, we're very willing to look forward three or five or seven years when we do these investments in acquisitions. And so, it's often the case that we're looking at a company that has released a successful niche game the previous year, and we know for a certainty that that company's revenue will be declining for the next couple of years because it will be digesting the success of the first game. And in many cases, we know the company will be loss making for a period of time because we're scaling up the team to work on a much bigger sequel, but we don't get to have the bigger revenue from the bigger sequel, and, you know, that fits us. It doesn't fit everyone.
    专注于游戏内容本身而不是游戏硬件;游戏工作室的收购基于后续的游戏内容

  • Well, in terms of the revenue position, we won't be super specific, but Weixin, as we said, is the majority. Within that, Moments is historically the largest contributor. Mini programs and official accounts are substantially smaller and roughly similar to each other. And video accounts have sort of leapfrogged over mini programs and official accounts because there's a lot of inventory in video accounts, and that's very high revenue per 1,000 impressions in video accounts.
    You know, outside Weixin, then the biggest set of properties is the mobile ad network, which has had quite a pronounced bounce back in the last six to nine months. And then beyond that, there's a drop off to the long-form video, music, news, and some smaller properties around QQ and Qzone. So, that's the composition. In terms of the margin profile, then basically, all of the owned and operated properties, with the exception of long-form video, are high gross margin, versus long-form video and the ad network are inherently much lower gross margin.
    ──原来朋友圈广告这么赚钱......

  • So, on games, I'm not necessarily going to answer your question because we don't provide guidance as to where the revenue lines are accelerating or decelerating beyond observing that, firstly, the strength in the first quarter was partly due to what I referred to as evergreen games, but we also saw very good growth from a number of games released in the last one to two years. And it's often the pattern with us that when the quarter we release a game, it generates relatively little monetization. And of course, that monetization doesn't translate into our P&L because of our deferral policy. But then, over subsequent quarters, if it's a good game, we see the monetization improving, we see the revenue -- the deferral starting to flow through into reported revenue.
    If it's a really good game, we see the daily active user count increasing. And so, recently, a number of games have fit that profile, including the Arena Breakout, which is the leading extraction shooter game in China, where we've seen good growth in DAU and in monetization as players become familiar with this new type of game that didn't exist before. We've seen very good growth in terms of both users and monetization from the auto chess game, Golden Spatula, that has over 10 million daily active users, which is, you know, rare for a new game in China and very rare for a game to build to that level over the course of two years as opposed to it gets that level very quickly and then fade away. And also quite good trends in the first quarter for our League of Legends: Wild Rift mobile game.
    Now, obviously, looking through the rest of this year, we have a number of big games that we intend to release domestically that we're excited about. But again, with some of those games, they'll monetize well from day one. With others, they'll monetize more gradually over time. Thank you.
    所以游戏这里的思路是:不一定在游戏刚推出时就拼命获利,有些游戏如此,有些游戏是先走DAU增长再获利

  • So, in terms of the big e-commerce companies advertising with Tencent, then we're certainly experiencing share gains. I think that those share gains arise for a number of reasons. One is changing perceptions about sort of ecosystem competition, meaning that in the past, there were some companies that chose not to advertise with us because they viewed us as an ecosystem rival, and they may have reassessed that perspective over time.
    And then secondly, we're deploying much more sophisticated machine learning now to our advertising targeting that is particularly beneficial for big e-commerce companies with gigantic range of SKUs because now we can ingest those SKUs and then display the right SKU to the right user within the e-commerce companies' advertisements. So, that's on the e-commerce companies advertising with us.
    ──朋友圈广告直跳淘宝

  • But in none of those do we need to spend a great deal of extra money to make them more popular. You know, we aren't incurring gigantic subsidies on providing community group buying food solutions. We aren't inducing users who would otherwise be on another short video service to install video accounts and spend time on video accounts instead. You know, we're just providing what we think are very attractive and somewhat differentiated experiences to users within our big apps, particularly Weixin, and letting users -- over time, you discover those, enjoy those, socialize those, and spend more time and ultimately more money on those.
    So, I don't think that we're in a sort of nongrowth mode and, you know, at some point in the future, we need to flick a switch and dramatically increase sales and marketing or subsidies to go back to growth mode. I think we're in growth mode right now. It's just that the nature of our growth drivers is that they do not require heavy subsidies, they do not require heavy sales and marketing activity. Thank you.
    值得向微信学习如何踏踏实实做产品
    不需要大额成本支出的增长,真是非常美妙!

