Can Baidu Thrive As a Stand-Alone Search Engine? (Tech Strategy – Podcast 76)

This week’s podcast is about Baidu and its key strategic questions:

  • Can it thrive as a stand-alone search engine? Against competitors such as Tencent and ByteDance.
  • What is its best growth opportunity? I think cloud.
  • Did it need to go into content? How far?
  • Can it compete in the attention market? Search plus feed.

You can listen to this podcast here or at iTunesGoogle Podcasts and Himalaya.

Here are my mentioned slides:





Related podcasts and articles are:

I write, speak and consult about how to win (and not lose) in digital strategy and transformation.

I am the founder of TechMoat Consulting, a boutique consulting firm that helps retailers, brands, and technology companies exploit digital change to grow faster, innovate better and build digital moats. Get in touch here.

My book series Moats and Marathons is one-of-a-kind framework for building and measuring competitive advantages in digital businesses.

This content (articles, podcasts, website info) is not investment, legal or tax advice. The information and opinions from me and any guests may be incorrect. The numbers and information may be wrong. The views expressed may no longer be relevant or accurate. This is not investment advice. Investing is risky. Do your own research.

——-Transcription Below

Welcome, welcome everybody. My name is Jeff Towson and this is Tech Strategy. And the question for today, can Baidu thrive as a standalone search engine? Now Baidu did a offering in Hong Kong a week or two back, raised some more money. You know, they’ve been listed for a long time in the US, but you know, like many companies, they’re doing a secondary in Hong Kong seems to be the thing to do. And they did a pretty nice filing 500 plus pages, which was some pretty good reading. And I thought it was a good time to sort of dig into this company because I haven’t really talked about it much. And in fact, I’ve kind of been dodging the question of search engines in general because I think they’re just a very unique business model. I mean, they’re really not like anything else. They’re this sort of strange animal, you know, out on the savanna. It’s, you know, it’s just a really weird thing. And also some of my thinking is not quite developed when you start thinking about learning platforms. which was, you know, this has always been sort of one of my five platform types I’ve listed, you know, marketplace, innovation, audience builder, coordination, payment, and learning platforms. And I’ve always kind of skipped over learning platforms for the most part because I’m not totally sure this is right. I’m not, I think I’ve got kind of half of what I’m trying to discuss described well, but I think this is probably going to change over time. But anyways, it’s the best I got thinking wise right now in terms of models to think about this, so I’m gonna stick with it for a while. But anyways, I thought this would be a good time to sort of dig into this, and I’ve been doing a little back and forth with the Baidu Media Relations people asking them for information, they’ve been sending me some articles, things like that. Anyways, for kind of a lot of time over the last week or two, I’ve been sort of writing and thinking about this. Those of you who are subscribers, I’ve sent you… Four to 5,000 words about ByDue thus far, and I’ve got another 2,000 coming tomorrow. I mean, I really kind of did a lot on this. I sent you about six PowerPoint slides mapping out their business with my little blue diamond charts and things like that. So hopefully that’s helpful. I think that was a good solid pass at their strategy now and some of the big questions they have going forward in terms of strategy, and ultimately, valuation, which is, you know. kind of the so what in all of this. So for today, I thought I would talk about what I think are the bigger strategic questions for Baidu going forward, how I see it, a bit about learning platforms, so some theory there, how to think about search engines, and kind of why they’re so unusual and why they’ve been becoming, I think, more and more complicated as business models over time. So that’ll be the topic for today. For those of you who are subscribers, more to Baidu coming tomorrow. Plus, I’m gonna start talking about two other companies that have recently listed, Jihoo, which just filed recently, and then Popmart, which has been out for a while. But that’s kind of an interesting business to think about, so talk about those two. So that’s on the way as well. Okay, now for those of you who are not subscribers, you can feel free to go over to, sign up there, free 30-day trial, see what you think. And my standard. Qualifier disclaimer, nothing in this podcast or in my writing or on my website is investment advice. The numbers and information from me and any guests may be incorrect. The views expressed may be incorrect or no longer relevant or accurate. Overall, investing is risky. This is not investment advice. Do your own research. And with that, let’s get into the content. Now, all of this information goes on the company library, which is on my webpage. Just click there and look for ByDoo. And the other tab, which is the concept library, the three ideas I kind of want to focus on today concept wise are learning platforms, which is one of my five platform types, and search engines, which really has two ideas there. One is a learning platform, and one is a standardization network effect. So on that concept library, search engine, learning platform, standardization network effect. Those are kind of the three concepts for today. All right, first thing to think about with search engines is, as I’ve said many, many times, I like competitive dynamics. I like competitive strength and defensibility. And the more of that you have in life, generally the better life is as a business. And sometimes you can make a lot of money, but not always. Obviously, unit economics are not guaranteed, but usually competitive. Defensibility is a prerequisite to really attractive economics because otherwise if people can do what you do They copy you and the economics aren’t as good Generally, okay. So I mean just high-level. What’s