TikTok and the Sorting Hat

TikTok is an exception to my cultural determinist worldview

My cultural determinist worldview:
- Skeptical any of the new tech companies out of China would ever crack the U.S. market;
- Not just because US had strong incumbents or the Chinese tech companies were still in their infancy;
- Default hypothesis was that the veil of cultural ignorance was too impenetrable a barrier:
  - Non-WEIRD countries would struggle to ship into WEIRD cultures. Even more skeptical of US companies competing in China or India.
TikTok which designed by two guys in Shanghai is an exception. Why?
- It turns out that in some categories, a machine learning algorithm significantly responsive and accurate can pierce the veil of cultural ignorance. Today, sometimes culture can be abstracted.
This has significant implications for the future of cross-border tech competition, as well as for understanding how product developers achieve product-market-fit.

Muscial.ly's success and invisible asymptote

An app offered users the chance to lip synch to the official version of popular songs and have those videos distributed to an audience for social feedback;
The Musical.ly team listened to their early adopters. They received a ton of product requests, helping to inform a product roadmap. That combined with some clever growth hacks helped them achieve hockey-stick inflection among their target market;
Still, Musical.ly ran into its invisible asymptote eventually. There are only so many teenage girls in the U.S. When they saturated that market, usage and growth flatlined;
It was suddenly looked more attractive to sell the company to bytedance.

How Bytedance supercharge TikTok's growth

The one-two combination of Bytedance’s enormous marketing spend and the power of TikTok’s algorithm came to the rescue;
To help a network break out from its early adopter group, you need both to bring lots of new people/subcultures into the app—that’s where the massive marketing spend helps—but also ways to help these disparate groups to 1) find each other quickly and 2) branch off into their own spaces;
Bytedance’s short video algorithm fulfilled these two requirements. In the two sided entertainment network that is TikTok, the algorithm acts as a rapid, efficient market maker, connecting videos with the audiences they’re destined to delight. The algorithm allows this to happen without an explicit follower graph:
- The usual path for most other social networks to scale is to encourage their users to follow and friend each other to assemble their own graph one connection at a time. It's slow and has to provide some reason for people to hang around and build that graph. The trick usually is "come for the tool, stay for the network".
One challenge to scale a social network is clashing of subcultures. The first generation of large social networks has proven mostly unprepared and ill-equipped to deal with the resulting culture wars. TikTok helped to keep these distinct subcultures, with their different tastes, separated. TikTok’s algorithm sorts its users into dozens and dozens of subcultures. Not two FYP feeds are alike:
- Twitter's content moderation problems are largely the results from throwing liberals and conservatives together in a timeline together.
Another challenge within a larger social network, even subcultures need some minimum viable scale and Bytedance paid dearly to fill the top of the funnel, its algorithm eventually helped assemble many subcultures surpassing that minimum viable scale. More notably, it did so with amazing speed:
- Some people still think that a new social network will be built around a new content format (a new proof of work), but just as critical is building the right structures to distribute such content to the right audience to close the social feedback loop.

What is TikTok comparing to other social networks? Why is it so powerful?

Three purposes which I used to distinguish among networks: social capital (status), entertainment, and utility:
- While almost all networks serve some mix of the three, most lean heavily towards one of those three purposes;
- A network like Venmo or Uber is mostly about utility. A network like YouTube is more about entertainment. Some networks are more focused on social capital.
TikTok is less a pure social network, the type focused on social capital, than an entertainment network: It is trying to figure out what hundreds of millions of viewers around the world are interested in:
- It consists of a network of people connected to each other, but they are connected for a distinct reason, for creators to reach viewers with their short videos;
- An entertainment network is as an interest network: TikTok takes content from one group of people and match it to other people who would enjoy that content.
Drawbacks of the old way of building the interest graph:
- The tradition way to build out an interest-based network using a social graph has always been a sort of approximation, a hack. You follow some people in an app, and it serves you some subset of the content from those people under the assumption that you’ll find much of what they post of interest to you;
- The problem with approximating an interest graph with a social graph is that social graphs have negative network effects that kick in at scale:
  - Twitter: the one-way follow graph structure is well-suited to interest graph construction, but the problem is that you’re rarely interested in everything from any single person you follow;
  - Facebook: the bidirectional social graph eventually fills your newsfeed with a lot of noise from people you don't know, manifesting itself in the declining visit and posting frequency on Facebook across many cohorts;
  - Snapchat: struggles to differentiate between its utility as a way to communicate among friends and its entertainment function as a place famous people broadcast content to their fans;
  - There’s a reason that many people in the U.S. today describe social media as work. Apps like Facebook, Instagram, and Twitter are built on social graphs, and as such, they amplify the scale, ubiquity, and reach of our performative social burden. They struggle to separate their social functions from their entertainment and utility functions, injecting an aspect of social artifice where it never used to exist.
- TikTok doesn’t bump into the negative network effects of using a social graph at scale because it doesn't really have one. It is more of a pure interest graph, one derived from its short video content, and the beauty is its algorithm is so efficient that its interest graph can be assembled without imposing much of a burden on the user at all. It is passive personalization, learning through consumption;
- In a two-sided entertainment marketplace, they provide creators on one side with unmatched video creation tools coupled with potential super-scaled distribution, and viewers on the other side with an endless stream of entertainment that gets more personalized with time.

Potentials of TikTok

Imagine an algorithm so clever it enables its builders to treat another market and culture as a complete black box. What do people in that country like? No, even better, what does each individual person in each of those foreign countries like? You don't have to figure it out. The algorithm will handle that. The algorithm knows;
Now imagine that level of hyper efficient interest matching applied to other opportunities and markets. Personalized TV of the future? Check. Education? I already find a lot of education videos in my TikTok feed, on everything from cooking to magic to iPhone hacks.

Ref: