Distribution is the new moat: Evan Spiegel on building Snapchat for 15 years

Executive overview

Most consumer apps fail not because of bad product but because of broken distribution. The app store gold rush is over — people download fewer new apps, and incumbents have structural advantages that are nearly impossible to overcome.

Snap survived by focusing on connecting users to their closest friends rather than everyone, then building defensible ecosystems — platforms, creators, developers, hardware — that cannot be cloned the way software features can.

The real moat is what can't be copied: ecosystems, platforms, and hardware — not features.

Why distribution beats product-market fit

• Mobile app stores in 2011 gave Snapchat free distribution; that tailwind is gone
• TikTok solved distribution with money — billions subsidising both creators and viewers to bootstrap supply and demand
• Threads solved it with Meta's existing install base
• Connecting users to their closest friends, not their full network, was Snapchat's early distribution unlock
• Software features are trivially copied; ecosystems with millions of AR lens developers are not

How Snap builds defensible modes

• Software is not a moat — Snap learned this 15 years ago when features were cloned repeatedly
• Ecosystem strategy: creator-Snapchatter relationships, AR lens developer platform, Snapchat+
• Hardware (Specs) creates a fully vertically integrated AR stack competitors cannot easily replicate
• Network effects alone are insufficient when software cloning is trivially easy

How Snap's design team drives innovation

• Tiny design team (9–12 people), flat hierarchy, no senior titles — modelled on Loon Shots' "franchise organisation" structure
• Designers present work on day one; velocity of making ideas is the core discipline
• "If you want a good idea, you have to have lots of ideas" — preciousness kills output
• Weekly design reviews with Evan: no gate on what gets shown, hundreds of ideas reviewed
• Designers rotate across products to prevent staleness and bring fresh perspectives
• Design is a deliberate bottleneck — nothing ships without design approval, ensuring product coherence
• Hiring based purely on portfolio: range signals a designer (vs. artist), story behind work signals process

Stories: how listening to users produces non-obvious products

• Users asked for a "send all" button; Snap heard frustration with social pressure and reverse-chronological feeds instead
• Stories solved the real problem: low-pressure sharing, no public metrics, 24-hour ephemeral timeline, chronological order
• The lesson: listen deeply, empathise, then invent something new — don't build what users literally request
• Screenshot detection (pre-Apple API) was an early trust mechanic that drove early growth

Product management and the design-engineering dialogue

• Snap waited until ~200 employees to hire its first PM; designers were expected to do that work themselves
• Early model: Bobby (engineering/stats) + Evan (design) in constant dialogue — mutual respect between disciplines
• At scale, PMs play a coordination role: synthesising data science, managing cross-functional sign-offs
• The danger is PMs reducing designers to pixel-pushers; Snap deliberately kept design primary

AI's impact on Snap and the industry

• Designers at Snap now ship code; AI removes friction between idea and impact at billion-user scale
• Snap uses automated code review with AI; ~10,000 bugs caught automatically
• Internal app shake-to-report: agents debug the issue, suggest a fix, and will soon implement it
• Evan's personal use: a Glean agent that combs company dashboards and documents daily to surface hotspots
• Go-to-market agent: one-shot pipeline from product idea → spec → sign-off stakeholders → risk/legal analysis → blog post
• AI enables the flat, fast-moving leadership structure Evan always wanted

The crucible moment and Snap's strategic position

• Snap is near Fortune 500 scale, near one billion MAUs, $6B+ revenue — but not yet net income profitable
• Investment in specs is the long bet: 12 years in the making, OS released in 2024 for developer builds
• Gaming: 200 million people play games on Snapchat monthly — a growing engagement driver
• "Middle child" position: bigger than Pinterest/Reddit, far smaller than Meta/Google
• This year must prove profitability foundation before launching the next computing platform

On AI and humanity

• Technology leaders underestimate societal resistance to AI adoption
• Humanity dictates how technology is adopted — not the other way around
• Expect significant pushback on AI changes; the industry needs to put human comfort and goals first
• Snap's hardware thesis is the same: computing should fit into human life, not pull people out of it