Livecast #1: WeChat 8.0 & Channels Recap with Dan Grover
TRANSCRIPT
We are grateful to Yalan Wu for editing the transcript.
Rui Ma: Thank you, Dan, for joining me, it's really a pleasure to have you. I know you are very busy with your full-time job and also you are my go-to expert when it comes to anything WeChat related, especially anything that is also around social because of your background. So for those of you who are not familiar with your background, can you give us an introduction?
Dan Grover: Sure. So I'm Dan, I've been an engineer before product manager been around a bunch of companies in the Valley, spent a few years in China on the WeChat product team. I was just looking at my Facebook memories and I saw I had posted that I was at the fifth-year anniversary, and now this is the 10 year. And I didn't realize how long it had been, but I was there about from 2014 to 2016 or so.
And I've been back in the U.S., But like you, I still like to follow this stuff. See what's going on with Chinese tech. I used to write blog posts about some of the trends in product design and UI in China. I wrote a post a lot of people read about why everyone was into chatbots here for a while and how things didn't quite work the way that people thought they did.
Rui Ma: Your posts are very high quality. Yeah. So I’d really suggests you go to dangrover.com. If you're interested in revisiting, maybe if you've already read it before, some of Dan's excellent posts, especially those comparing and contrasting the strengths and weaknesses of China tech.
Now I'm going to talk about WeChat.
It's now used by almost 1.1 billion people daily and that's of course, primarily in mainland China, about 30% of these people or 330 million use video chat. Actually, I was really surprised to hear that stat about 72% view their moments feed. Effectively your private network that not everyone can see, but if you are friends with this person, you can see their moments.
11% post to moments, about a third read articles in WeChat through the public account program, which is very important actually to the whole Channels product. Almost 40%, so about 400 million people use Mini Programs, which is also something really interesting, and we've definitely talked about it on Tech Buzz in detail.
So you can revisit our episode on that if you're interested and about 200 million people use the product that we're going to talk about in depth today, which is channels, that's almost 20% of WeChat's total daily users. Maybe you didn't know this product has been around for exactly one year basically. On WeChat's 10th anniversary, Allen Zhang, who is the creator of WeChat.
We're going to be quoting him and talking about him a lot today. Allen gave us really good background on why WeChat decided to do Channels - this video format - and that's because he noticed in the last five years, users are sending 33 times more video messages within the chat and 10 times more videos into their moments feed.
Clearly videos is a huge use case. Also he notes that not everyone can write long form content, right? So the public accounts thing that we were talking about earlier [which by the way, inspired Substack] is something that's for people who are really into writing content. And that's just not for everyone, but far more users can obviously shoot a short video.
What happened was that in 2017 or so when the team was looking at this and saying that, okay, we really need add video into our product. They realized that would require a whole new account system, since right now the way it works is that within WeChat, you have your sort of private feed with your friends that you're connected with. So, it's not just a pure follow function versus the public account distribution system of content. It's really like a follow system, like on Twitter, for example, it's more broadcast oriented. The video product that they were thinking of didn't quite fit into either of these systems. So, they realized they really needed to create a whole new account system.
And that was a lot of work. So, they didn't really get around to it until 2019 or so. And video is really the next big thing. It was super obvious with the rise of ByteDance and Kuaishou, and they simply felt that they couldn't be avoided. Let me describe what is Channels for those of you who are not on WeChat. Anyone with a WeChat ID can sign up for a Channels account.
Within Channels, you can now create videos, or long videos and then you can also livestream. If you're in China, so if you're abroad, you cannot quite livestream. Channels will then show up in your profile as a separate section. And it'll also show up if you go to the discover button and then you can look up Channels and look up video content there.
When you do that, and you go into the Channels feed, you'll see that there are actually three separate feeds. You default to a feed called friends, which means that you'll be seeing what your friends have been watching and have liked. You'll be able to see actually, which one of your friends have liked each video.
And you'll see how many have liked each video. And then in 8.0, they actually made it so that your like can be private if you want to. So I guess like I could like a video and then it will show up with my friends, I presume, but then actually they won't see that it was Rui that liked it. That's the default function.
Then you can look at accounts, you follow, and then there's a section called hot. I think it's basically a purely algorithmic machine learned content. So it just tries to figure out what you like to watch based on your preferences, but it does not have anything to do with what your friends have done or your explicit follows.
Finally, also livestreaming, like I said, is a big part of Channels. And what they did in actually starting past couple months actually, is that you can now see who is livestreaming near you. So, WeChat has always had a people nearby function and now that will default to basically people who are livestreaming near you.
I think the most important thing that everyone here, if you don't use WeChat, really have to understand is that WeChat actually doesn't really allow you to link to videos just in general content outside of the platform. They do, but it's not in a native format; it's it looks very different.
Dan Grover: You can, but nobody would, right? Cause like the way official accounts works, they give you an entire CMS and they host all this stuff for you and they add lots of interesting features like comments and being able to tip people. There's a lot of good reasons just to use all of the hosting built into the WeChat account system, instead of trying to host a blog somewhere else and send out links using the chat.
Rui Ma: Yeah, I think one of the main strengths, which we'll talk about later, with WeChat Channels, is that it's native to WeChat. So it's easily sharable within WeChat, that's different from other content from other apps.
Dan Grover: So, for folks who don't know, the workflow that you would use if you wanted to share a cool Douyin video on WeChat, is you'd have to actually download it to your phone and then upload it into WeChat as if it was your video.
So that's why they did the whole watermark where everything has the brand on it. If you want to share stuff with your friends though, it's not a great user experience.
Rui Ma: No, it's not. Yeah. And the last thing I wanted to add is that Allen wants to do all this and grow WeChat Channels without using any money.
He basically said we're not paying for content and he's very dismissive of sort of growth hacking tactics that a lot of other Chinese companies are.
Dan Grover: This is a consistent thread across a lot of their new products.
Rui Ma: Yeah, so that's the recap. Now Dan, let's talk about what happened. We both watched the WeChat pro conference which is their developer conference, annual keynote.
I know that we were chatting the entire time and I was finding it uninspiring, but then Allen gave this speech and then I personally felt a lot better. What's your reaction?
Dan Grover: Yeah, that was my reaction too. Watching the product development the past few years, I was starting to get a little skeptical of some of the stuff they were doing.