the simplest way to check if a company has a strong competitive position? Well, you can look at stability and market share and you can look at return on invested capital Now return on invested capital kind of depends on unit economics, but generally you can compare within an industry, let’s say retail. If you’ve got five retailers and one of them has better ROIC than the others, that’s an indication that that one has some defense that the others don’t have. So that ROIC is kind of relative to other people who live with effectively the same economics you have. So that’s kind of a relative marker within an industry. And it does change over time, but market share stability is kind of a simpler metric. Look, if one company has a very stable market share for five or 10 years, then almost by definition, they’re doing something that’s keeping people from taking their business. That’s kind of what defensibility is. Look, I’ve got a bunch of business, no one can take it. My market share is not swinging up 30%, down 30% year after year, it’s pretty stable. Now if you do see market share swinging up and down, that’s a good indication that, okay, you’re not that defended. Now there’s a caveat to that, which is oftentimes you will see an industry where two or three companies really dominate because they have basically the same competitive advantages. So you see a giants and dwarfs scenario. Hey, here’s a business like art auctions. And then you see two or three companies, Christie’s, Sotheby’s, and they’re the majority of the market, they’re the giants, and then everyone else is much smaller, the dwarfs. Okay, now within that, the collective market share of the giants should be stable. You should see the market share of Christie’s and Sotheby’s together be about the same year after year. But between them, it can shift back and forth. So you kinda gotta break that out a little bit because yeah, you’re strong and I’m strong, but we can take from each other a bit. but none of the smaller players can touch us. So the collective market share of the giants should be stable and hopefully the individual company. Okay, so we look at that and first thing that jumps out at search engines is, oh my God, the competitive picture appears to be pretty unbelievable. I mean, not just good, unbelievable. Google search, depending which numbers you believe, 85% of the world’s search engine is one company that much market share, that’s an incredible market position. And that includes all the suppliers. That includes Baidu, that includes NAVR out of South Korea, Yand, I mean, that’s everyone. And one company’s got 85% of the world’s search. I mean, the numbers are a little bit different, but it’s 80, some people say 90, it’s up there. Okay, maybe that’s just Google. Nope, I mean, who’s number two and number three? Well, Bing and Yahoo are two and number three. They got like 3% and 2%. I mean, that is real dominance. Now, we can pull that down to sort of, I wouldn’t say country specific, although it tends to play out that way. It tends to be language specific, like NAVR, which is the number one search engine in South Korea. One, the searches in Korean, but also the content, knowledge, and information they’re indexing and ranking is also Korean. So it’s almost more language specific than it is country specific, although it plays out that way. So if you look at NAVA and South Korea, they got 70, 75% of the market, same picture. We look at Baidu in China, 70, 75% of the market, Chinese language. Yandex in Russia looks pretty much the same. Okay, that should jump out at you as, all right, something powerful is going on here. That’s very unusual. Try to think of another website that has 85% of the world’s market for anything. I mean, there’s a couple of them. Dominance tends to be local, which is a good thing to look for, because you can find it all over the place. Dominance at a national level tends to be pretty rare. It’s hard to dominate a whole country. Dominance globally, I mean, that almost never happens. So it’s a real outlier what Google has done. Okay, anyways, it should jump out at you. Okay, there is a fantastic competitive picture here. Does it have attractive unit economics as well? You take a look at that and it’s like, yep. It’s got that too. These are really great businesses, search engines. There are other great businesses in the world that always get my attention. As you probably know, marketplaces, like companies like Shopee and Taobao, they totally get my attention. I think those are great business models. Competitive position, market size, unit economics, I mean, I really like them. I like casinos too. I think casinos are quite attractive. Alcohol, Coca-Cola, not hard alcohol, well, beer is a little bit different, but like liquor, Johnny Walker, Diageo. Cigarette companies are actually awesome businesses, although they’re not good for people. They’re actually great businesses. There’s a handful of businesses that always get my attention and Search Engines is on that list. All right, but what’s the business model? Now, I think these are platform business models. But I mean, people can, it depends if you believe that a learning platform is a type of platform business model. So if you look in the show notes, or for those of you who are long time listeners, I always talk about five platform types. Some people think it’s two, some people think it’s nine. I generally say five describes most of what I’m looking at most of the time, but there are exceptions. And number five on my list has always been learning platforms. Now, my definition has been a platform business model. requires more than one user group. And the primary purpose of the platform is to enable interactions between those user groups. Those could be transactions, like in a marketplace, they could be communication, they could be viewing, they could be a lot of things. But you have some type of interaction and that’s the primary purpose of the platform. Now a learning platform like a search engine, we could say, well, one user group is clearly the person searching, right? And You know, platform business models are about lowering CoSien coordination transaction costs, which can be price negotiation, information asymmetry, coordination, and