Innovation Monitor: The TikTok backstory

Innovation Monitor: The TikTok backstory View this email in your browser

Welcome to the Innovation Monitor. This week, we are going to take a look at one of the most compelling tech business stories of the moment that also encompasses geopolitics and culture: The Rise of TikTok.

While there are hundreds of pieces on TikTok, I wanted to help our readers understand a fuller story, especially with a focus on Toutiao, the news app created in China in 2012 by TikTok’s parent company Bytedance.

TikTok’s rise rests on a news aggregator app that could be described as a mashup of Flipboard x Apple News x Taboola. This app grew to reach hundreds of millions, and Bytedance’s valuation grew to tens of billions. However, few in the U.S. knew of its rise and reach. It was barely mentioned in the U.S. news ecosystem.

While this edition will feature no speculation on the near-term future of the product, as it’s a volatile situation, we anticipate that the conversation will become more complex and politically (and culturally) fraught. We think it’s important to understand the TikTok origin story.

So we’ll begin with some of the best early English language literature on Toutiao’s success in order to better understand it’s algorithm and the difference between content-driven recommendation systems vs. social graph recommendation systems. We’ll then explore the winding path from Toutiao ⇒ Douyin ⇒ Musical.ly ⇒ TikTok, and whether growth was fueled by a magical algorithm or just a lot of ad spend.

While this newsletter seeks to give you the backstory, the current political situation around TikTok is clearly of interest. I’d recommend the Washington Post’s “Can He Do That?” podcast for some great context on the potential ban.

On a separate note, please push your networks to fill out the 2020 Census (2020census.gov)! It can be done entirely online, takes only a few minutes, and it does not ask about one’s citizenship. The new Sept 30th deadline makes it more important than ever for everyone to fill it out.

As always, we wish you and your community safety, calm, and solidarity as we support each other through this unprecedented time. Thank you again for reading!

All best,
Erica Matsumoto ByteDance, Toutiao and TikTok: The history and context

TikTok has been saturating the news with recent talks of a Microsoft acquisition, bans by various countries, and calls for algorithmic transparency. How did ByteDance, TikTok’s parent, go from a news aggregator operating exclusively in a China-only market to top 10 in global app downloads — and becoming a US social media juggernaut, coveted by big tech and targeted by global governments?

It’s not a simple story. It’s full of conflicting ideas and practices, doublespeak, insane amounts of ad spend, and an unprecedented underdog story. It’s messy, and it starts eight years ago.

A shy Zhang Yiming founded ByteDance in 2012, launching news aggregator app Toutiao as the company’s first product. An adept engineer, Zhang approached the problem of content curation at scale algorithmically, and it turned out to be a hit: Toutiao hit 1M DAUs in four months. It wasn’t just serving people what their brains wanted (celebrity gossip and other inane content) but also fortuitous timing: “mobile penetration increased from nearly nothing in 2010 to 65% by 2014,” wrote YC partner and Toutiao investor Anu Hariharan in a 2017 piece.

Hariharan’s Toutiao long-read is one of the best, early pieces on the algorithm that would one day power TikTok — definitely set it aside as it gives a prescient glimpse of events that would follow. But I’d like to point to one passage in particular:

“Toutiao is chipping away at their end goal, which is essentially to wipe away the concept of search and just serve up aggregated, hyper-relevant content. We’ve seen ‘content aggregators’ come and go in the U.S. but it is possible that they are an idea whose time is yet to come — and that better algorithms will be the catalyst for success.”

And that time has indeed come, complete with mind-blowing acquisition offers and government bans in tow.

But back to Toutiao. Around 2014 ByteDance secured $100M in funding for the news aggregator — which is crazy, considering that:
A) it was a news aggregator, and
B) it wasn’t affiliated with the BAT (Baidu, Alibaba, Tencent) squad.

The secret tech sauce was Zhang’s vision of an algorithmic-powered feed. And what stood out was that is wasn’t powered by a social graph.

Satisfaction=F(content, user, context)

The quintessential example of the social graph is Facebook: data gathered from what you like, what your friends like, your news feed habits, and myriad other nodes, is fed into the company’s algorithms, which crank out content you’re likely to click, like, share, consume.

Instead, Toutiao doesn’t need to know any of your personal information or who your friends are. Your actions speak louder: “taps, swipes, time spent on each article, time of the day the user reads, pauses, comments, interactions with the content and location,” has been the app’s input, writes Hariharan.

This is a big deal. It laid the foundation for behind TikTok and other chart-dominating apps. This explainer from 2018 succinctly (read: coldly) summarized the TouTiao algorithm as “Satisfaction=F(content, user, context)”. No friends required.