A lot of it seemed very reactive and very much driven by sort of corporate strategy and not the typical thinking that we see from WeChat, you know, like they had their stories feature. We can talk about how they change that. They have video feature. They have a Toutiao knockoff. They have a search section called Sou Yi Sou and they just have all this crazy stuff they keep adding.
And I was thinking as I was watching these presentations in the morning there, it was if you ever remember watching the classic, like Apple keynotes, where Steve Jobs would be out and then you'd say, and now we have CEO from, AT&T and then they'd get the guy like reading index cards for 10 minutes.
And you're like, when can we get Steve back? Cause it was that was like, every presenter was like that guy for a whole day. And then it's okay, good. Finally, we can get out and then you can hear what's going on. So, it's interesting. They do it in a kind of reverse order.
Which I think is because it's a developer conference first and foremost, like they're trying to talk to people in industry and show them the whole ecosystem and what they can do with WeChat. So makes sense. It's just less exciting for us to watch sometimes.
Rui Ma: What was something that Allen said that made you go, "Oh, wow that's really insightful."
Dan Grover: Yeah. So, he spent, it was a good two hour talk at the end. So he spent a lot of time just setting up the story of why they started the Channels feature, how they evolved it, what they learned. And I realized, okay, they've still got it.
So, there are a few interesting things for me. One is like he had this insight that the kind of content they were showing people would affect the quality of the content that people would produce. I've developed a lot of products where you have this kind of growth goal and you're like, "Oh, we need more people watching."
You would think, you've got all these levers of let's get more people into the feature. Let's get celebrities on there. Let's make our recommendations better. And he had the insight to say maybe the content sucks and maybe we just need better content. And how can we inspire people to produce better content, which is it's a hard thing to gauge, but I thought it was neat that he thought of that. There's also an interesting thread about how he made this prediction where he wrote it on a blackboard of ratio would be of stuff that you follow versus stuff that your friends like versus stuff that's recommended by the algorithm. And I think he expected there to be it was a one to two to ten of those, and it actually ended up being completely not the case. And they ended up leaning in further to the sort of the social context. So, I thought that was pretty interesting, like they could have just gone totally balls to the wall on growth.
And instead they're watching it, tinkering with it, seeing what, seeing how people use it. Yeah, so I liked that it was because people just didn't have friends that were liking other videos. And from the stats you cited earlier, it's only 18% were using Channels. So, if you're expecting that to be the driver, then obviously you're not going to see it until the feature has greater saturation.
Rui Ma: What Dan's basically referring to is that Allen said, remember I explained earlier, there are three separate feeds in your Channels. You have what you follow, you have what your friends are watching, and then you have what they call hot, but it really is just going to be machine recommended content.
And he made the guess that the distribution of time spent on this content would be something to the ratio of one to two to ten. So, for example, for every 10 minutes that you're watching content from people you already follow, or accounts you already follow, you're probably going to watch double that from content that your friends have liked and therefore is being recommended to you.
But you're going to be watching 10 times that for content that the machine recommends to you. His logic, by the way, for guessing this was, he felt that what you follow and what your friends recommend to you are going to be high quality content, because your friends are not going to want to like, or recommend, content that is very uninspiring or very common, very banal. Whereas the machine is going to have no problem recommending this content to you and it's to be very brainless and we're probably going to be spending a lot more time on brainless content, but the content where we're going to get the most value from is going to be from your followers and friends. So.....
Dan Grover: This is kind of an interesting reversal too, because I remember watching the keynote they did. And what was it? It must've been 2017 when they first did Mini Programs and he had this whole side rant about how they're never going to do any kind of ML in the product. This was before the video accounts came out.
Rui Ma: It sounds like Allen has basically, he's not totally reversed, but he's become more nuanced right? In the value of machine recommended content.
Dan Grover: And as you said, it's not the default tab when you go in there and as it is with Douyin. So, this was back in 2017 when they're first releasing Mini Programs.
And it was an interesting talk because it was this super hyped up thing, like everything had leaked before they actually announced it. And people were like, "Wow, this is just going to totally change the ecosystem. It's going to replace the app store. It's going to disintermediate Apple." It's going to do all these things.
So we have this, his two hour talk. Yeah, it was a lot of cooling water, pouring water on that sentiment a bit. And then talking about some of the deliberate limitations of what they had designed at the time, basically saying we're not going to have an app store. We're not going to be trying to recommend Mini Programs to people because we think that people will be naturally sharing them in context, when it is like situationally appropriate and he had some analogy about how, if a friend tells you that something is a really good movie, that's way better than a recommendation algorithm. And he said something to the effect of people are a natural recommendation algorithm, which I think is pretty insightful. If you think about it, we have sort of purpose-built circuitry in our brains for judging who to trust and which recommendations to look at. And it is actually hard for ML to beat that.
Rui Ma: I really liked that quote. That basically really leans into WeChat's whole like core asset, which is this extremely large social graph, dense social graph that they've built that no one else has, it's proprietary and they obviously see it safeguarded very closely.
Do you think what Allen has said, even though he is now acknowledging the value of machine recommendation, do you think the social graph - the defaulting to what your friends are watching - is really what's going to make WeChat Channels successful?
Dan Grover: I think there's validity to what he says, but let's say he's wrong.
And let's say that by any metric, if you're trying to just look at raw time spent watching videos or raw enjoyment, let's say that social context isn't the way to get that, and actually ML can be better. Even if that's true, he has to talk his book, right? That's, that is their advantage going into this.
And it is precisely why, when Douyin was coming out, they didn't have this advantage. Tencent had the social graph totally locked up. So, of course they had to make a really good ML algorithm for recommending videos cause they had no other way to drive engagement. I'm sure they wish they could have had your WeChat contacts in Douyin, but they didn't. So, it's whether or not it's better, that is the path that they have to take if they're going to win. And that is how they have to win.
Rui Ma: I totally agree with you. And of course, ByteDance has already tried to make social, because machine learning as great as it is, there are core advantages to having that social graph.
Yeah. For me personally, like the friends thing, it is interesting, because I am curious what my friends are watching. So even though a lot of my WeChat contacts are people I've met professionally. What they're watching is just really interesting to me because it helps me stay in touch with what the circle of Chinese venture investors or Chinese internet entrepreneurs, what are they doing, what are they talking about? It's like being on tech Twitter, right?