search, searching costs. Well, it’s a search engine. Its goal is to help you find information as a searcher you could not find otherwise. So, I mean, almost by definition, the fact that it’s a search engine means that it’s focused on lowering search costs, which are a type of CoSien transaction cost. and that’s the purpose. Okay, so that side of the platform’s pretty easy, and then we have a second user group, which is advertisers. Fine, we know that business model well. You know, you have newspapers use advertisers. That’s a second user group, and you know, you charge them, and then you give it to the searchers for free. Fine, that’s advertising-based media, we know that. It’s the third user group that gets complicated, because in the early days, what you were really searching, you know, this is back in the… 2000, 2005 range when a search engine was basically just going online and helping you find web pages. Because the web at that point was a bunch of web pages with text and some photos and that was pretty much it. And it wasn’t on phones, it was just you could get there on a browser on a PC. So you can kind of search everything with a web crawler that could just search the world’s web page, rank it all, connect it all, and then tee it up. Those were kind of the earlier simpler days of search. Okay, fine. So who is the other user group? Is it web pages? Well, I suppose you could say it’s web pages or you could say, well, it’s a content provider. Who is the content provider? Well, it could be an individual or it could be a company that has put up a web page with information and that’s what we’re searching. Now, I think in the early days, you could call it sort of a content provider. Who is putting up web pages? One of the reasons this gets complicated when you start thinking about learning platforms is I’ve never said in any of these talks, well at least I didn’t mean to, that a user group has to be people or companies. All of that implies a human agent. You know, companies are a group of people. So if it’s a marketplace like Taobao, you’re connecting buyers with sellers. It’s implied that the sellers are people, they’re human agents. But one of the things that’s happening is we’re seeing digital agents, software, AI. You know, when a web browser, when a search engine searches the world’s information, is it really searching information put up by people or is it kind of a combination of human agents and digital agents that are creating and showing? I think it’s kind of a mix. And now, okay, true, the software, I mean, maybe the software is creating the content itself, which is happening more and more. Or maybe the software is just creating the presentation and putting the information out in the world. So it’s kind of a mix of human and digital agent. But one of the reasons learning platforms and search engines is complicated is when you start to think about what really is a user group. You know, if I’m connecting people searchers with information on webpages, is webpages, is that a user group? Is that a combination of content providers and software? Is it? human plus digital agents, how do you think about that user group? And this is a problem we’re going to face more and more because a lot more of what’s happening online is created by digital agents that we interact with. If you call a customer service rep, there’s a decent chance you’re not talking to anyone anymore. You are talking to a digital agent that sounds a lot like a person and most people can’t tell anymore. We could be moving towards a time where we could be seeing platform business models where one user group, the searchers. the buyers, are digital agents. And they are connecting with another user group, the sellers, the information, that is also digital agents. We could have platform business models with no human interface, no webpage, no mobile app, no nothing. It’s just one group of digital agents interacting with another group and we’ll call that a platform business model. And we’re kinda getting there already. So anyways. Search engines kinda tee all this up, which is why it gets a little more complicated than just talking about buyers and sellers in a marketplace. Now jumping to the so what here, what is a basic search engine, a Bydo, their core mobile app, Bydo app, Bydo Search, circa 2005, when search was a lot simpler than it is now. I think it’s a learning platform with a standardization network effect at the same time. Now I’ve defined a learning platform as a business model where the more users and or interactions, the smarter the service becomes, and therefore the better the service becomes. The primary purpose of the platform, it’s defining characteristic, is that it learns as people use it. And that’s just a characterization. Now it kind of sounds like a network effect in a platform. idea mixed together there. So I’m not sure that definition’s gonna hold long-term, but it’s pretty good. When I look at a search engine, I mean, the primary characteristic is helping you find information and knowledge. And that’s how Baidu describes itself in their filings. They say, we are in the business of searching for information and knowledge. Now, information used to be, okay, I just need to find this company or find this text. You’re just searching text. But clearly information has gone from information to actually human knowledge. Help me answer this question. What is the best way to fix my broken window? That’s not searching for information. That’s searching for answer to a question. So, but I think that characterization, information and knowledge is a good description of what search engines do. And I think, you know, a learning platform where the primary characteristic is that process gets smarter the more people use it is a good way to think about it. I think everything else is secondary to that. And that clearly has a network effect implied. The more people search as a group, the smarter it gets, which is pretty much how search engines work. And the more you search personally, often it will become better just for you because it knows more and more what you like. So it kind of gets smarter on two levels, but it’s mostly in the aggregate. And it’s a pretty powerful network effect