Read on Twitter
Since there's still some confusion on why TikTok is so disruptive / successful ... this is old, old news but ...
There are a few main ingredients but the core is the Algorithmic Distribution of Content. It's not social OR interest graph based.
There are a few main ingredients but the core is the Algorithmic Distribution of Content. It's not social OR interest graph based.
User needs to do nothing except interact with the stream of content. User does not need to have friends on platform, follow anyone, nor even know who they want to follow on platform, nor indicate any interests they have to get started.
Think abt how diff that is from other apps.
Think abt how diff that is from other apps.
Again, the core value prop is that the content finds the user, vs. the user finds the content.
Of course, that means ByteDance is highly focused on / good at this recommendation portion, because otherwise, your user would churn pretty fast.
Of course, that means ByteDance is highly focused on / good at this recommendation portion, because otherwise, your user would churn pretty fast.
But you can't just be good at recommending when there is nothing TO recommend. So TikTok is also exceptionally good at getting users to become creators as well.
It makes it amazingly easy to create (try it) & gives you both ideas (campaigns) & incentives ($ subsidies) to create.
It makes it amazingly easy to create (try it) & gives you both ideas (campaigns) & incentives ($ subsidies) to create.
BTW, you don't just win with more creators of the same type, you also need diff creators / more dimensionality in your content. ByteDance is also great at this. TikTok is much earlier than Douyin in this respect, but you already see more diverse creators / content on platform.
Besides being SO EASY for users, other effects of algorithmic distribution:
1) Creators: you don't need a ready audience / existing reputation (as much).
2) Brands: Users feeds are much cleaner / tailored. Ads can be more relevant. And hello, campaigns.
1) Creators: you don't need a ready audience / existing reputation (as much).
2) Brands: Users feeds are much cleaner / tailored. Ads can be more relevant. And hello, campaigns.
And obviously, for the developers, algorithms are math. A good amount of the work is now not reliant on your cultural understanding. It can be data driven.
THAT BEING SAID, a TON of operations / growth DOES STILL depend on cultural understanding / being close to your users.
THAT BEING SAID, a TON of operations / growth DOES STILL depend on cultural understanding / being close to your users.
But that's why as Zhang Yiming says, global product, local operations.
TikTok's success (in China in particular) is not just because of its algorithmic distribution, although that conveys many advantages as cited. It's also damn good at operations & knowing what to localize.
TikTok's success (in China in particular) is not just because of its algorithmic distribution, although that conveys many advantages as cited. It's also damn good at operations & knowing what to localize.
As I posted earlier, @eugenewei's post on TikTok was awesome, and goes through some of this in more detail. Highly recommended. https://www.eugenewei.com/blog/2020/8/3/tiktok-and-the-sorting-hat
One last thing I rarely see mentioned:
Why did a Chinese co come up with this? Smartphone penetration, competitive landscape aside ... Chinese users are newer to posting AND discovering content online / "lazier". They needed a lot more handholding. And voila! Here we are.
Why did a Chinese co come up with this? Smartphone penetration, competitive landscape aside ... Chinese users are newer to posting AND discovering content online / "lazier". They needed a lot more handholding. And voila! Here we are.