Dan Grover: There is a risk though, which is this whole - you've probably heard of the concept of context collapse - which we've all observed in the West where, some people will say they have just their close friends on Instagram or WhatsApp, but they have everybody they've ever met on Facebook. And so the stuff that they see in their newsfeed, it's every random person from their high school and everyone they've ever met. I have that problem a bit with WeChat. I have probably a thousand contacts on WeChat and never mind the fact that I don't live in China anymore, but I do think they reach a point where that becomes less of an advantage because they've done such a good job at making it this utility where you can add everyone in your life on it. I think every time a social network has tried to prevent context collapse; it never really works. WeChat tried to do lists and Google Plus had their circles and Facebook has something like that but there is a danger of that starting to peter out at some point if they think that's what their advantage is.
Rui Ma: Well, I'm already seeing it. I have my old Ah-Yis, my real estate agents, they're also on it, anybody you ever wanted to pay through WeChat Pay is on it.
So, for me, a lot of the value of knowing who is watching this and therefore ending up in my recommendation stream is that then it's a really easy excuse to go talk to them.
Dan Grover: The Channels feature, it tells you: these are videos that your friends have liked. WeChat is still, it is a communication app first and foremost, it is an excuse to go socialize with that person.
And if you think about it, like regardless of the context, humans always need excuses to socialize. Even though we're social beings, whether it's going to watch the game- it's not really about watching the game-, seeing someone's status message, or playing Fortnite or things like that. So, I thought that's kind of neat.
Rui Ma: Yeah. I think actually in Kan Yi Kan this happens to me where if I recommend an article, my friend liked it, he can actually thank me for it. Pretty much only have one friend that does that though. So, I don't know if that's like it often used function. Doesn't seem like it, but yeah, that, that is a way to start a conversation.
Dan Grover: This is a bit different from what folks are used to on Facebook and Twitter, where you have a single feed of everything. And occasionally something will pop up in there because someone liked it. This is just a whole tab of stuff that your friends liked, which is a neat experience.
Rui Ma: Recently I've been seeing in Chinese chatter, right? And this could be again, because I am in these quote unquote internet circles of investors and entrepreneurs who unfortunately happen to skew slightly older male. And I've just been basically seeing all this chatter about how they feel that Channels is attracting these quote unquote old uncles "zhongnian dashu." as both creators and viewers. Let's go into that slightly because I think this is potentially a place where WeChat Channels could be different. I'm not really sure. On the creator front, we have Douyin, right, ByteDance's algorithm, which basically rewards content that's very popular and yes, it gives you a personalized feed.
But what it's really good at, which has been covered a lot in Chinese media, maybe less obvious in English media, but I think now with TikTok, people get it. It's really good at identifying mass trend, so something that goes completely viral, TikTok/Douyin is really good at that.
WeChat, again, remember it doesn't default to this defaults to your friend's feed. So, there is a sense you have to be like really jumping on the viral trend of the moment to be seen in Douyin or TikTok. Whereas perhaps on WeChat because it defaults to your friends, then maybe you can just rely on your existing social graph to share your content more. You don't really have to care about what's quote, unquote, really hot at the moment. There's a sense. That's why I'm getting, again in China, that the old uncles are like: “Oh, yay. We can also get an audience because our friends care about what we're doing or what we like. We don't have to compete with the cosplay girls with their anime filters.” So what do you think of that, Dan?
Dan Grover: Yeah, I think there's some validity to that. I think having this focus on your friends and the kind of stuff that people around you are looking at and maybe publishing, I think it makes the whole thing a lot more human scale. It seems to be better because their goal isn't to suck you in for the greatest amount of time on video alone, their goal is to suck you in on the whole app and the whole ecosystem. And maybe that means talking to other people because you saw a video, maybe that means joining more group chats. Maybe that means some synergies with the other features in the app.
It makes sense to take whatever temporary hit they can take in terms of not trying to show the most engagement grabbing thing all the time as like the primary focus of the app, if it means that the whole app is going to work better in that it stays true to the purpose of being a communication tool and a social network. It's completely anti-social you can just sit alone for 10 hours and scroll through the thing and not talk to anybody. That's like the old David Foster Wallace quote. It's something like that.
Rui Ma: No, it's so true. It's so funny when people like, say TikTok is a social media and I'm like, what's really that social about it?
Dan Grover: Right, which is all fine and good, but there's already an app for that. So they have to do something different, right? About all the Ah-Yis and stuff, I have less experience with that these days. I have my wife's folks and my landlord and a few other older folks on there. From my limited experience, it seems like old people use WeChat in the same way that old people use Facebook in the U.S.; it's eerily identical behavior.
I have gotten a lot of videos or anytime it's a holiday or get some Merry Christmas video. Like the old e-cards. I see a lot of those forwarded, there's a lot of like stuff people share on. I think the term for it in Chinese is “ji tang”, like a lot of chicken soup, just like utterly banal, cliche stuff. It's the exact same stuff that older folks in my life share on Facebook. It wouldn't surprise me if the video section is a lot of the kind of videos that they're already sharing.
Rui Ma: Again, I don't have any insight data, unfortunately, so it's not like this is completely a data-driven comment.
Just anecdotally, my friends and family, just older people. They're actually, I would call them super sharers. Right. They're just like share a ton of content. It's all this like very, mish-mash like content that has nothing to do with whatever's hot or trendy at the moment,
Dan Grover: It could be a danger for the product. Because if you think about it, they're not sharing content they created, they're not sharing their opinion. They're not sharing commentary. They're just forwarding on this like cliche stuff, right?
Rui Ma: Yeah. Maybe like the chatter I'm hearing is really like the chatter I was referring to earlier, like these old uncles saying “Oh my God, I am excited to create on WeChat Channels.” That's still a pretty small segment. I, for example, can see my parents, maybe consuming that type of content, but I can't really see them creating it. So there's that dichotomy, the creators are probably still going to be, relatively speaking, like younger people.
Dan Grover: I could see people doing vlogs and things like that.
