Lore

distribution strategy

How to win when software is not a moat | Evan Spiegel (Snapchat CEO)

Evan Spiegel argues that software features are not a durable moat — a lesson Snap learned 15 years ago and that the whole industry is now relearning because of AI — so lasting consumer products require solving distribution and building hard-to-copy ecosystems and hardware, while ultimately human willingness to adopt technology, not technical capability, determines how any of it plays out.

Lenny's Podcast · 2026-04-26 · English

Key ideas

  1. Consumer tech overweights product-market fit and underweights distribution, which Spiegel argues is the harder, more decisive problem, especially as AI makes software features trivial to replicate.

  2. Software is not a moat: Snap learned this 15 years ago as competitors copied Stories, swipe navigation, and Snapchat Plus (via Meta's 'Instagram Plus'), a lesson Spiegel says the whole industry is now rediscovering with AI.

  3. Because features get cloned, Snap shifted its defense from patents and network effects toward things that are hard to copy: full ecosystems connecting creators and developers, and a vertically integrated AR hardware stack.

  4. Organizationally, large-scale reliable operations and small-scale innovation require fundamentally different structures (hierarchical vs. flat), and success depends on maintaining healthy dialogue between the two rather than one dismissing the other.

  5. On AI's effect on design/PM/engineering roles, Spiegel rejects the framing of a disciplinary 'standoff,' arguing AI lets designers ship code directly while design retaining its role as a quality 'bottleneck' remains essential to a cohesive product experience.

  6. Spiegel frames the CEO role as evolving from hands-on craft work toward communication, culture-building, and strategy, casting communicating so people understand their role in the mission as a core, often underrated leadership skill.

  7. Spiegel calls the coming year a 'crucible moment': positioning Snap as tech's 'middle child' — bigger than Pinterest/Reddit but far smaller than Meta/Google — that must prove profitability while launching Specs after 12 years of investment, defining its identity for the next chapter.

  8. Spiegel's closing thesis: technology's trajectory is shaped less by what's technically possible and more by human willingness to adopt it, so he expects societal pushback against AI and argues the industry should prioritize humanity's interests over pure technological advancement.

  9. Close friends over total network size — Snapchat's early insight that most of a social network's value comes from connecting users to a handful of very close relationships (best friend, partner) rather than maximizing total friend count. Apply: Design growth mechanics around deepening a user's connection to their closest relationships instead of optimizing purely for total connections added.

  10. Ecosystem/platform-building as a moat — A strategy of building relationships between creators, developers, and users (e.g., Snap's millions of AR lenses built by outside developers) because ecosystems are far harder to copy than individual software features. Apply: Invest in platforms and third-party creator/developer tooling around a product rather than relying on features alone to stay ahead of copycats.

  11. Vertically integrated hardware stack — Snap's strategy of building its own full AR hardware stack (Spectacles/Specs) as a defensibility layer that is inherently hard for competitors to replicate. Apply: When software moats erode, shift investment toward owning hardware or infrastructure layers that require years of vertical integration to copy.

  12. Two-sided marketplace subsidization — TikTok's distribution technique of spending billions to subsidize both sides of its video marketplace, paying to acquire viewers and paying creators simultaneously. Apply: Bootstrap a two-sided marketplace by directly subsidizing acquisition on both the supply (creator) and demand (viewer) sides at once.

  13. Spectacles hardware roadmap iteration — Snap's staged hardware development path for its AR glasses: single camera, then a second camera for depth, then a display for overlaying digital objects, then a full OS for third-party developers ahead of consumer launch. Apply: Sequence hardware platform development by adding one core capability at a time (capture, depth, display, developer OS) before a full consumer release.

  14. AR content-anchoring-in-world design principle — A design principle for AR glasses that anchors digital content directly in the wearer's field of view/world rather than sticking a heads-up notification display to the face. Apply: When designing wearable AR interfaces, place content in the user's spatial context instead of routing it through a corner notification overlay.

  15. Loonshots dual-structure framework — A framework (from the book 'Loonshots') holding that large organizations need hierarchy and operational rigor to deliver reliably, while innovation requires small, flat, low-risk-averse teams, and leadership's job is fostering healthy dialogue between the two. Apply: Deliberately maintain a large structured org for scaled delivery alongside a small flat team for experimentation, and actively manage the relationship between them so neither dismisses the other.

  16. Fast thinking / slow thinking framing — An internal Snap analogy, echoing Kahneman's System 1/System 2 distinction, used to describe the relationship between the large operational org (slow, reliable) and the small design team (fast, innovative). Apply: Use the fast/slow-thinking framing to explain to teams why both quick experimental work and slow deliberate execution are needed and valued together.

  17. High-velocity design/critique process — A design practice combining Stanford's empathy-driven human-centered design process with art-school-style constant output and brutal regular critique. Apply: Pair frequent customer-empathy research with a cadence of rapid output and regular tough critique sessions to keep a design team's ideation velocity high.

  18. Deep-listening customer research method — A qualitative research approach favoring 1-2 hour in-depth conversations about how people use products in their lives, rejecting the 'survey model' of customer listening. Apply: Replace or supplement surveys with extended one-on-one conversations that probe how customers actually live with a product, then synthesize insights rather than building literally what's requested.

  19. Screenshot-detection technique — A workaround Snap engineers found to detect screenshots by inferring, from a touch event showing the finger lost contact with the screen during a hold-to-view gesture, that a screenshot had occurred, since no official OS API existed at the time. Apply: When a platform lacks an official API for a desired signal, look for an indirect device-level event (like a touch-state change) that correlates with the behavior you need to detect.

