Snapchat

Subscriptions on a messaging app grow slowly, so launch them early. Snapchat Plus passed 25 million subscribers and a $1 billion revenue run rate. Peter Sellis says penetration on messaging products grows slowly and roughly linearly because users must form a daily habit before they see the value.

Sellis wonders whether launching it four years earlier would have eased pressure on the ads business and left more room on content to compete with TikTok. "time is your enemy with subscription products"

Lenny's Podcast▶ 1:03:20

Snapchat grew by connecting people to their few closest friends. The old belief was that the biggest network is the stickiest and cannot be beaten. Snapchat beat much larger networks by linking people to a best friend, partner or spouse, where most of a network's value sits.

Evan Spiegel: "It wasn't about having the most friends. It was about having your close friends." The new app store helped, but he credits this way of growing with doing the real work.

David Senra▶ 42:11

Distribution, not product, is now the hardest part of consumer apps. Snap launched when the app store was new and people downloaded many apps; they download far fewer now. The two recent consumer hits had a distribution answer: TikTok spent billions subsidizing viewers and creators, and Threads borrowed Meta's reach.

Spiegel says founders spend too much time on product-market fit and too little on how people will find the product. AI speeds up building but cannot do distribution, and new platforms like mobile, and next glasses, are where it reopens.

David Senra

Users asked for a send-all button; Snap built Stories instead. Snap heard the pain under the request: pressure and judgment around sharing, plus confusion over reverse chronological feeds. Stories gave easy sharing without spam, no public likes or comments, 24-hour expiry and chronological order.

Spiegel skips surveys and suggests one or two hours with a person on how technology fits their life, and rejects Keith Rabois's view that consumer builders should skip customer talks. "We didn't build exactly what they asked for. We empathized and then came up with something new."

David Senra▶ 46:42

Screenshot alerts answered the main objection to disappearing photos. Early users said disappearing snaps made no sense since you could screenshot them. Snap noticed that a screenshot taken while a finger held down a snap triggered a touch-lost event, and used it to notify the sender.

Spiegel calls it one of the first things that got real traction. "They didn't mind if somebody saved one of their snaps. They just wanted to know."

David Senra▶ 38:01

Software is copied in days, so Snap invested in ecosystems and hardware. After rivals cloned Stories, swipe navigation and more, Snap built a platform of millions of developer-made lenses and creator relationships that has been far harder to copy. Instagram Plus launched after Snapchat Plus passed 25 million subscribers.

Network effects and patents were necessary but not enough; rivals still iterated on the ideas. Spiegel says copying beats making things nobody wants to copy, and that everyone is now learning with AI what Snap learned 15 years ago. Snap is still not net income profitable and its stock has lagged despite these defenses.

David Senra

Specs is the end of a step-by-step hardware roadmap. Snap started with a camera off the phone, added a second camera for depth, then a display, then in 2024 an operating system for developers, with the consumer Specs launch this year.

Specs anchors shared content in the real world for hanging out with friends rather than putting notifications on your face. Spiegel calls this year Snap's crucible moment: it must prove a profitable core business before Specs can win.

David Senra

Nine to twelve designers show Spiegel hundreds of ideas each week. There is no gate on what gets shown, new designers present work on day one, and hiring is based on portfolio range alone. The team's saying: "if you want to have a good idea, you have to have lots of ideas."

The habit of constant output and brutal critique comes from art and design school; Spiegel says the critique is where so much of the learning happens. Design approval is an intentional bottleneck that keeps the product cohesive, even as AI lets designers ship code.

David Senra

Leaders must keep the fast small team and the big company talking. Snap runs a public company serving almost a billion people alongside the small design team, drawing on Safi Bahcall's Loonshots: scale needs hierarchy and rigor, but hierarchy makes people avoid risk.

Each side tends to mock the other, and Spiegel says the dialogue between engineers and designers is where innovation happens.

David Senra

Subscriptions on a messaging app grow slowly, so launch them early. Snapchat Plus passed 25 million subscribers and a $1 billion revenue run rate. Peter Sellis says penetration on messaging products grows slowly and roughly linearly because users must form a daily habit before they see the value.

Sellis wonders whether launching it four years earlier would have eased pressure on the ads business and left more room on content to compete with TikTok. "time is your enemy with subscription products"

David Senra

Snap mapped its core jobs to decide where AI agents fit. It listed jobs for users and advertisers, such as getting people to download the app, add close friends and use lenses, or onboarding advertisers and configuring campaigns. This showed where to use agents, where to build AI-supported cross-functional teams, and how to track each job against business outcomes.

"If we can define a job to be done clearly enough that we can build an agent to do it, that can really create a lot of lift." Spiegel also expects human comfort, not the technology, to set how fast AI spreads, though he offers little evidence beyond unfavorable polls.

David Senra▶ 1:22:30

Snapchat's paywall keeps native style but adds its own brand flair. Its paywall avoids the default SwiftUI blue while still looking native. Jonathan Parra found a native SwiftUI-style paywall beat every custom variant by a long shot for Pyometer Plus+ and worked again in a second app.

Lenny's Podcast▶ 1:02:18