Rui Ma: I'll just play devil's advocate here. The thing that a lot of people have been saying is that, hey, everyone -not just in China, of course, but like definitely in China, as well as the U.S.- everyone's going after Gen Z. Everyone's like “What's going on with Gen Z? What do they want?” and we're building for them. But WeChat because it's a 1.1 billion user channel with a lot of friends and family and older people on it, then I guess the unique opportunity where they can definitely build for like slightly older population at scale. And by the way, those older people, I am one of them, older people in China, just 40 and above, so it's not, like, that old.
Dan Grover: Yeah, and I think like that doesn't get enough appreciation and praise. Because if you think about it, there's hundreds of millions of people where that is their first time, probably the first computer, their first time being on the internet. And they've made it all simple enough to work well for them.
Rui Ma: I think WeChat for sure has a monopoly on that population. Earlier, Dan, you had a really good point about how WeChats Channels. WeChat is this like incredibly - honestly Allen calls it simple - but really, it's a very complex app.
Dan Grover: It's getting more complicated than it used to be. Yeah.
Rui Ma: So it doesn't have to be like ByteDance’s Douyin, like you said, it doesn't have to win by just keeping you watching videos, right? It can do all these other things that can make money off of you, or at least benefit from your presence in all these other ways. And do you see Channels as being a really powerful standalone thing? Because people, when they compare it to Douyin that's really the question they're asking. In my opinion, it's really more of this additional thing that makes the entire ecosystem stronger. What do you think of that? Being an integrated platform?
Dan Grover: It's a reasonable enough argument just to say that, hey, we missed the boat on video, more and more content is going to be video. We need to experiment with video features in the app until we get it right. Which I think is what they're doing. I don't think they have to be Douyin nor do they have to be Netflix. They don't have to be anything.
Rui Ma: Do you think the integration - I guess this integrated vision - do you think it's an asset or actually a liability because some people are arguing that, the pure play Douyin is going to do better with video.
Dan Grover: I think it's easier for WeChat to do video than it is for ByteDance to build out the whole ecosystem of other stuff that Tencent has done. I think it's very hard to go the other way around.
Rui Ma: It sounds like what you're saying, especially with your comments earlier, is that this is a little bit of a defensive move, but it's pretty easy for WeChat to defend. Because of all these moats that they built previously with Mini Programs with of course the communication app. And then yeah, of course they have a really dominant payments [system].
Dan Grover: I think the angle that I'd be curious to analyze it from is if you think about it, - like going back to 2013, 2014, that time - there wasn't anything really competing with official accounts. For folks in the U S who remember Google reader, it's like this alternate universe where Google reader really took off and everybody had on their phones and people are telling you to add your feed to Google reader.
And they had this advantage of just being a place where a lot of content was consumed, just because there wasn't any other good way to publish to people like that, where they'd actually read it. And then ByteDance came around with Toutiao, and then they came around with video. So they're catching up on, on both of those fronts. They had this, despite all the other things they had accomplished with just being a good social network and a good communication platform - they have this advantage of just being a good place to go to consume content. I don't know how integral that ends up being to all the other parts of the, what they call the lifestyle. Right?
Rui Ma: Wait, when you say they, you meant WeChat or ByteDance?
Dan Grover: WeChat.
Rui Ma: You're saying that people probably forget that WeChat's already built this whole content ecosystem already.
Dan Grover: Yeah.
Rui Ma: So, you're saying that after WeChat launched public accounts and got all this traction off of it, Toutiao was actually coming in and trying to take it away by not having the social network. So, therefore, they had to do the whole recommendation thing and then they applied that to video. And now, WeChat is like: Oh hey, we built that whole text content that was really successful. - 360 million people still open it daily. I am one of them very loyal. After Douyin's success, we're going to apply our same product logic and leverage a lot of the same core assets, same core advantages, like the social network and apply it to video. Even the names of these products sound very similar, right?
Channels is shi-pin-hao and public accounts is gong-zhong-hao. Maybe a comment here would be the WeChat products are even named very clearly to be focused on the creator. Whereas Douyin, TikTok, etc. at least when you think about it, it's highly focused on the content.
Dan Grover: And I think that is still the experience they're trying to preserve. Like the way that gong-zhong-hao worked in contrast to Toutiao was, it's not trying to be this thing where it just like shovels content down your gullet for as long as they can.
It's something that is used as a tool, it works deterministically. You go out and subscribe to things that you want to follow and you go back there because you know what you're going to get which is very different from Toutiao and Douyin. So they're trying to do more with ML but they're still trying to preserve this nature of being a tool and working in a predictable way.
Rui Ma: Do you think like maybe the final winner is actually somewhere in the middle, like a hybrid, like what Channels is trying to do now versus, public accounts is deterministic and follower-based, not algorithmic. Do you think that the hybrid model is actually what's going to win?
Dan Grover: I'm not sure. I guess like we have to define like winning at what, right.
Rui Ma: Even Allen has already acknowledged that machine content because when it serves you up, the brainless entertainment content you're more likely to be spending time on it.
Dan Grover: Yeah. Like Allen, I have this uneasy attitude towards ML and raking. It's like the MSG of product design. Where you can always improve something by having better suggestions like you can see the same patterns being used in every app where, if you added an account, let's suggest four accounts you can add. If you join a group, let's suggest more groups you can join, if you add a friend, let’s suggest more friends again.
So you can apply ML and try to heighten the engagement in any feature, just by getting people to do more of the things that you've seen them do. And that will improve any product’s metrics, but the question is: is that worthwhile in the long run? Does it actually make these products better? I don't know, you'd be dumb to not be doing it. If you're trying to compete with ByteDance, of course you have to try something with that.
Rui Ma: What do you think is the biggest downside of that?
Dan Grover: I don't think we understand the downsides as a society. And I think what Alan seems to get, from what we saw in the talk, is that these social networks are complex systems where any little change you make in the UI has not only like the direct effect on, whatever metric it is or how people use the feature, but the sort of second order and third order effects, when you get to the quality of the content created. And I think there's a case, like from what we've seen in the U.S., where like people are joining, some of these you know, conspiracy groups and getting all sorts of weird idea. Maybe there's some butterfly effect where some of the things that all the product managers along the line changed to make their features better, like somehow lead, 15 steps later to, storming the Capitol. I don't think we have a concrete way to think about it because it is such a complex and unpredictable system.