 

Popular posts from 产品随想的博客

产品随想 | 周刊 第116期:Great things in business are never done by one person.

Cromite   https://github.com/uazo/cromite Cromite a Bromite fork with ad blocking and privacy enhancements; take back your browser! awesome-shizuku   https://github.com/timschneeb/awesome-shizuku Curated list of awesome Android apps making use of Shizuku KernelSU   https://github.com/tiann/KernelSU A Kernel based root solution for Android Love, Hate or Fear It, ​​TikTok Has Changed America   https://www.nytimes.com/interactive/2024/04/18/business/media/tiktok-ban-american-culture.html?unlocked_article_code=1.mE0.DpEZ.VWmNssw5B6_c "My model for business is The Beatles.There were four guys who kept each others, kind of, negative tendencies in check. They balanced each other, and the total was greater than the sum of the parts. And that's how I see business. You know, great things in business are never done by one person. They're done by a team of people. "Our motivation is simple--we want to provide the most advanced and innovative platform to our developers, and we wa...

产品随想 | 读《中国是部金融史》:第一章 此朝无钱胜有钱(周朝)

第一章 此朝无钱胜有钱(周朝) 姬旦选择了当时最有效、最简单的统治方式—一封建:为确保对国家的统治,姬且共封七十一国,其中姬姓五十三国,封国在地方替周王管理属国。 封建封建,封而建之。 对权力没有任何制约的西周,“王德〞是封建统治者最后的底线。这种“王德”对天命的畏惧延续了周王朝近千年的生命,此后,无论春秋五霸如何强盛,都只能挟天子以令诸侯,再强的诸候都必须获得王室认可。这一点,颇类似于西欧或者日本王位干年一系,绝不是“皇帝轮流做,明年到我家”的强盗逻辑。 司马迁之所以大骂周厉王,原因还在于周厉王推行了“专利”制度。其实,西周铭文甚至后世典籍始终都没说清楚究竟周厉王的“专利”是个什么东西,只是说他“专山林川泽之利”。周厉王很可能把诸候的山林川泽收归国有(自己所有),凡是在山上砍柴的、打猎的都要向王室纳税。看起来,周厉王敛财是确定无疑的,所以,司马迁认为此人是一个暴君。不是周厉王贪婪,而是王室实在太穷了。 西周王室失去了财富、失去了王德,也很快会失去江山。 失去“王德”的社会,人们只会臣服于强人,只有利益才是真正的信仰。也就是说, 这个时候的西周几乎是一种丛林社会,只有赤裸裸的暴力才能制约这些诸候了。对诸侯来说,周总王不过就是一个名号,没-起喝过酒、没一起拼过命,凭什么要听你的命令? ──复用下作者的句式:失去“法律”的社会,人们只会沉醉于拜金。 管仲告诉我们,强国、弱国最大的区别是强国的钱都在黎民百姓手里,弱国的钱却在国王的钱箱子里。这一点,颇类似于穷国和富国的区别,穷国只想抢老百姓的钱,富国却在想养如何让老百姓赚钱。 ──思考:我们的钱,在哪里? 周幽王死后,周平王为了躲避犬我,东迁至东都洛邑,中国开始了春秋时代。但无论春秋诸侯多么强盛,都要举起周王室大旗,所以,春秋在中国历史上也被称为“东周” 真正的强国,震慑敌闲的不只是万乘之师,还有无法超越的经济实力。强国经济之强,必然源自园内公平的经济环境,能为国民提供一个自由创造的空间。即使弱国拒绝承认强国的地位,甚至试图与强国平起平坐,也永远无法护绝强国对本土的经济渗透。 ──瞬间想到美国 无论弱国多么不情愿,强国货币都一定会流进他的地盘。在古代,一种曾经广泛流通的货币必然有大量文物存世,春秋货币则主要包括布币、刀币和蚁鼻钱三个体系。在中国史籍中经常提到“春秋五霸”,不过“春秋五霸”有很多种版本。顺着本书的逻辑, 我们...