because search engines have a nice, huge, long tail You know, if you go on Bing and Yahoo and Google and you search for like, what’s the weather today, pretty much any search engine can give you a good answer to that question that is sufficient. But if you ask for something bizarre, like, you know, what was the most popular type of fruit in 1632 in Belgium? That’s very weird, long tail content. Well, a more robust search engine that more people use and that indexes more content will be better at those quirky niche. type questions, information, slash knowledge. Now that sounds a little bit like what people call a data network effect. This term has been floating around for a while. I’ve really stopped using it. Data network effect is hey, you go on Netflix and it knows what to show you next as a video because it knows what you just watched. You go on Amazon or Alibaba and it shows you a very personalized store for you because it knows what you’ve bought before and your Amazon looks different than my Amazon. One, I don’t like the term data, because I think it’s, nobody really knows what data means. Like, text on a spreadsheet, text on a Word document, photos, video, what isn’t data? It’s a very nebulous idea. And I don’t think it’s a, I think that’s just personalization and customization in a data-rich world. I don’t think that’s anything terribly special. I think that’s just. table stakes in pretty much any business you’re gonna be in now that they’re gonna personalize and customize. So I put learning platforms at a level way above that in terms of its power. I think everyone is gonna have to personalize and customize, I don’t think it’s great. So I don’t use the term data network effect very much. But I mean, I think you could hopefully agree with me, the primary characteristic of a search engine is it gets smarter and more accurate. And that’s really what you go to a search engine for. Okay, now. I put in the notes my basic, my slide for a basic search engine, which is a, you know, one of my blue diamonds. You can see three user groups, searchers, web pages by content providers and advertisers. You can see two network effects between web pages and searchers and between searchers themselves. And that all gets you demand side scale, which is always the name of the game. However, I’ve also put a third network effect on there, which is that infinity sign next to webpages by content providers, which I’ve called a standardization network effect, which I think is a different thing. So when I look at a search engine, here’s my takeaway. I think it’s a learning platform on one side and I think it’s a standardization network on the other. Why? because anyone who has a webpage in this world has to do search engine optimization. You are creating and tailoring your website to make it easy for Google in particular to search and rank it. Everyone’s doing that. Every webpage does SEO, there’s tons of consultants and marketers. I get emails that are unwanted every single day from SEO consultants. Your webpage is not ranking well. And it’s because Google has set the standard for how information on the back end should be displayed on a webpage such that it is easy for search engines to see it and rank it. And as they change their ranking algorithms, everyone changes the SEO they’re doing and everyone wants to be the, you know, the highest ranked for any term that is searched. You know, the standard joke, which is a pretty good joke, is the easiest, you know, the best place to hide a dead body is on the second page of a Google search result. because nobody ever clicks on the second page. So, you know, you want to be ranked high. Now this is a standardization network effect. The more people that use the same standard, the better it is for everybody. The fact that you use Google, I’m sorry, you use Microsoft Word and I use Microsoft Word makes it better for both of us. We can send each other documents, we can train it in the same way, we can all learn one program instead of 10 programs, we can have other things that are built to support Microsoft Word. So people, you know, they compete early on in a lot of businesses for standards. They want to become the standard. They want to be PDF, that’s Adobe. They want to be Microsoft Word. They want to be Microsoft Excel. They want to be whatever, MPEGs, JPEGs. And a lot of the discussion coming out of China right now, which is interesting, is the fight to become the next wave of standards globally. Usually they’re talking about protocols and formats and things like that. Anyways, I think it’s pretty obvious that Google has set the standard for how you present your webpage so that it can be ranked by their search engine, everyone’s doing it, that helps everybody. So I think that’s a clear network effect on the other side. Anyways, if you look at the chart there, you’ll see my blue diamond, you’ll see the network effects, one, two, three of them. And on the supply side, which is down below, which is the yellow boxes, it’s pretty basic. You just have a lot of spending on tech and web services. That’s gonna be your primary fixed cost. And because it’s mostly a fixed cost, you’re gonna get nice operating leverage, which is exactly what you see in Google’s financials and ByDue’s financials. That is really the first point for today. That’s how I would describe basic search. You can see my graphic there. And it covers sort of two of the main concepts for today, which were learning platforms and standardization network effects. However, Search, unfortunately, platform business models, they’re always developing. A marketplace platform like Shopee is different today than it was two to three years ago, and it’s gonna be very different in three or four years from now, because you are in the interactions business. And as the user groups evolve and change, the nature of their interactions also evolve and change. Taobao and today in China is very different than five years ago. It’s got content, it’s got live streaming, it’s got short video, people’s behavior. is changing, how they shop is changing. So it’s kind of a living thing. Well search engines are the same way, that they have evolved with not necessarily how people search that hasn’t changed that much, but