Rui Ma: Yeah, I totally agree with you. It goes back to the point that you said much earlier, which is that, the way you set up the system, right? Whether whatever tweaks you do to the algorithm, or it's a hybrid - friend's recommendation plus machine - it will actually drive different content creation. Which will then in turn drive different content consumption.
Dan Grover: Because that's in our nature too, is we mimic what we see other people doing. Right?
Rui Ma: It's like this whole cascading effect and you can't like disentangle the creation from the consumption basically. At least he has the awareness, I'm not sure he has the answer, either.
Dan Grover: And, and like people always have, there's kind of two ways you can reason through these things. You can either do it from a purely objective metrics, point of view, or you can do it from the sort of intuitive creative line. And I think both of those are like, are crappy yet confronting this kind of question.
Rui Ma: Wait, isn't Allen more of the latter where he's a little bit more.
Dan Grover: I think so and I think that's helping them. But I think like software is getting so complex and so intertwined in so many ways with every aspect of our society, that it is hard to predict intuitively.
Like a good example is something like read receipts, in a chat app, right? Like you could either say, we're going to, we're going to test it and see if it improves responsiveness or improves message sends, or you could do what they did at the time, which is to reason through it intuitively and say, I don't think read receipts are a good idea because it's infringing on the privacy of the recipient.
And we think that the privacy of the recipient is more important than the sender having this convenience of knowing their message has been read which is a fine and good way to, to reason through it. But then you have to consider maybe, in environments where people have five different messaging apps like in the U.S. or some of these other countries they're trying to grow it into. That's actually a big downside to using the app, if you don't know what messaging apps your friend checks, then you probably would like to have that that read receipt. And it's actually going to make the product not do as well and not be as useful for the people who are on it.
So, I think there's times in making product decisions where you can get so wrapped up in the metrics that you don't have that kind of empathy. And you can't predict it, but then empathy also has limits too because like you can't really, it's very hard to when you have a product that's a billion users. I can empathize with people who are like me, we're in my cultural background. It's very hard to empathize with someone, all the way across the country from a completely different walk of life. So anyway, this is a whole other tangent, but. It's getting harder to make good product decisions at scale in tech companies in general, whether that's in the US or China.
Rui Ma: And Alan's like sort of reliance on intuition could be seen as a strength. When it's the right thing to do and you're one of the users that he understands well, but maybe you're not one of them then it feels very like uncomfortable. I think this happens with any billion user plus app, but basically when WeChat 8.0, came out there's like a lot of gripes.
Dan Grover: There’s also some pretty snarky stuff from some of my Chinese friends. I didn't quite have the time to understand it. I didn’t quite understand where it was coming from, I’ll say it that way.
Rui Ma: But anyways, I do want to ask we've talked so much about Channels and I think some of the product philosophies and it's just really hard to have a very clear-cut answer if which way of thinking about algorithms or content or whatever works.
But. What do you think? Maybe the here's an easier question, what do you think are the biggest risks to Channels right now? Could be anything. For example, I think some of their biggest risks are WeChat has never really made a product where they're going to have a lot of, in my opinion, if they're going to have live streaming, they're going to need a lot of staff doing moderating.
Dan Grover: That was, I think one of the reasons they kept gong-zhong-hao so locked down, to get an account, like you have to, in some cases, if you're a company, you have to send them your registration and you have to give them your shen-fen-zheng card and you have to do all this paperwork to be able to get an account.
And I always thought that was like a government mandated thing. Or just a thing to minimize their compliance risk, but then it seemed a bunch of companies came around and made it very easy to publish. And didn't have so much rigamarole to get started. And then they were left in this position where they were trying to be very conservative and not run afoul of those risks. And now everyone else is just running circles around them in terms of democratizing people publishing.
Rui Ma: Yeah. So, I consider that one of the big risks, how are they going to do trust and safety with everyone- 1.1 billion people now able to open up a Channels profile?
Dan Grover: Although I guess ByteDance has demonstrated that can be scaled and I guess not run afoul of the relevant regulations. I don’t know how they're doing it, but I guess if Bytedance can do it, then Tencent can do it.
Rui Ma: With a really large trust and safety team, which I think runs counter to what Allen says. I think the maybe this is good time to talk about his bias against operations, right? Strictly speaking of operations as yunying 运营 which doesn't really include trust and safety per se. It's really more about user growth and user retention and a lot of the very manual methods or at least capital-intensive methods that Chinese companies often use.
And I think Allen in his speech was very pointedly dismissive of these methods because he felt that the products should be able to live on its own merits. Do you think that kind of bias though, especially in today's very competitive China tech market, or even, or like that kind of purist thinking?
Dan Grover: I think it's a little black and white. Like I think I do see, in my line of work, I see teams make this mistake all the time. Where whatever it is that you're launching, maybe you're launching a new developer platform or a new content publishing thing or whatever new feature it is in interact.
Like the first tendency of a lot of people will be to, sign some exclusive partnerships and build some, like one-off things with them and try to bootstrap it somehow. And what that does, is it blinds you thinking about what are the fundamental mechanics and value proposition of the thing I'm making and can it eventually stand on its own legs.
So, I don't think it's right to say, Oh, we're never going to do any of that stuff. And it's always a crutch but it does mean you have to be able to walk and chew gum at the same time. And it's, it can be very hard for a lot of teams to balance these kinds of one-off. Growth efforts and partnerships with like just being very rigorous about the fundamental mechanics of whatever feature they're trying to launch.
Rui Ma: Yeah. Like maybe what Allen will end up saying in a few years is that, Oh, Operations is not so bad. Just like how he's come around into machine learning. A little bit worried, maybe it will be more of a hybrid approach to many things. I do think it's a little absolute, it sounds great as an ideal, right?
Oh, I'm a product genius and therefore the product should really drive- it's not even a that he’s a product genius- but like the product should drive everything. Maybe it's a little bit of a wishful thinking at this point
Dan Grover: Getting back to like how do you police the content now that you're opening it up to so many people, there is this like fundamental trade-off between the high operational costs of doing something like that and your ability to directly monetize the content. On Facebook, on Toutiao and Douyin, it's very simple, the whole thing is just this big old feed of content and like you can just police that stuff pretty linearly and you can monetize it pretty linearly in that you're just, one in every five pieces of content you show you're showing an ad. And WeChat has been a lot slower with that and a lot more thoughtful in the way that they monetize.