So this helps us get some context on ByteDance’s algorithmic curation philosophy. They leveraged this algorithm in 2016 into a new short-video app called Douyin, which was effectively a clone of the trending app Musical.ly. With four years of user data and curation experience, ByteDance now had its flywheel honed. From Hariharan’s piece above:

But Douyin’s popularity centered around China’s market. And ByteDance had bigger goals — the world beyond China. And it finally secured a foothold in the U.S. market, something even BAT couldn’t do, with their ~$1B acquisition of Musical.ly — the very app that inspired Douyin — in 2017. It was a notable feat as companies like Facebook and Disney were also considering buying the app.

“In August 2018, Bytedance dropped the Musical.ly name in favor of TikTok. It took off in the U.S., and also signed up hordes of users in India, Japan and Southeast Asia,” writes the WSJ. A few months later, in Q4 2018, TikTok reached #1 in the Apple app store. At last, a Chinese giant had made its way to America’s shores and gained a significant chunk of the Gen Z market.

But TikTok didn’t get to a $50B valuation solely through the prowess of its engineering team. In 2018, the company was spending $1B on ads, according to the same WSJ piece. That included pervasive advertising on Snapchat, Facebook, and Instagram.

TikTok even considered acquiring Twitter, Quora, and Snapchat at certain points. This is coming from a company that was accused of taking ideas from Musical.ly and Snapchat. (Evan Spiegel actually visited Bytedance in China, “where he was impressed with how Toutiao served people content based on what they had clicked in the past, rather than friends’ recommendations.”)

If “ruthless” comes to mind, you’re not far off. In this Intercept piece, journalists found that TikTok’s moderators were instructed “to suppress posts created by users deemed too ugly, poor, or disabled for the platform.” Here’s the “Ugly Content Policy” the publication obtained. Despite a spokesperson saying otherwise, it’s clear what TikTok is aiming for:

This kind of environment is not that suitable for new users for being less fancy and appealing.

The TikTok representative in the Intercept piece noted the policies aren’t in use anymore. Whatever the reality is, TikTok is trying to put on a friendly facade. In June this year they actually posted an official blog on how their recommendation system works:

They even had clearly learned the language of Silicon Valley: “We’re building a global community where you can create and share authentically, discover the world, and connect with others.” That blog was followed by another: a well-timed call-to-action from the company’s chief executive Kevin Mayer. As Vox reports:

“Hours before the House antitrust subcommittee hearing featuring testimony from the CEOs of Facebook, Google, Amazon, and Apple, a blog post from… Mayer proclaimed that all platforms should ‘disclose their algorithms, moderation policies, and data flows to regulators’ and challenged the app’s competitors to follow suit.”

Concerns Around National Security

TikTok isn’t just attacking homegrown competitors — it’s also trying to distance itself from allegations that its parent company ByteDance posed a security risk: “Earlier this month, Secretary of State Mike Pompeo even threatened to ‘ban’ TikTok, though it’s unlikely the Trump administration could actually do this on its own; Joe Biden’s presidential campaign also recently instructed its staff to delete the TikTok app from their phones,” wrote Vox. Furthermore:

“‘From a national security perspective, there’s concern around using that data for espionage purposes, blackmail,’ Kiersten Todt, a scholar at the University of Pittsburgh Institute for Cyber Law, told Recode. ‘Artificial intelligence is only as good as the data that goes into it. So if the Chinese government has the most data of any other country in the world, then what it can produce from an AI perspective could potentially give it a tremendous advantage.’”

Subsequently, the app is attempting to distance itself from notions of China’s ruling communist party or the controversies surrounding big tech. This year TikTok even set up a Transparency Center in LA and plans to open one in DC, trying to become a model social media giant, all while Microsoft attempts an acquisition and the company deals with an India ban.

However this unfolds, we’re curious to see what’s next for TikTok and what this could signal to other social media platforms and tech companies around the world. THIS WEEK IN BUSINESS HISTORY August 9, 1995: The Netscape IPO

Barely a year old and never having shown a profit, software/web browser company Netscape holds its initial public offering (IPO). Five million shares went on sale for $28, double the initial estimate. By the end of the day the stock hit $78 and closed at $58. It was considered unusual at the time for a company that was unprofitable to go public, but Netflix market cap quickly doubled from $2.9 billion to $5 billion. The event has been cited as the very beginning of the Dot-Com Bubble.

Copyright © 2020 NYC Media Lab, All rights reserved.


Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.