《Becoming Steve Jobs》Chapter 15 The Whole Widget

What the world did see was an effective and visionary leader at the height of his powers. These were complicated years for Apple, but Steve handled almost every challenge in exactly the manner he wanted. He had fallen into leadership at such a young age, but he was comfortable in that role now, and justifiably sure of his capacity to guide Apple’s tens of thousands of employees to the goals he set for them. During these years, he would ensure the company’s continued success in personal computers by engineering a deft switch to a new kind of microprocessor; ruthlessly and successfully managing some major transitions in his executive team; and optimizing and building upon the efficiency and ambition of the company’s product development “treadmill,” as Tim Cook describes it. This is also when he delivered what is likely to be remembered as the most notable product of his life, the iPhone, and then improved even that by pivoting once again into a strategy he personally had not wanted to pu...

产品随想 | 周刊 第117期:He saw the intersection of art and science and business and built an organization to reflect that.

He disliked biography attempts. “I regard my scientific papers as my essential biography,’’ Land said. “I pour my whole life into the scientific project I’m investigating. I leave behind the things I’ve done in the past to do the work in the present.’’ “The purpose of inventing instant photography was essentially aesthetic,’’ Land said in 1947, announcing the process’s invention. “We live in a world changing so rapidly that what we mean frequently by common sense is doing the thing that would have been right last year.” — Edwin Land, Statement to Polaroid Corporation employees (25 June 1958) The worldview he was describing perfectly echoed Land’s: “Market research is what you do when your product isn’t any good.” And his sense of innovation: “Every significant invention,” Land once said, “must be startling, unexpected, and must come into a world that is not prepared for it. If the world were prepared for it, it would not be much of an invention.” Thirty years later, when a reporter ask...

Hacking Team武器库

之前百度盘上公开了部分Hacking Team工具,自己一直没有看懂,今天在泉哥博客上看到他对于武器库的几篇文章,mark一下

产品随想 | 周刊 第51期:Never let a good crisis go to waste

Products Paperless-ngx   https://github.com/paperless-ngx/paperless-ngx A community-supported supercharged version of paperless: scan, index and archive all your physical documents 自架设服务,文档聚合 Tube Archivist on YouTube   https://github.com/tubearchivist/tubearchivist Your self hosted YouTube media server 自托管YouTube流媒体播放 Emby Server Emby Server is a personal media server with apps on just about every device. 自己掌控流媒体 Pointless   https://github.com/kkoomen/pointless An endless drawing canvas desktop app made with Tauri (Rust) and React 无限画布白板工具,Tauri构建,跨多端 PWA LIST   https://www.pwalist.app/ 一些好玩的PWA应用,有些还不错 Pomofocus 番茄钟 Song Search “Find me a song by lyrics.”   https://songsear.ch/ Nanopi Openwrt   https://github.com/klever1988/nanopi-openwrt Openwrt for Nanopi R1S R2S R4S R5S 香橙派 R1 Plus 固件编译 纯净版与大杂烩 Project ImmortalWrt   https://github.com/immortalwrt/immortalwrt An opensource OpenWrt variant for mainland China users. China用户专用......心情复杂 YAOF   ht...

