What this covers
Patrick interviews Emmett Shear, and the conversation moves between Twitch's history and general advice for founders. Shear's startup counsel is consistent: talk to customers before forming ideas, because early ideas are hard to revise; pick a narrow initial focus, as Facebook did with Harvard and Amazon with books; and track next month's growth rather than building multi-year spreadsheets. He says he asks about market size only to gauge a founder's ambition. He also tells beginners that their weak first efforts reflect taste exceeding grasp, and that only making many things closes the gap.
The middle of the conversation covers Twitch's culture and economics. Shear describes how subscriber-exclusive emotes produced private in-joke languages inside channels, and why he sees subscriptions as a way to join a club of supporters, not as payment for perks. He extends the point beyond games: chess, music, fitness and meditation streams work because the topic is an excuse to gather. Just Chatting gets a comparison to social networks. He admits being surprised by how many streamers wanted to earn money, though most channels made about $5 a month. The Viennese cafe, where expensive newspapers drew people into shared discussion, is his analogy for what convenience can erode.
The forward-looking and more contested material concerns the unbundling and rebundling cycle. Shear predicts that subscription fatigue will drive demand for all-you-can-eat bundles, with creators already banding together. He also argues that a lot of interesting creative work is moving into private spaces to avoid outrage, and that Amazon works because it leaves acquisitions like Twitch independent.
Shear argues Twitch succeeded by treating streamers rather than viewers as the customer, building products from observed streamer behavior, and that live, person-centered interaction is a return to humanity's historically interactive entertainment.
- Viewers follow streamers, so the platform that attracts streamers wins the audience
- Products like community gifts and emotes came from watching a few power users rather than from up-front design
- Live streaming is about who the person is, not the content, which distinguishes it from other social networks
Streamers, not viewers, are the customer whose needs should drive Twitch's product.
- Small streamers most often say they need help growing and getting discovered, and Twitch spent a year or two making channels easier to find, though customers often describe problems rather than solutions.
“this is the number one thing we hear from small streamers, "Help us grow. Help me get bigger."”
- Twitch viewers segment into community loyalists, who watch mainly to belong to a group, and strategists, who watch to improve their own play and appreciate high-level performance, with community being the larger group.
“You've got community loyalists and you've got strategists.”
- Streamers want to know how much they will earn before month end and to get paid faster, and Shear treats solving these practical payment problems as the job for which the company gets paid.
“I really wish I had a better way to get to know exactly how much money I was going to make this month before the end of the month. I wish I could get paid faster.”
- Justin.tv treated the viewer as the customer because it was a show, but Twitch reversed that after data showed the streamer side had been neglected.
“So in Justin.tv, our customer was the viewer, because we were making a show.”
- Choosing the streamer as customer was a deliberate decision that clarified Twitch's product, and the team built features by working backward from what streamers wanted more than viewers, especially early on, though both must be balanced.
“starting from the streamers, thinking about it, streamer first was so powerful for us because it gave us that clarity.”
Improving the core product directly beats indirect fixes and long-range spreadsheet modeling.
- Twitch's competitive advantage comes from endlessly improving its core value proposition, since every 5 percent improvement makes it 5 percent harder for a competitor to match, with better audience growth, monetization, and community safety as the drivers.
“At the end of the day, every 5% better we get it that, it's 5% harder it is for someone else to come be better than us at it.”
- Companies are often distracted by indirect routes to improvement, reasoning that one change will enable another, and while that sometimes works, it is usually better to just improve the thing directly.
“a lot of the time, you'd have been better off just making the damn thing better in the first place.”
- The way companies get big is by focusing on the growth number coming up next month rather than a theoretical three- or five-year model, and a back-of-the-envelope model is enough, since detailed spreadsheet projections are a waste of time.
“It's always next month. What's the most important growth number to hit? The one that's coming up right now”
Creators selling directly to fans and bundling together is a trend that will keep growing.
- Once fans accumulate too many creator subscriptions, they will want a single all-you-can-eat option, and creators will band together into shared bundles, which is already happening among small creator groups on Twitch.
“We're seeing small groups of creators come together and begin to make shared bundles.”
- Shear, an investor in Substack, saw it at demo day as the same model as Twitch applied to newsletters, namely supporting individual creators directly.
“Oh, this is just Twitch, but for newsletters. It's the same idea.”
- Regardless of which company wins, the trend toward a marketplace where creators sell directly to fans of those creators will not go away and will become a big deal.
“I don't think that's going anywhere. I think that's going to be a big deal.”
Twitch's founding history and early choices explain its direction today.
- A 24/7 live reality show is intrinsically a poor idea because reality TV's appeal depends on editing weeks of footage into 30 to 60 minutes, and watching reality unedited is dull.
“a live reality television show is an intrinsically poor idea because it's only interesting when you take weeks of content and edit it down”
- Justin.tv's founders debated two directions: Michael's analytical bet on mobile video, which became Socialcam, versus Emmett's personal preference for gaming content, which he pursued because he enjoyed watching it himself.
“I reflected in the other direction. I would say that's like an analytical approach to the problem.”
- Emmett worked roughly 80 to 100 hours per week for the first two years, and the intensity crammed a large amount of learning into that period, though he believes such intensity is not meant to last forever.
“For those first two years, I probably worked 80 to a 100 hour weeks, most weeks.”
- Justin.tv was originally about non-gaming content, so the return of chess, music, and fitness on Twitch is a return to its roots rather than a new departure.
“Originally, we were Justin.tv. Justin.tv was all about things that weren't gaming.”
- Co-founder Michael's analytical bet that video would be big on smartphones led the team to build a mobile video app for recording, live streaming, and sharing, which became Socialcam and was later sold.
“that turned into Socialcam which wound up selling shortly thereafter”