This article features Emmett Shear, co-founder of Twitch, explaining how a failed livestream experiment (Justin.TV) became the top gaming platform. His key insight: startups should pick one narrow customer—like Twitch focusing solely on game streamers—and win that niche, then grow month by month. He's bullish on gaming and subscription models. Top holdings mentioned: Twitch (grew from 3% of viewership to 80M+ monthly users, acquired by Amazon); Substack (compared to Twitch for newsletters, seen as promising); OnlyFans (paid $600M to creators, noted as a trend).
The episode featured Twitch founder and CEO Emmett Shear, discussing how Twitch, as the world's leading game live-streaming platform, helps streamers monetize (through subscriptions, tips, etc.) and its development trajectory after being acquired by Amazon in 2014. Key points included: the necessity
Guest: Emmett Shear, Co-founder and CEO of Twitch
Main Thread: How Twitch evolved from a failed life experiment into the world's leading game live-streaming platform, and how the "emote language" it created reshaped online interaction.
Core Judgment: Emmett Shear believes that the most critical decision in early-stage startups is not market size, but "pick a customer"—make the product extremely narrow to win a defensible niche, then expand boundaries through sustained growth in the "next month."
Emmett Shear believes that the failure of Justin.TV was precisely the prerequisite for Twitch's success—because the "live reality show" itself was a pseudo-demand, but the technical tools that supported it unexpectedly found real users.
Justin.TV's original goal was to "live stream founder Justin's life 24/7," but Emmett reflected: "live reality television show is an intrinsically poor idea because it's only interesting when you take weeks of content and edit it down to 30 minutes" (meaning: a live reality show is essentially a bad idea because it only becomes interesting when you edit weeks of content down to 30 minutes; a 24/7 live stream is just "staring out the window"). However, the technology supporting this project—real-time video streaming + chat rooms—was discovered by other users, who began using it on their own.
After the 2008 financial crisis, the team was forced to focus on profitability, and subsequently faced a strategic choice. At the time, there were two directions: one was Michael's proposed mobile video (which later became Socialcam and was quickly sold) ; the other was Emmett's gut-driven insistence on game live streaming. Emmett pointed out: "The only content on Justin TV that I enjoyed watching was the video game content, particularly the Starcraft content. It was less than 3% of our total viewership. But I thought it was great." (meaning: The only content on Justin.TV that I enjoyed watching was gaming, especially Starcraft; it accounted for less than 3% of total viewership, but I thought it was great.)
His estimation logic was: IGN had 80 million monthly active users at the time, so the game live streaming market could at least reach that scale. This judgment was later validated—"We were right; there are indeed many people like me who want to watch game live streams online."
Emmett Shear argues that Twitch’s true innovation is not live streaming technology, but the creation of a "new internet language" – emotes – and the paid club model built around it.
Origins of Emotes: Initially, they were just avatar expressions for the Justin.TV core team (Justin, Emmett, Michael). An intern named Josh uploaded his own face, which accidentally became Kappa – now a globally recognized "sarcastic/ironic" symbol on the internet. Emmett calls it: "It's a shared in-joke language" (i.e., a shared coded language). Later, Twitch allowed streamers to create exclusive emotes for subscribers, further fostering a "channel-specific coded language culture."
The Economic Logic of Subscriptions: Emmett emphasizes that subscriptions are not "paying to unlock content," but rather a "club ticket." The core motivation for users to subscribe is "want to be part of the club of people who like this stream and want to support this streamer" (i.e., wanting to become a member of the club of people who like and support this streamer). The rewards received (such as badges, exclusive emotes) are more akin to "PBS tote bags" or "nameplates on a statue" – symbols of belonging, not functional exchanges.
Key Business Discovery: Emmett admits that what surprised him most early on was: "how much streamers really wanted to be able to make it something that they did professionally or semi-professionally" (i.e., how much streamers longed to turn it into a professional or semi-professional pursuit). Even if the income was only $5/month, they still viewed it as "beer money" – but it validated the value of their effort.
Emmett Shear believes that the most fatal mistake a startup can make is "trying to serve everyone." The right approach is to "pick one customer and go all-in for them" — this matters more than any market size assumption.