If it is that important for them to start opening up video and getting more content produced by people and more stuff being shared and if that carries a cost, in terms of policing it, then you could start to see the screws tighten a bit on monetization,
Rui Ma: A monetization of the, what do you mean of the whole app?
Dan Grover: Well, because they've been pretty conservative, right? Like they have this started to have the ads in in peng-you-quan, but not very many. They do a lot of very indirect things, like they make a little off of payments and they have their whole payments business selling financial services, they have their games business, but I don't think they could at this point make- I haven't seen the latest financial returns, so maybe this has changed a lot- but I think it's a lot harder for it to stand on its own as an independently profitable product for them.
Rui Ma: Yeah.
Dan Grover: That could start to change.
Rui Ma: Yeah, it just seemed I mean, Allen's whole philosophy is really about not bothering the user that much. Do you think that's really compatible with the Channel's product? I guess the way he's doing it is it's still relatively indirect in the sense that it's not like you open up the app and it’s pinging you with a bunch of content and recommendations. You do have to go tap into it or your shared the content from friends. So maybe it's either just like trying to shift responsibility of the recommendation off of the app itself and just onto the shoulders of your friends. So, if you have really annoying spammy friends, then you'll have a bad experience.
Dan Grover: Yeah, I think that'll work out well for them. I still think they've been relatively restrained, like they're not trying to do growthy things for you to go check out the feature. They do put the red dot on it when there's some new content, but other than that it's pretty restrained. And I think that approach can work because it's worked so well for, peng you quan and some of the other features like they never needed to do that stuff.
Rui Ma: Yeah, it really still just goes back to the fact that WeChat is a 1.1 billion user app that everyone opens basically every day. And there are so many messages flying through the system, so many opportunities for content to be shared. I don't think anyone else could get away with this kind of design. Would you agree?
Dan Grover: Yeah.
Rui Ma: Yeah, WeChat really has been very restrained about forcing you to use a product. It's more saying Hey guys, there's this new thing, it's here, use it if you want, we'll make it really easy for you to use it. But if you don't want to use it, that's okay.
Dan Grover: And in, in a lot of other companies, there would be this pressure on every one of those product teams where they go: Oh no, not enough people are using video, it's like, how do we… and instead of seeing that as a by-product, just a thing to observe where it's, Oh maybe the feature's not very quick and that's why people aren't using it. A lot of teams and companies would say, alright, we need to fix that, we need to start sending notifications, we need to, etc. So they tend to take this sort of laid-back approach where they're not trying to juice every metric for every feature. They’re looking at that as the by-product of the efforts and figuring out, okay, only a third of people use Mini Programs, I guess we have more stuff we can do.
Rui Ma: Yeah, again, I know that's only like only WeChat could do that because of how strong they are. Do you think that they are that strong? Do you think WeChat's strength - they're able to retain and how do I say this? Do you think WeChat’s long-term competitive advantage is actually this restraint?
Dan Grover: I think it has been, I don't know if it will continue to be. Like I think what they realized, and before things got totally crazy and our industry around that stuff, was that every one of these AB tests that shows some great results on how many more users you can activate has these kinds of externalities that you could never measure in the test, around things like trust and how annoyed people feel with those kinds of things. And what sort of muscle memory people are developing using the app. So, there's a lot of these kinds of intangible, sort of squishy things that are very hard to deliberately optimize for. And I think they've been good at keeping this sense of trust and control and sort of psychological safety when using the product that maybe you get a little less of with some other products that are constantly trying to grow every feature. They also have been protected to some degree from competition and so maybe it's a, it is a luxury, maybe it isn't, I don't know.
It's nice to say that when you're already on top, you can go and give these two-hour keynotes about how tasteful and restraint you are. It's hard when you're not on top to do that, right? But I think you still have to hand it to them. Like they have the courage to be that way from the very beginning. There's a great old talk, I forget which year it was, when Alan was talking about when they were version 1.0 or 2.0 or something, when they are trying to decide, how are we going to integrate with QQ?
And they could have just said they could have just built a feature where it just imports your whole QQ contact list. And all of a sudden you have a hundred friends and every metric would have shot through the roof in, you know, 2011 or 2012. And then it would have petered out and it never would have gotten to the extent that it has.
So, I think saying that now and saying oh, we have our way of doing things. Sure. that's you're bored on third and think you hit a triple, but they did have that restraint and that courage early on, I think they still deserve the credit there.
Rui Ma: I do think that's the thing I respect the most about Alan is that he has been, as far as I can tell I don't know him personally and I've only been observing from afar, he has been fairly consistent in his beliefs. He's been very stubborn. Of course. Like some people think that he's a dinosaur and maybe he should change some of his beliefs and get with the times.
But I really respect how strongly he continues to hold onto them. We've seen some movement with this 8.0 and especially earlier, we talked about his more openness towards algorithms, but I think for the most part, he really is the same product manager that I was listening to, like 10 years ago.
Dan Grover: Yeah. It's very unusual in our industry. It's as if Facebook acquired Kevin Systrom before he created Instagram, like that kind of thing never happens, right?
Rui Ma: Oh yeah. Oh, you were talking about how Allen ended up at Tencent.
Dan Grover: Yeah. So, there's not a lot of people in his position you know, he could retire tomorrow, if he wanted to right.
Rui Ma: So, I think we've talked a lot about channels but I want to get a little bit of your thoughts on just the 8.0, upgrade in general, were you excited by it? What's your sort of general feedback?
Dan Grover: Yeah. So, I think they had a good collection of good additions to the feature lineup. So obviously the focus this year was on the Channels. They also have a new status feature. There's been some evolution on Mini Programs they've changed the way that you can have a picture in picture mode.
Rui Ma: What do you think of statuses actually? So, for those of you who don't know, basically now you can set a status, which is like they have these default categories, which a lot of people were making fun of because slacking as a category, but the icon was someone sitting on a toilet. But just so you basically start off with I think 10 or 12 categories. You can then input your own status, with words, with a photo, with a location, etc. and then you can click on it and you can see who else in your friend circle is sharing the same status.
