I recently wrote about Baidu as an example of learning platforms (one of my five platform types). Located here, I said:
“My definition for learning platforms has 4 key points:
- The primary purpose of the platform is to enable interactions between user groups within a greater ecosystem. It is, first and foremost, a platform business model. It is in the business of lowering Coasean coordination costs between user groups. And this is pretty obvious for search engines, which are about decreasing the searching costs for finding information.
- The interactions on the platform improve in their accuracy and quality with more users and / or activity. The “learning” of the service is directly dependent on the interactions between user groups.
- The learning can happen in the aggregate and / or at the individual level. More overall activity can make the service better for everyone. Or an individual’s own usage can make it better for that one person. For example, YouTube customizes to you specifically. But search engines learn from everyone’s activities.
- Users can be human and/or digital agents(cameras, IoT sensors, web pages, etc.). This is kind of the big point.”
For Baidu, I talked about search engines began as relatively simple type of learning platform that enabled interactions between searchers and content created on webpages. But they have since grown increasingly complicated. This was a natural result of their objective: to catalog and make available all human information and knowledge. That is now a goal far beyond just ranking webpages.
This is why Baidu expanded from basic search into creating in-house content like videos (on iQiyi). This is how they expanded into multi-modal search, with search being done via voice and camera. This is how they merged (somewhat) into the entertainment and attention business. They are now a search engine that is somewhat competing with Tencent gaming and ByteDance videos for the attention of Chinese consumers (unfortunately).
Much of this is predictable. Organizing all human information and knowledge is, by definition, going to be increasingly difficult as the supply, format and types of knowledge increase. We are long past the days when information was just webpages with text and photos.
Which brings me to Zhihu (ZH), which has just gone public. It, like Baidu, is a learning platform. But it is very specialized and dominates the smaller and simpler Q&A niche. It has some great lessons on learning platforms. And it is easier to understand as it doesn’t have the increasing complexity problem of Baidu.
An Intro to Zhihu
Zhihu is often compared to Quora in the West. It’s a question and answer site that specializes in matching user questions with high quality, trustworthy answers (provided by other users). And those are its distinguishing characteristics on the user side. It offers higher quality and more trustworthy answers to questions.
The company began as a Q&A site in 2010, reportedly as a copy of Quora. And in the early days, it was invitation only. So highly regarded professionals and quasi-celebrities made the site feel more interesting and special. It was also a younger community that found this as a place to interact online. So yes it is a Q&A site. But it also had a community-aspect that was an important part of its early days. This was important in attracting the content creators (who work for free).
Zhihu has since evolved into a more broadly-used utility, and less of a special community. Today, it is pretty standard on every smartphone sold in China. And Zhihu is synonymous with Q&A.
The company has always struggled with monetization. It didn’t start placing ads until 2016. And paid content and memberships didn’t launch until 2018-2019. The balance between monetizing by ads and encouraging activity and community has been difficult to find. In many ways, Zhihu is more similar to Wikipedia than anything else. It is a crowd-sourced content site. Note: Wikipedia gave up on being a profitable company and is owned by a foundation.
The Zhihu IPO filing shows only two years of financials. And those two years have operating losses. 2020:
- Revenue 1.35B RMB
- Gross profit 758M RMB
- Sales & marketing expense 734M RMB
- R&D 329M RMB
- G&A 296M RMB
- Average MAU 76M
- Average monthly paying members 3M (2%)
- Revenue 670M RMB
- Gross profit 312M RMB
- Sales & marketing expense 766M RMB
- R&D 351M RMB
- G&A 351M RMB
- Average MAU 57M
- Average monthly paying members 1M (2%)
The two years do show growth in traffic and revenue. But let’s not kid ourselves here. These are big operating losses and the other years’ financials were not included for a reason.
A Cool Example of a Simple Learning Platform
The company uses lots of unusual language to describe itself. It says Zhihu is:
- A “Q&A-inspired online community”.
- A place where users come to “share knowledge, experience and insights. And to find answers”.
- One of China’s top 5 “online content communities”.
- A place to “accumulate trustworthy content through creation and engagement.”
- A “trustworthy content community” with “wisdom of the crowd.”
- A “marketplace of answers”.
There is a really interesting strategy question here. What exactly is this business?
It dominates this smaller online space and clearly has some competitive strength and defensibility. But even the company can’t tell you exactly what it is.
- It’s a “community”. A “content community”? A “Q&A inspired online community”?