the nature, I mean their goal in life is to present the world’s information and knowledge. Well how information and knowledge is stored and presented has been changing quite dramatically since 2005. It’s no longer Everything’s a webpage. On the webpage you have text and a couple photos. That’s not how it works anymore. You have mobile apps. Well, that’s a lot of information. How do you search a mobile app? You can’t just send out the software to crawl. I mean, the companies that have these apps don’t want you doing that. And a lot of the world’s information and knowledge is now within these sort of walled gardens, something like WeChat. You know, Baidu, they can’t search inside of WeChat. Tencent won’t allow that. They have their own search engine internally. So the shift to mobile kind of changed where the information went. We started to get from photos and text to video. Video is a massive source of information and knowledge. How do you capture the knowledge in a video? It’s actually very hard to understand what’s happening in a video if it’s not a human being looking at it. I mean, if I’m watching a video of how to fix a broken window, software doesn’t know. I mean, software, it’s very hard for the software to know. That’s what’s the content here. So how can they sort of tee up and catalog and rank that type of information and knowledge when it’s in videos or songs or podcasts? I mean, is it going to listen to every podcast and understand what I’m talking about? So the world’s information has gotten into sort of more complicated formats. and it’s harder for search engines to access that. And you can see Baidu and Google and these companies struggling with that in terms of the nature of information. And then if you actually look on the searching side, as opposed to the knowledge and information side, it’s also getting complicated. Okay, you can go in and you can type in how to fix a broken window. But often people are starting to do sort of voice interaction. You know, these AI speakers where, you know, what’s the weather today, Siri? And you know, it’s sort of a voice to search. So it’s kind of multimodal search. Where you search is not necessarily, I’m just going to Baidu and typing in my question, or I’m just going to Google homepage and typing my question. You know, these search fields are embedded in lots of websites and apps. So if you go into any regular app, it’s probably using Google or something else. So it’s kind of extending out into other webpages and apps. It’s going into things like smart devices, smart TVs, smart speakers. It’s just gotten a lot more complicated on the searching side and on the sort of information and knowledge side. And I think that’s gonna, you can do video search now. You can do photo image search. You can just point your phone at something like a cat and. it will search based on what the camera sees. And then there’s sort of AR and VR search. What would that look like if I’m in a VR environment? How do I search there? Do I say it? Do I type it in somewhere? Do I just point with my finger? And then you got languages too. I mean, what about the fact that are you searching in Chinese, are you searching in English? Or is it searching in English, but it translates it to Hungarian, and then it also searches all the Hungarian websites to get me an answer in any language. So it’s just getting more and more complicated to catalog and understand all the world’s information and knowledge and how you interact with that. It’s a very big, difficult tech challenge, and it’s dramatically harder than it was back in 2015. So in the show notes, I’ve put a slide for complicated search, and I’ve sort of flagged some of these issues like open versus closed gardens human versus digital agents, webpages, mobile apps, other types of content and service providers, how you search. And the one thing I’ve added on sort of the supply side is I’ve broken out a second core capability you need, which is AI and machine learning, which I think is obviously different than web and tech spending. And they really, these search engines really were the first in the pool in terms of developing AI. Because AI is cheap and fast prediction. Well, that’s exactly what a search engine is. It’s predicting an answer to the question and then seeing if it gets it right. And then on the backside, it’s predicting advertising placements and things. But it’s mostly AI. So I put that as a second capability. And that brings us to the question of Baidu. Okay. Going forward with Baidu, I wanted to kind of tee up their core business, which has always been search. And it went from a basic search in 2005 to a more complicated search 2010. And it’s still evolving. I mean, search is a constantly evolving sort of technology. It’s pretty amazing if you think about it. Okay. But, you know, when people talk about Baidu, what they’re talking about now is usually they’re talking about growth, because Baidu has had fairly flat revenue growth for quite $16 billion in advertising revenue, and that’s how they make their money is advertising last year. Okay. And that used to be pretty good when they were the same size as Alibaba and Baidu. I’m sorry, Alibaba and Tencent, and it was the bat companies 10 years ago. Well, and those other companies have now grown and they’re much, much bigger now. Tencent is in 2020, $65 billion in revenue versus Bydu at about 16. And even ByteDance, which is a fairly new company, they had $16 billion in revenue a year ago, but they just logged in and they recorded $27 billion in 2020. So they’ve just sort of blown past Bydu. And as they’re all kind of making their money on advertising, although Bydu is more paper click and paper performance, not attention-based advertising. But generally speaking, there’s a fixed pie when it comes to digital advertising spend. And ByteDance in particular has been taking that from Baidu in particular. So there is this question of, okay, is being a search engine enough? And they’re not doing anything wrong. It’s just that the nature of their business is, it’s a very good business with good economics. But it just doesn’t grow because it doesn’t. I mean, it’s just, you know, there’s only so many searches and they’ve got that and that’s