So, if you, for example, picked insomnia as your status, you can click on it and then see who else in your friend circle is also suffering from insomnia. But is this also replacing the stories thing?
Dan Grover: Yeah, they talked about that a bit in the keynote. So, I think he said - the what were they calling it?
Rui Ma: Right, time capsules.
Dan Grover: They had this kind of, back when everybody was adding storage to their product, they had this kind of, bolted on stories feature where you can shoot a little video and it shows up on your profile screen, you could see in chats that people had a video status. And I think he said that only about a million people were using it. So, it was basically a flop. So, I think they're trying to figure out what else can we do for that. And I think the intention is let's make it more lightweight for people to share. If your impulse is to share something that's going on in your day, how many choices do you have for that? You could, uh, message the group chat. You could post it in peng-you-quan. You could post a video about it and your Channel or you could use this.
So, there's obviously a lot of ways you can share on the app, so why would they need one more? Um, but if you think about it, like peng you quan, at least a lot of the content that I’ve seen on there, it's a lot of like polished and like very performative kind of stuff. So I understand why they would want to lower people's inhibitions and make it a little more casual.
I think the design of the feature is neat. I like it, it reminds me of something that Path would have shipped in like 2011 or something. I like all the little icons, you can upload a picture and they have they have a button that's unique to this feature to blur the picture that you upload, which I thought was an interesting choice.
I haven't seen a lot of my contacts using it and it still seems a bit like bolted on and disjoint from peng-you-quan. Like I would probably want to see more of that stuff in other areas of the app. So, I haven't quite figured out how does this square with the rest of the app? But it does seem like a better attempt than the last year.
Rui Ma: Yeah, I agree with you. I like that idea as well. If you're messaging someone and they have a status that's active, you can click through and see what their status says because it'll show up right next to their name. So maybe it is meant for only sort of people you're actively engaged with.
Dan Grover: they have they've full circle because if you think about it back when we were transitioning from like the desktop messaging apps, like one of the big changes that WeChat and all the other messaging apps made on that were mobile first was we're not going to have status, like they're not going to have online and offline and away and do not disturb because you always have your phone; it's always with you. So it is, it's funny that we've gone full circle back to AIM and QQ and things like that.
Rui Ma: I know that is, that's actually pretty hilarious. I wonder how many of our listeners actually know what AIM is, but anyway. Yeah, okay, let's do wrap up and predictions. So just, this is just like purely summarizing all the things that we talked about today and specifically regarding Channels. So, let's talk about what you think will be the ultimate outcome.
I'll give my take as well. What happens in five years? Okay. Wait too long. Two years. Okay. We'll probably already have a really good idea. Cause we're talking about WeChat the app that basically runs like the internet in China, so in two years, what do you think is going to happen? Do you think Channels is going to die off like the time capsules function that we were talking about earlier? Any chance of that happening? Or do you think it's going to survive and thrive more like mini apps?
Dan Grover: It wouldn't surprise me. It's only at 18% now, so it wouldn't surprise me if that grew to at least surpass what's happened with official accounts. I think it'll do just fine as a feature. I don't know if you were to look at raw time spent or any other metric. I don't know that it would go head-to-head with Douyin.
Rui Ma: Got it. Okay. So unsure whether or not it will beat Douyin on pure video, but you think usage will probably be comparable.
Dan Grover: I think so, in terms of just being another draw to keep people on the app and connecting with their friends, but I think it'll do fine.
Rui Ma: Oh, that's fair. I actually agree with you. I think that it will. Yeah, I guess my question.
Dan Grover: You’ll noticed though, in the stats that they released they told you what percent posts to Moments, they said 11% posts to Moments. What do you think it's going to be for Channels. I don't think it'll ever get to 11% for Channels.
Rui Ma: That's a really good point.
Dan Grover: What does that mean for their content ecosystem and the flywheel that they're building there? What's your prediction?
Rui Ma: I just looked up Douyin, Douyin has 600 million daily actives and about 22 million paid creators, something like that, the timeframe is not exactly the same, it's so maybe 3%, that sounds actually pretty reasonable.
Dan Grover: Yeah.
Rui Ma: So I agree with you. Yeah. If over a hundred million people are posting to Moments, which is a really like low friction and low pressure thing, because it's to your private network, then I would guess a much smaller portion of that.
Dan Grover: It doesn't have to either because Moments is still going to be there and you have statuses and you have all these other ways to share.
Rui Ma: Yeah. Yeah, exactly. So yeah, I would guess less than 5% of the people who were using Channels are actually going to be creators, that makes sense and it seems comparable to other platforms. Okay. I guess we already answered what kind of DAU are we going to see? So, let's ask you the question explicitly do you think the DAU is going to be greater than Mini Programs in two years? We're talking about viewers, we're not talking about creators, obviously.
Dan Grover: I have a theory on Mini Programs that I don't know if it's accurate but I think a big part of that, that 37% or the 400 million- I don't think it's like the stuff that we always see in the keynotes where it's like O2Os stuff and actual apps- I think a lot of Mini Programs that people are using are content Mini Programs and I think Mini Programs is cannibalizing from gong-zhong-hao and probably from a video counts. So it wouldn't surprise me, if a lot of Mini Programs out there that are just doing video that are just replicating Channels. So, it could be that both of them go up and gong-zhong-hao goes down. I don't know, like to a certain extent they're bleeding into each other and they're doing some of the same jobs.
Rui Ma: Got it. Yeah. And we already talked about in tech buzz, how actually WeChat Mini Programs is going to allow live streaming directly from the Mini Program. I guess the whole direction WeChat is going in is that you can just live stream and video from every which way inside of WeChat.
Dan Grover: Yeah. I think they're trying to see what sticks.
Rui Ma: Yeah. As you said earlier, it's this more integrated vision that they have. And it's less about does Channels itself do it really well, but does it add strength to the whole ecosystem? And right-now it seems like we're saying yes. What kind of content do you think will be ultimately thriving on Channels? For prerecorded asynchronous video, do you think that would be longer videos or do you think it will be like Douyin and Tiktok really short and meme-y stuff?
Dan Grover: I think it'll be longer videos. If you look at the interface now, it's not really very well optimized to the sort of like channel flipping that you do on Douyin. I think they're trying to get you to discover and subscribe to creators or people where you actually watched everything they'd have to say, and you want to say, I think it's going to be longer form content, but I'm not sure.