- Users “share knowledge” and “find answers”.
- Content “accumulates” through “creation and engagement”.
Blah blah blah.
Here is my take:
- Zhihu is a platform business model. It is clearly 100% in the business of enabling interactions between user groups. That’s a platform. The three user groups are searchers, content creators and advertisers.
- Zhihu is a learning platform. I think the defining characteristic of the platform is that it increases in intelligence, accuracy and quality with more users and interactions. It is clearly not a marketplace (which enables monetary transactions). It is not an audience-builder (content creators aren’t getting lots of followers). And it is not a coordination or innovation platform.
- Zhihu has both direct and indirect network effects. The more content creators, the more valuable it is to users. And vice versa. Plus the more searches that happen, the more accurate it becomes at providing answers. Like Baidu, there is lots of long-tail content which makes the network increase linearly with volume for a long time.
- Zhihu’s primary intangible assets are its users, data / content and their interactions. This is how it achieves demand side scale.
- You could consider its accumulated catalog of content as a separate asset. Or you could put this under data.
I would chart all this out this way.
Think about how different this learning platform is than a search engine (my Baidu slides are here).
Yes, it does search and must match the inquiry with the best content. But it only has to look at the content (i.e., answers) on its own platform and in its own designed formats. It does not have to search all the information and knowledge of the world, in all its locations and evolving formats. This is closer to a search engine on Wikipedia than Baidu searching the world.
According to the filing, the primary users come to the app to:
- Ask questions
- Seek inspiration
- Make decisions
- Find solutions
- Have fun
Yeah. I don’t buy that.
I think the Q&A format means users are mostly coming to:
- Ask questions
- Find solutions
- Make decisions
Their questions can be basic questions. Maybe about games and gadgets. Or travel destinations. Which can lead to discussions. But questions can also go be more advanced, like medicine, physics and the law. This second category is more interesting. There are lots of communities for chatting. But there are few places to ask questions of qualified individuals and professionals – and get higher quality, more reliable and more trustworthy information and answers than you could get with a general search.
Overall, Zhihu is a nice specific, niche service. And according to the company, it had 75M MAU in Q4 2020. And this resulted in 675M monthly interactions and 25M daily searches. So they are getting traffic.
The IT is basic matching of questions and answers. The algorithms route questions to existing answers or to individuals. The ongoing Q&A interactions create, accumulate and revise the user generated content over time. The company says it has 353M pieces of content, of which 315M are questions and answers. This covers +1,000 verticals and +571,000 topics.
The bigger challenge is keeping the engagement of the content creators. Why do people spend so much time working for free on sites like Wikipedia and Quora?
On YouTube and TikTok, content creators can get status and build audiences. And they can express themselves creatively (the passion economy). And maybe they even get some monetization (rarely). But why would people spend time answering questions on Zhihu? Note: there have been rumors about large numbers of editors leaving Wikipedia.
Zhihu says it systematically “supports” and “incentivizes” content creators. I think this mostly means giving them helpful tools. But with revenue at only around $200M after ten years of operations, there isn’t a lot of profit sharing going on.
Ok. Next topic.
I Really Like Cheetahs. And Zhihu Looks Like One.
As many of you know, I like cheetahs. It’s weird how many cheetah documentaries I have watched. I think most all of them.
Cheetahs are an example of hyper-specialization. Which can be really effective in both business and natural ecosystems. Cheetahs are big cats, like lions and tigers. But they are specialized for acceleration. They can go from standing to 60mph in three steps. And this let’s them do what other big cats cannot do, which is to catch gazelles and springboks (in some situations).
But to achieve this type of acceleration, they had to give up a lot. Their claws don’t retract, as they function as cleats. Their collarbones don’t connect with their spines or rib cages, so they can hyper-extend when running. They have small jaws. They are not really that strong. They gave up a lot for their specialization.
I like businesses that are like cheetahs. That give up a lot in order to achieve specialization in 1-2 dimensions. This is how I think about Costco, Coca-Cola and China Vanke.
Compare Baidu to Zhihu as learning platforms. Baidu is all over the place. As mentioned, I like the core search function and the cloud. But the rest of the company looks like a tiger with wings, gills and a shell. It’s a mess.
Zhihu looks like a cheetah. It does one specific thing: Q&A via text interactions between searchers and content creators. And those interactions create an accumulating and evolving library of long-tail content. Simple.
Zhihu is also conveniently immune to many of the problems that plague Baidu.