that. And there’s nothing wrong with that. It’s just who you are. So, you know, that’s been kind of this growth question that’s been circling this company for quite some time. And the company has, you know, we’re going into self-driving cars. Why growth? We’re an AI company. This is how they now describe themselves. We’re an AI company. Why growth? Why did they go into IGE? Why did they go Baidu Waimai, which was their competitor to Meituan and Ulema, which they subsequently sold? You know, this growth question has been around them for a long, long time. Now my questions when I think about them going forward is really three things. First of all, what is their best source of growth, both in terms of size and probability of success? talk about that either in this podcast or I may have to break this into two because it might go long. Second question is why did they start going into content creation? I mean their biggest holding is iQiyi, which is the number one or number two sort of video meet streaming service of China. Lots and lots of users but it loses money because it’s a content platform. It’s not like YouTube. I mean… It’s like YouTube, but it’s also like Netflix where you spend a lot of money on content because that’s how you differentiate from Tencent video and Tudou, is sort of in-house content. So basically Tencent video, Tudou and iQiyi are all spending a ton of money on content. And it basically loses money every year. Okay, that’s content. So question number two is why did they move into content creation and did they have to do that? And then third. Why did they go from search to search plus feed? If you look at their filings, they will describe their core products, the ByDo app, the Search app, pretty much everything they do as this phrase, search plus feed. And I’ll talk about that in a minute, but I think those are kind of the three big strategy questions I’m thinking about. Okay, so they describe themselves as we do, quote, search for information and knowledge. And I like that language. Search is very clear. We know exactly what we’re doing. Information and knowledge is very crisp language. I like that. They’re not entertainment. It’s not, hey, come to Netflix and just chill and watch for a while. It’s not, hey, just spend time on bite dance watching videos. You know, that’s entertainment. It’s, I don’t know, learning things. It’s just killing time. It’s storytelling. There’s a lot of things you can do online. They are talking about searching for information and knowledge. Okay, I like that. But as kind of mentioned, you know, where you access that information and knowledge is not as easy as it was back in 2005. It’s within videos now. Well, what’s the difference between a video that’s information and knowledge and a video that’s for entertainment? I mean, if I go to YouTube or Youku or whatever, those are all kind of mixed together. Am I only searching and indexing the ones that teach me how to fix my window? Or am I also getting all the videos of, you know, Blackpink dancing and whatever? I mean, it’s kind of hard to separate out the information and knowledge from a lot of the world’s content, especially music, podcasts, videos, things like that. So you can see they’re kind of starting to bleed into this idea of general content as opposed to just information and knowledge. And then a lot of that content maybe isn’t available to them directly. If it’s in WeChat, or ByteDance, they can’t index that. I mean, it’s behind a walled garden. So they start to think, well, maybe if it’s really critical information that our search engine requires to be effective, we may have to develop that in-house. So they start to do in-house content, which doesn’t mean they’re hiring people and writing themselves. It usually means they’re creating some service that allows users to generate the content. So the ones you’ll see are like Baidu Wiki, which is basically like Wikipedia. Well, Wikipedia is a big source of information and knowledge, so you’d almost have to build that yourself and offer it as Baidu Wiki. And you get users to generate the content, so it’s UGC, but you kinda gotta have that if you’re gonna be an effective search engine. Baidu knows, which is like Juhu or Quora, where you’re asking specific questions. I mean, you can see how they kind of blended. from information and knowledge into general content. And then they also blended from, hey, we’re just gonna index what’s out there to, hey, we may have to create this content ourself if it’s critical to our search engine. And then you end up with IGE, which is arguably the biggest jump in this direction. But they did kind of go from search to, search plus in-house UGC user-generated content plus IGE. And you know, 2015, that’s really what they looked like. They looked like kind of a mix of a search engine and a content creation company. I mean, was that really necessary? Maybe you think it was, but it looks to me like, I’ve got a, at a certain level, it looks to me like, I’ve got one platform business model, which is a search engine, which is very profitable and attractive. And I’ve got some other platform business models, in particular, IG, which is an audience builder. which gets a lot of traffic and has some real competitive power but is losing a lot of money and probably will forever. Do I really have to be in that business? Can’t I get there by partnership? Can’t I M&A that with Alibaba’s 2DOE and just maintain the rights to search and index that information without running the company anymore? And I think there’s a lot they’re doing within content creation, UGC, IGE. that could be better done through partnerships and get that off the books. This is not a complementary platform business, which you probably know that I really like complementary platforms. I like how Ant Financial complements Taobao. I like when you get multiple platform business models supporting each other because it can be quite powerful and almost impossible to reproduce. In my little strategy pyramid, that’s top of the pyramid, complementary platforms, because I think it’s effectively impossible to reproduce usually. Okay, this doesn’t strike me as particularly powerful. It strikes me as a really good business paired with not a very