The shift from Justin.TV to Twitch: In the Justin.TV era, their customer was "the viewer" — because they were producing a show. But after pivoting to Twitch, Emmett observed through data: "the viewers would go to where the streamers were. Whatever service convinced the streamers to go there, that's where the viewers would show up." So he redefined Twitch's customer as "the streamer."
Extreme narrowness: In its early days, Twitch focused on only one thing: "live streaming video games." Emmett stressed: "The biggest mistake that startups make is trying to have it all. I'm going to be ridiculously narrow because I'm going to win this niche."
Critique of market size: As an investor, Emmett said that when he looks at a founder's market assumptions, he pays more attention to the fact that "ambitious founders sort of conceptualize themselves as being part of a space bigger than what they're currently doing." But he believes the real way to grow is "next month" — "The way you get big… is next month. It's always next month," not a three- or five-year model.
Failure case: Emmett recounted that his first startup was a calendar app. When asked "Who is your calendar for?" he answered "Everyone" — "I didn't want to limit it. I feel like if I said it was smaller, I was somehow limiting my startup's potential." This is exactly the kind of mistake he later came to see as the most typical.
| Target | Guest View | Key Data |
|---|---|---|
| Twitch | Bullish (on its own business) | Early game content accounted for only 3% of watch time; later grew to 80M+ MAUs (comparable to IGN); acquired by Amazon in 2014 |
| Substack | Bullish (from investment perspective) | Users describe the model as "This is Twitch, but for newsletters" |
| OnlyFans | Neutral (trend reference) | Has paid $600M to creators |
| Socialcam | Neutral (historical comparison) | Mobile video project led by founder Michael, later sold quickly |
1. Emmett Shear: Startups should "choose one customer and serve them extremely well" rather than "serve everyone."
During the Justin.TV era, the customer was "viewers"; during the Twitch era, the customer was "streamers" — because "viewers follow the streamers." This clarity is more important than any market size assumption.
2. Emmett Shear: Emotes are the new language of the internet, and the way they are created determines their value — they are "bottom-up," with users able to spontaneously upload and spread them.
Kappa "was born" from intern Josh's face, essentially a "shared coded language." This contrasts with the earlier top-down standardization of mobile emojis.
3. Emmett Shear: The core of the subscription model is not "pay to unlock content" but "buy a club membership pass" — badges and exclusive emotes are merely symbols of belonging.
The primary motivation for subscribers is "wanting to become a member of the club that supports this streamer," not a functional need. This aligns with the logic of PBS's "eco-friendly tote bag" and statues' "nameplates."
4. Emmett Shear: The way to grow is "next month" rather than "a three-year model."
"The way you get big… is next month. It's always next month." (Becoming big is always about next month.) He opposes spending a lot of time on Excel models, advocating focusing on customer growth in "the next month."
5. Emmett Shear: Listen to customers first, then think of ideas — if the order is reversed, customer conversations are worthless.
"You can't have your great ideas and then go talk to them. Talk to the customers first and then you have the ideas." (You cannot have ideas first and then go talk to them. Talk to customers first, then generate ideas.)
6. Emmett Shear: A live-streamed reality show is essentially a terrible idea because "streaming 24/7 is like staring out the window."
Justin.TV proved: what people need is "edited content," not raw, unedited 24-hour footage. But the technical tools that supported it (real-time video + chat room) unexpectedly found real users.
7. Emmett Shear: The most interesting startup trends are often not the trend itself, but "that weird company that looks like nothing else."
"The trend is usually not the place to look. It's that weird startup in the batch that's just not doing anything like anyone else is doing that often winds up being successful." (The trend is usually not where to look. It's the weird startup in the batch that does nothing like anyone else that often ends up being successful.)
8. Emmett Shear: In the early stages of a startup, "everything you build is terrible" is normal; the key question is "are you improving?"
"The first three things you build kind of suck. The important question is, are you getting better at it?" (The first three things you build are quite bad. The key question is: are you getting better at it?) He suggests watching Ira Glass's talk on "taste and production."