- They don’t have to get into the entertainment and attention business. Baidu has gone from search to “search plus feed”. And into entertainment. And increasingly against Tencent and ByteDance in things like short video.
- Zhihu doesn’t have to do that. They also don’t have to struggle to access all the new information emerging in apps and within walled gardens (i.e., WeChat) all over the world. They only use their on-platform content.
- Zhihu don’t have to create expensive content. They rely entirely on UGC. Not big production costs like we see at iQiyi.
Overall, Zhihu has a nice niche, specialized learning platform. A small cheetah.
Great Services Don’t Mean Great Businesses
Content is a great service. But they tend to be bad businesses. For example, books are a fantastic product. You can read what Benjamin Franklin wrote and learn the important lessons from his entire life for $5. That’s amazing as a product. But books are also a bad business.
Zhihu is a great product. But it is struggling to make money and become a great business. That is common in the content world.
In my article on iQiyi, I asked “Could iQiyi (IQ) Quickly Become a Great Business?”:
“…Hamilton Helmer is looking at tech stocks with three characteristics sort of mixed together:
- A big market (existing and growth potential).
- A strong competitive position.
- Attractive unit economics and operating profit after capital costs (i.e., value creation).
I focus mostly on #2, which is a prerequisite but not a guarantee of #3. You can have a strong position in a bad or mediocre business. And there are investment strategies for that. Hamilton focuses on the small subset of businesses that are wealth generating.”
iQiyi has multiple big competitors involved in an expensive money war right now (Tencent Video, Tudou). But Zhihu looks pretty great for #2. It has clear competitive strengths and already dominates its niche.
But it is not a big market or growing market (#1). It’s just not. After ten years, it has around $200M in revenue as the market leader. And that is not going to change that much. And that’s ok. There is nothing wrong with being a dominant small company. Small giants can be great businesses. Growth is often over-rated. And its almost always over-priced.
So for Zhihu, the issue is #3. Can it get to unit positive economics? Can this company become a profitable small giant?
If it becomes profitable, it could be a nice investment at the right price
If it doesn’t, then it will likely need to become a service for a larger platform (like Mobike did). Or it will need to be a foundation like Wikipedia.
So it’s a good company to put on the watchlist. Just keep an eye on their unit economics over time. It’s easy to value.
Final Question: Does Zhihu Have a Network Effects Trap?
Ok. One last thought. This is probably my biggest worry with regards to Zhihu. They are spending a lot of money on marketing and sales.
In 2020, Zhihu spent $734M on marketing and sales, which was greater than their gross profit. That could just be the company spiking the growth before the IPO. But big marketing spend is always a red flag for a company dependent on network effects.
“I’ve mentioned many times that network effects can work in reverse just as fast. If merchants stop taking American Express cards, they are less valuable to consumers. If fewer consumers then carry them, they are less valuable to merchants. And so on. We can get a negative feedback loop.
And what do you do if you are American Express? What would you do if you are a platform based on network effects and you are seeing fewer consumers using your app each month? Or if the cost of attracting them is increasing? Or if they are spending or posting less when they do arrive? Or if your retention is falling?
This is the network effects trap that nobody talks about. Network effects make you highly dependent on keeping activity on your site.
You can see lots of network effect-based businesses getting frantic as they lose activity (and altitude).
- Ctrip and Expedia are increasingly worried about their low frequency of usage and search engines and meta search sites cutting them off from their consumers.
- Entertainment platforms like TikTok are worried they are a fad and that viewers will switch to a new popular media type, like live streaming.
- And most all online platforms have some degree of user churn. They are a leaky bucket where you constantly have to be adding new users. But for most, this is getting harder and more expensive in an age of abundance with endless choices but limited attention.”
Ok, That’s it for Zhihu. I think there are some good strategy lessons here for learning platforms. And it’s worth keeping an eye on their operating margin over time.
- Baidu’s Search Engine Explained in 3 Slides (pt 1 of 3) (Asia Tech Strategy – Daily Update)
- Baidu is Struggling in Content Creation, Push Feeds and the Attention Market (Pt 2 of 3) (Asia Tech Strategy – Daily Update)
From the Concept Library, concepts for this article are:
- Learning Platforms
- Search Engines
From the Company Library, companies for this article are:
I write, speak and consult about digital strategy and transformation.
My book Moats and Marathons details how to measure competitive advantage in digital businesses.
I also host Tech Strategy, a podcast and subscription newsletter on the strategies of the best digital companies in the US, China and Asia.
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.