李录推荐阅读书单

李录推荐阅读书单 李录在《文明、现代化、价值投资和中国》的最后,列出了他推荐阅读的一些书目。这个书目的含金量非常之高,是培养一个多层次有深度思维很好的参考,特整理如下。 一. 科学、哲学、进化、人类文明史、人类历史 1. 《枪炮、病菌与钢铁:人类社会的命运》,贾雷德·戴蒙德 2. 《西方将主宰多久》,伊恩·莫里斯 3. 《文明的度量:社会发展如何决定国家命运》,伊恩·莫里斯 4. 《群的征服》,爱德华·奥斯本·威尔森 5. 《无穷的开始:世界进步的本源》,戴维·多伊奇 6. 《真实世界的脉络:平行宇宙及其寓意》,戴维·多伊奇 7. 《理性乐观派:一部人类经济进步史》,马特·里德利 8. 《科学发现的逻辑》,卡尔·波普尔 9. 《开放社会及其敌人》,卡尔·波普尔 10. 《自私的基因》,理查德·道金斯 11. 《人类简史:从动物到上帝》,尤瓦尔·赫拉利 12. 《文明》,尼尔·弗格森 13. 《当下的启蒙》,史蒂芬·平克 14. 《心智探奇:人类心智的起源与进化》,史蒂芬·平克 15. A history of knowledge, Charles Van Doren 16. 《神的历史》,凯伦·阿姆斯特朗 17. 《为什么佛学是真实的》,罗伯特·赖特 18. 《思考,快与慢》,丹尼尔·卡尼曼 19. Creating the Twentieth Century, Vaclav Smil 20. Transforming the Twentieth Century, Vaclav Smil 二. 中国文明、历史、文化 1. 《先秦诸子系年》,钱穆 2. 《中华文化十二讲》,钱穆 3. 《史记(白话本)》,司马迁 4. 《白话二十五史精选》,李解民等 5. 《四书章句集注》,朱熹 6. Waiting for the Dawn, William Theodore de Bary 7. 《中国的自由传统》,狄百瑞 8. 《万古江河——中国历史文化的转折和开展》,许倬云 9. 《黄宗羲全集》 10. 《余英时文集》 11. 《思想和人物》,林毓生 12. 《曾国藩全集》 13. 《万历十五年》,黄仁宇 14. 《天安门:知识分子与中国革命》,史景迁 15. The Search for Modern China, 史景迁 16. 《中国官僚政治研究》,王亚南 17. 《中...

《深度学习推荐系统》By 王喆

  推荐序言 深度学习/机器学习,爆发于2015年 算法能力+工程能力 1992年,施乐创建了应用协同过滤算法的推荐系统 本书for:推荐、广告、搜索领域的从业者/爱好者 第1章 互联网的增长引擎──推荐系统 不同业务模式的公司定义的具体推荐系统优化目标不同,例如,视频类公司更注重用户观看时长,电商类公司更注重用户的购买转化率(Conversion Rate,CVR ) YouTube推荐系统的主要优化目标就是观看时长,而不是CTR 工程论文:Deep Neural Networks for YouTube Recommenders 大致流程:构建深度学习模型,预测用户观看某候选视频的时长,按照预测时长进行候选视频的排序,形成最终的推荐列表 淘宝首页:千人千面(即使是1%的CTR提高,也会到来数十亿元收入增加) 推荐系统逻辑框架:f(U,I,C), User, Item, Context 推荐系统数据部分:实时数据、准实时数据、离线数据 推荐系统模型部分:召回层、排序层、补充策略与算法层(排序层是工业界与学术界研究的核心) 第2章 前深度学习时代──推荐系统的进化之路 UserCF: 基于用户相似度进行推荐的协同过滤算法 互联网用户数的增长会导致矩阵维护复杂度急剧上升 用户的历史数据可能稀疏(导致UserCF不适用于酒店预订、大件商品购买等低频应用) 适用于发现热点、以及跟踪热点的趋势 ItemCF:基于物品相似度进行推荐的协同过滤算法 更适用于兴趣变化较为稳定的应用 协同过滤的天然缺陷:推荐结果的头部效应较明显 其它模型:矩阵分解、逻辑回归、FM、FFM、GBDT+LR、LS-PLM (细节已经要看不懂了) 第3章 浪潮之巅──深度学习在推荐系统中的应用 Google Play的推荐团队,基于对业务场景的深刻理解,提出了Wide&Deep模型 这张属于看不懂那种类型 第4章 Embedding技术在推荐系统中的应用 Not Interested 第5章 多角度审视推荐系统 从7个不同的角度切入推荐系统 推荐系统如何选取和处理特征 召回层的主要策略有哪些? 实时性的重要性体现在哪里?有哪些提高实时性的方法? 如何根据具体场景构建推荐模型的优化目标 如何基于用户动机改进模型结构 冷启动问题的解决方法有哪些 什么是探索与利用问题?有哪些主流的解决方法 第...