good business. And I’d like to get off the books and, you know, if the only goal is to access the information and get it indexed or to get the users, then let’s do that some other way. But I think that tees up the related question, which is, am I in the search business or am I in the attention business? So my next question is, they describe themselves as search plus feed. Right? Baidu Search plus Baidu Feed. And here’s how they describe their feed. Quote, Baidu Feed provides users with personalized timeline based on their demographics and interests. Baidu Feed complements our core search product, leverages Baidu AI recommendation algorithms and monetization platform, and contributes to user engagement and retention. That sounds a lot like TikTok. That sounds a lot like Twitter. That sounds a lot like Facebook, where the primary product is a newsfeed. You know, you sit on Facebook and you just scroll down and it tees up one thing after the next that it thinks you might wanna see. That’s TikTok, that’s Twitter, and they get there by different mechanisms. You know, Twitter has an interest graph where based on who you’re following, it knows what to show you. Facebook gets there by a social graph. By who you know, they know what to show you. ByteDance just uses its algorithms and figures out what you wanna see. But you basically sit there, and all three of those businesses, I would argue, are in the attention business. Their goal is to get your attention, to keep your attention, and monetize it by basic advertising. Now, I think that’s a different business than search. I don’t think that’s the same. I think search engines are a very specific type of business, and they monetize by basically pay-per-performance. advertising, you pay for search results. This is different. This is about my goal is to keep you staring at this newsfeed all day long. And the more of your time I can get, the more of your attention I can get, the more I can monetize by just putting up display ads and other stuff. So they kind of slid from this idea of we’re a search engine to we’re search plus feed, which that’s a big jump. It’s not clear to me they want to be in that business or they have to be in that business. Maybe they do. Maybe if you’re dominant in feed, let’s say Tencent, WeChat feed, that those companies can now add a search engine and come at this from the other direction, which is what Tencent is doing. I mean, they’re building a search engine, SoGo, which is number two search engine in China, and it already has inside of WeChat and external. And ByteDance says they’re building a search engine. So maybe the ultimate user experience is search plus feed and Baidu is trying to get there by jumping from search to feed and bite dance is coming the other direction as is WeChat from feed to search. Okay maybe that’s true. I’m not convinced it’s true. I’m not convinced you can’t just be a standalone good search engine and even if you strategically decide I have to be in the attention business I’m not sure they are set up to win that game. and that they’re terribly competitive against Tencent and ByteDance, who are designed from day one to capture and hold your attention. This is why Tencent, they have video games, they have music, they have videos, they have chat, they have a full suite of things designed to keep your attention. That’s historically not what ByteDue is. And, you know, ByteDance is trying to get there as fast as they can. So that’s kind of my other question is, you know. Did they need to go from search to search plus feed? Because if they did, that’s a big strategic move. Which is why my question for this podcast, the title of this podcast was, can Baidu thrive as a standalone search engine? Not just survive, can they thrive? Is it enough to say we are the best search engine, we don’t need to do much content beyond the basics? and we don’t need to go into feed and we don’t need to go into attention. It’s enough to be the number one search engine, that’s it. Or, yeah, we really do have to do this. It turns out that these other companies, Alibaba, Tencent, ByteDance, they are generating so much attention and data and cash flow. from these other businesses they have, which for Alibaba would be e-commerce, from ByteDance it would be short video, and from Tencent it would be WeChat and gaming, that their search engine is a complement to much larger and more powerful businesses. And we don’t wanna be the standalone player against these giants who also do search as a complement and don’t care if they make money on it. I think that’s the strategic question that sits at the center of this. So let’s say that’s true. Let’s say, yeah, you don’t wanna be a standalone search engine in a land of giants. That’s just not, don’t do it. Okay, then we’re back to growth. What is the best, the largest, and the most likely growth opportunity they have going forward? Now, they say it’s, well, they say it’s a bunch of things. I mean, they’re kind of all over the place. They’re saying, we’re going into AI, we’re an AI company. That’s how they describe themselves. They say we are doing self-driving cars, which they’re doing, they’re Apollo program. I don’t have strong feelings on that. I think self-driving cars, I like that they’re trying to do the software, the AI, and not building cars, which is what Xiaomi has said it’s gonna do, which I think is crazy. I think building cars is awesome if you’re a billionaire and you wanna do something light, you know. that’s dramatic and important in the world. I don’t think it’s a great way to make money. You know, if the shareholders of those companies make money, it’ll be kind of a fluke. I don’t think that’s their highest priority. I think they wanna put a dent in the universe. I don’t think they wanna make shareholder money. So, you know, whenever I see a tech company say, we’re going into electric cars, I’m like, oh man, that’s not gonna be good. But I like that they’re approaching it only from the software side. I like that they are starting with AI. Because at the end of the day, they’re doing what we’ve seen many companies do, which is they’re gonna take their core capability and they’re gonna externalize it and exploit it. This is how Alibaba created Ant