Rui Ma: I think I agree with you again, it's an intuition thing. I don't have any data to back it up whatsoever, but it feels especially with my belief that older people are going to want to see this content, it's probably going to be slightly longer.
And also it's just like a really good opportunity, right? Because that's an area where, I guess, you have Bilibili and Xigua also trying, Bytedance hasn't really quite cracked it. And Bilibili is still not nearly as big of an audience as WeChat has. What do you think is going to be the demographic that WeChat channels will be most popular with?
Dan Grover: I couldn't say I'm so out of touch with that side of things now, I haven't lived in China for five years now. I think it is going to, it is going to be a broader demographic than it is on the other routes for sure. Because that's where they are.
Rui Ma: Yeah, exactly. They're just like they have everyone. Yeah. So I'm going to say that it's going to, it's going to skew older. How do you think people will create on it? Do you think they'll be mostly creators who are only using Channels as their main source of creation, just like how on Douyin people are dedicated to that one format?
Or do you think it will be more like content creators who are already using Public Accounts and Mini Programs and, maybe selling through their Moments.
Dan Grover: I think there's going to be a lot of people using all of these. And I think Allen even mentioned in the talk, he had something in the early vision about maybe having a Mini Program running underneath the video.
We already saw that to some degree where people would launch a Mini Program and then they would use the Mini Program to get you to subscribe to the Official account, to get marketing messages, right? It wouldn't surprise me if creators and businesses use all of these surfaces in every possible way to get traffic.
I think it is neat to think of we're very used to the idea on things like Taobao and the e-commerce sites, there are these live streams where you can, they're doing like QVC style things or you can buy stuff in the stream. It wouldn’t surprise me if there was a bunch of experimentation with people doing live streams and prerecorded videos where there's some kind of interaction with the video that can be done through Mini Programs, whether it’s buying something or voting on something or playing a game or something like that. I think we'll see some experimentation.
Rui Ma: Yeah, I think it was in the gaming thing where they said that there were going to give a lot more personalization in the Mini Programs. Basically WeChat has all these components and that's both a pro and a con. It's very complex for people to navigate but if you are really serious, then there are so many tools you can use even for individual creators.
And I might be the sort of outlier here, but I agree with this one venture capitalist in Silicon Valley called Hunter Wok at Homebrew. And he basically said that he believes creators are going to be multi-skew instead of, just newsletters or just whatever. As you can see, I'm already doing that, I have podcasts, newsletters, trying to do forums, and I will randomly pop in on Clubhouse and do these YouTube videos as well. I really believe in more of a multi-skew approach. Of course, what I'm missing is that I don't have a WeChat like platform where I can integrate all of this, I would love to see someone build that whereas right now I’m in a lot of these disparate places; I actually really envy the creators in China.
Dan Grover: If they had this insight a few years earlier they wouldn't have had to make a whole separate video account. They would just have some like really good video experience in gong-zhong-hao like maybe you have a video tab inside of that. It is going to be weird because a lot of these creators they're going to want you to keep coming back and get your continuous attention. They're probably going to publish videos both on the shi-ping-hao and the gong-zhong-hao
Rui Ma: Yeah. WeChat will make the rules so it’s not spammy, but that doesn't prevent you from being spammed by your friends if or whoever you follow. If you are unfortunately following a lot of really spammy accounts, then you will be spammed.
Dan Grover: That’s on you.
Rui Ma: Unsubscribe right, unsubscribe. One final question I do want to ask just because everyone's going to ask this question and I already polled people on this. I know that we're talking about WeChat channels as part of an integrated product and I actually think that's the right way to look at it. I don't think we should be comparing it to Douyin like an individual product, but if you had to pick, just short video as a singular product, let's call it user generated video as a singular product, who do you think will win in five years?
Dan Grover: So again, I guess we have to define the criteria of winning. Is it more screen time?
Rui Ma: Let's just make it easy and say it's time spent.
Dan Grover: Boy, I think it's going to be really close because, I think I forget what the stats were from Douyin and Tiktok, I remember it being pretty impressive. What they're able to get.
Rui Ma: It's 80 to 90 plus minutes per day. Yeah.
Dan Grover: Yeah. I think that's going to be really close to call. I don't think WeChat necessarily loses if some of the video consumption time leaks out to other apps but it wouldn’t be surprising if they're neck and in the next year or two.
Rui Ma: What you think that WeChat channels in particular or do you mean WeChat overall?
Dan Grover: I think Channels is always going to be below in terms of time spent not only because Douyin has this -you think about the interaction right don't even you go in the app and you just consume video for 20 30 minutes-WeChat your friends are messaging you're going in and out. It's cannibalized by all the other WeChat features that do video. I don't think Channels in terms of pure time to watch videos I don't think that will surpass Douyin I think WeChat as a whole we'll be ahead of them I don't think we'll see it declining rapidly
Rui Ma: Got it, got it.
Dan Grover: Even if Douyin beats them on video
Rui Ma: That's probably clear in Allen's head but I feel like for a lot of other people looking at it they're really trying to look at Channels and as directly competitive with Douyin. The better question might be to ask if Bytedance is going to build up a complete ecosystem is that going to be competitive with WeChat
Dan Grover: I don't think it's ever going to happen. I don't think they can do it I don't think it's in their DNA. The time for them to start building an ecosystem would have been like five years ago when they were launching all the news and joke apps, I think it's going to be really hard for them. It's not something you can do overnight.
Rui Ma: As someone who's spent more time actually studying ByteDance and Tencent or WeChat I actually agree with you. I don't think that is how the company thinks about it. Yeah.
Dan Grover: With something like payments, they're going to have the same problem, the same challenge as Facebook, where if Bytedance does payments it's going to be to support the rest of the app and to support e-commerce and to support grip. It's not going to be its own free-standing payments business like Ali and Tencent have tried to do. So, it's a lot harder to do it in that way.
Rui Ma: Cool, all right, awesome! I have taken up too much of your time, thank you for doing this on a Saturday night and whatever is the next big news on WeChat or anything Social we'll have to talk again.
Dan Grover: sounds good.
Rui Ma: Okay. Thanks Dan!