Bilibili Q2 2023 Earnings Call Transcript

Moving on to our community growth. Our DAUs increased by 15% year-over-year to 96.5 million in the second quarter, driving our DAU to MAU ratio up to 29.8%. In the second half of this year, we plan to shift our video watching metric from video views to video time spent, which will help our users discover more high-quality content on Bilibili. In addition, we are exploring new live broadcasting products to create more paying channels for our users, including live celebration events. We expect our ongoing integration activities to support our revenue growth in the second half of 2023. Our top five advertising verticals in the second quarter were games, e-commerce, digital products and home appliances, skincare and cosmetics, and food and beverage. 五大Top广告垂类 In the meantime, we have seven new titles that we plan to release in domestic and overseas markets in the second half of the year, including the highly anticipated game Pretty Derby 期待7款游戏能红火一波 We cut sales and marketing expenses by 2...

产品随想 | 陪读《芯片战争》:一、二章

作者其它书也值得读读 已出版《食用油营销第1书》《金龙鱼背后的粮油帝国》《鲁花:一粒花生撬动的粮油帝国》《手机战争》等财经书籍。 第一章 从晶体管到芯片 肖克利要创业的消息,就像17世纪的牛顿宣布要建立工场一样引起了轰动。那时候,美国还在草创时期的半导体产业主要集中在东部的波士顿和纽约长岛地区,许多科学家和工程师精英都慕肖克利之名前往美国西海岸,不远千里地聚集在肖克利旗下。可以这么说,肖克利给硅谷带来了最初的火种。 ——我也認可這個說法,不是斯坦福或其他校園,純粹是人才的原因 芯片发明后的六年间,政府对芯片项目的资助高达3200万美元,其中70%来自空军。同期美国半导体产业的研发经费有约85%的比例来自政府,政府的支持成就了美国在半导体领域的技术优势。“华盛顿通过支付技术研发费用和保证最终产品的市场份额,将原子弹最终制造成功的间隔缩短至六年,晶体管缩短至五年,集成电路缩短至三年。”不过,这些半导体企业实力壮大以后,往往不愿再参与美国政府出资的研发项目,因为那意味着专利权归政府所有,而且还得受保密条款的约束。 ——這纔是更合理的產業支持政策 诺伊斯采用激进的价格政策,将主要芯片产品的价格一举降到1美元,不仅是市场上的主流芯片价格的零头,还低于当时芯片的成本。这不是传统意义上的亏本倾销,而是第一个以反摩尔定律为定价依据的案例。反摩尔定律认为,同样的芯片在18个月后价格就会跌一半,所以按照几年后的价格为当前的芯片定价是有一定合理性的。市场被迅速打开,芯片很快在民用市场得到越来越广泛的应用,仙童的营收和利润都迅速上升,还带动了其母公司的股价上涨。摩尔后来评论:“诺伊斯以低价刺激需求,继而扩大产能、降低成本的策略,对于芯片产业的发展而言,其重要性堪比芯片的发明。” ——看到這裏的時候,真是非常震撼,好厲害的定價策略 东通工用磷渗透法研发出了高频的晶体管,于1957年做出世界上第一款袖珍收音机,并在这款产品上启用了索尼商标。盛田昭夫到美国去推销袖珍收音机的时候,德州仪器刚刚轻率地放弃了这个市场。美国人对盛田昭夫说:你们为什么要制造这种小收音机?美国人都想要大收音机。盛田昭夫回答:单单纽约就有20多家广播公司,同时就有20多套节目在播放,每人使用一台小收音机收听自己喜欢的节目,岂不更好?索尼用“一人一台”的宣传成功打破了美国人全家共用一台大型收音机的观念,成为全世界最畅销的收音机...