Financial. They built a core capability, which was payment to support their primary platform, but then they externalized it, offered it to everybody, the whole market, and then they exploited it and built it into something much bigger, which became Ant Group. I think we saw Amazon do the same thing. We have a lot of servers. Let’s begin to sell all those servers to the market. We’ll call it Amazon Web Services and we’ll build that into its own business. I like that they’re building off of core capability and their core capability is AI, so that’s great. So let’s externalize and exploit AI. And they’ve got a bunch of projects under that bucket and they call themselves an AI company. Fine, fine, I think that’s all reasonable. My feeling on that is I think electric vehicles are very far from commercialization such that they’re gonna see a lot of cash flow in the near future and I think that’s what they need. I think if the goal is like we gotta become bigger, we don’t wanna be a standalone search engine in the land of giants. Okay, we need revenue and cash flow now. What’s the shortest path to really big revenue and cash flow? Is it self-driving cars? No. Is it a GitHub type open AI platform which they’re building? No, I think those things are both cool. The one that jumps out at me is cloud. Cloud. Alibaba says their biggest opportunity going forward, their next giant business is cloud, full stop. And Baidu is already number, let’s say four in cloud and they can come at, so you’ve got Alibaba, you got Tencent, Huawei’s in there and then you got Baidu. And they’re already showing revenue and it’s growing at 40 to 45% per year, even though it’s small. I mean, they’ve, they’ve got serious revenue. It’s in the low billions and they can come at cloud. So we, we know cloud works. We know it’s a real business model. We know it can be a giant because we already see Amazon web services. We see Google cloud. We see Azure. We see Alibaba cloud. And against that clear proven business model, we have a nice core. differentiating strength, which is we can do cloud with a lot of differentiation on AI. That to me strikes, you know, that’s what I’d go for almost for sure. I mean, you’d have to do a full analysis, but I’d be surprised if that doesn’t jump to the top of the list is like, here’s how we become a digital giant in China once again, is we are a cloud business and a search engine. Bam. And that’s kind of how I take apart this company. I think there’s a lot of cool stuff here. I think the car stuff is really neat. I think their GitHub type open AI platform is neat. I think their content initiatives, IGE, Post, all this stuff, I think is pretty neat. But the two questions that I use to sort of project what I think is gonna happen in the future based on business model, not management performance, but business model is, does the search engine Can it thrive as a standalone business versus the search engines coming from Tencent and ByteDance? And two, how fast is the cloud business growing? Like those would be my two questions that would pretty much tell me what I think’s gonna happen in 18 months and two years. And everything else is interesting, but I don’t think it’s gonna move the needle financially in the near future. So anyways, that’s kinda how I take apart strategy at this point. Let me see that I’m not. cover anything? No, that’s pretty much it. The core concepts for today, five platform types, learning platform, and search engines, basic search engines, as a combination of learning platforms and standardization network effects. And I’ve put the slides in the show notes, including the last one that basically shows this business as two complementary platforms. One is the search engine, one is the cloud platform, which is an innovation platform. But, you know, two complementary platforms. And I’m sorry, I know these slides don’t show up in the show notes within the podcast. I’ve tried to paste them in, they don’t seem to show up very well. You have to click over to the webpage and you can see them there. But I’ve tried to put them in the actual podcast, doesn’t seem to work very well. And that is pretty much my take on Baidu thus far. As for me, I’m having a pretty nice time. The weather here is great, the pollution’s gone, which I really enjoy that a lot. In a moment of stupidity, I bought Red Dead Redemption 2, which was just going to kill my schedule. It’s another open world PS4 game. I really liked it. It’s by the same group that did Ghosts of Tsushima, which it’s kind of funny to think about this. I’ve always been sort of a global nomad. I’ve lived everywhere. England, Sweden, New York. California, Saudi Arabia, Beirut, Paris, China, Bank. I mean, I’ve kind of slowly explored the whole world over my life. And what type, what video games do I really, really enjoy? Well, it’s the open world games. It’s the ones where you play Ghost of Tsushima and you explore all the islands and you explore all the adventures and Grand Theft Auto where you explore that world and now Red Dead Redemption, which is again, it’s an, I don’t think that’s a coincidence. I think the fact that I enjoy these open world games so much has kind of a personality trait. It does seem to be a running theme in my life. Anyways, that’s, it was a stupid idea to get this game because it’s just gonna kill so much of my time I should be doing other things with, but it is so much fun. So, riding horses and shooting people on a revolver, which is, is that better or worse than running through Tsushima and killing people with a sword? I don’t know. but that’s how I spent my morning. It is a pretty spectacular game. I mean, it really is. You know, the level of video games are so impressive compared to what I used to play when I was growing up. But I think that’s it for me. If you have any suggestions on companies you’re looking at that you think would be worth taking apart, feel free to send it my way. If you have any recommendations for video games, please send those too. It’s really, I’m having some fun with that. But otherwise, I hope everyone is doing well. I hope you’re all having a good spring and that this is helpful. That’s it for me. Have a great week and I will talk to you next week. Bye bye.

Leave a Reply