  20. Portfolio 'range' test — A design hiring heuristic distinguishing 'design' (empathizing with an audience to build something that resonates) from 'art' (personal style/self-expression) by checking whether a candidate's portfolio pieces look different from one another. Apply: When evaluating design candidates, look for varied, audience-driven work across a portfolio rather than a single repeated personal aesthetic.

  21. 'Tell me the story' interview technique — A design-hiring interview method where every candidate is asked to pick one portfolio piece they feel strongly about (love or hate) and explain why they made it and what they learned. Apply: In design interviews, ask candidates to narrate the reasoning behind one specific piece of work to reveal their process rather than just judging the output.

  22. First-day presentation ritual — A Snap design-team onboarding practice where new hires present their own work on their very first day. Apply: Have new design hires present work immediately upon joining to set a cultural expectation of high output velocity from day one.

  23. Deliberate design rotation across products — A Snap practice of moving designers across different product verticals (e.g., chat, maps, AR) rather than letting them stay on one surface for years. Apply: Periodically rotate designers to new product areas to keep perspectives fresh and prevent burnout or staleness on a single surface.

  24. 'No gate' weekly design critique meeting — A weekly design review at Snap where any idea, regardless of perceived quality, can be brought and shared without a filtering or approval step to get on the agenda. Apply: Remove gatekeeping from recurring critique meetings so unpolished or early ideas can still surface and be evaluated rather than being filtered out beforehand.

  25. AI-driven engineering guardrails — A combination of automated code review (catching close to 10,000 bugs) and an in-app 'shake to report' feature where AI agents debug issues and can suggest or soon implement fixes. Apply: Pair AI code review with an in-product bug-reporting pipeline routed to debugging agents so a broader set of contributors can safely submit code at scale.

  26. Jobs to be done framework — A framework Snap applied separately to Snapchatters (e.g., download the app, add close friends, use lenses) and advertisers (e.g., join the ad platform, configure campaigns) to organize where AI agents and teams should focus. Apply: List out the discrete 'jobs' each audience segment needs done, then use that list to decide where to deploy agents, form cross-functional teams, and track progress against business outcomes.

  27. All-hands + Q&A communication discipline — A recurring leadership practice of running company-wide all-hands and open Q&A sessions rather than relying only on written communication like email. Apply: Institute regular open Q&A sessions with the full company to build communication skill and surface employee concerns directly, even if initially uncomfortable.

  28. Glean-based personal briefing agent — An AI agent Spiegel built using Glean that integrates Snap's dashboards, documents, and data to comb through company activity and alert him to what needs attention. Apply: Build a secure enterprise-search-based agent over internal dashboards and documents to generate a personalized daily briefing for a leader.

  29. Weekly 'three things' reporting structure — A reporting format where company leaders send Spiegel weekly summaries of three things from the past week and three things looking ahead. Apply: Ask direct reports for a standardized short weekly summary (three past items, three upcoming items) to quickly surface hotspots and priorities without long reports.

  30. Go-to-market agent workflow — An agent-driven workflow at Snap that takes a product idea and in one pass writes the spec, identifies people needed for signoffs, performs legal/trust-and-safety risk analysis, and drafts go-to-market materials like blog posts. Apply: Chain an agent through spec-writing, stakeholder identification, risk analysis, and content drafting to automate an entire product launch workflow end to end.

Insights

Snap's automated code review has already caught close to 10,000 bugs, which Spiegel frames as what makes it safe to let 'everyone and anyone' contribute code at near-billion-user scale.

Screenshot detection in early Snapchat wasn't from an official Apple API — Spiegel and Bobby Murphy discovered they could infer a screenshot from a touch event showing the finger lost contact with the screen mid-hold, then notified the sender.

Stories wasn't literally the 'send all' button users asked for; Spiegel says it synthesized three separate pieces of feedback (send-all requests, complaints about social pressure/permanence, and confusion about reverse-chronological feeds) into disappearing, chronological, non-public-metric content.

Spiegel says Bobby Murphy started Snap's Gen AI lab about ten years ago, tied to the real-time on-device face-swap lens — years before generative AI's mainstream boom — and that whole businesses later built around face-swap tech 'all went away.'

Snap deliberately rotates designers across products (chat, maps, AR) rather than letting them specialize on one surface for years, arguing fresh eyes matter more than deep tenure.

A board member reportedly overruled Spiegel's preference to avoid public speaking, telling him 'this is your job,' pushing him into the all-hands/Q&A habit he now says he genuinely enjoys.

Spiegel disagrees with Keith Rabois's advice (from a separate Lenny's Podcast episode) to avoid talking to customers when building consumer products, insisting on deep 1-2 hour conversations over surveys, while still not simply building what customers literally request.

Spiegel's specific objection to heads-up-display glasses: glancing at the corner display while talking to someone means visually looking down toward the other person's crotch, an ergonomic critique of notification-based smart glasses.

Snap reportedly operated with around 200 employees before hiring its first product manager, with designers doing PM-type work themselves in the early culture.

Spiegel frames Snap as tech's 'middle child' — big enough for scale advantages but overshadowed by Meta/Google and sometimes losing attention to newer platforms — and says defining that identity is the point of the current 'crucible moment.'

«You guys have a billion monthly active users. Why is it so freaking hard to build a durable, lasting, social consumer product?»

— 00:00

«15 years ago, we essentially learned that software is not a moat, which is something that everyone is discovering today with AI.»

— 00:38

«It's certainly better than making stuff that people don't want to copy.»

— 09:56

«It's very easy to copy hardware or software features, it's very hard to copy or to replicate a full ecosystem or a platform.»

— 11:19

«They're kind of all hunched over like gremlins.»

— 15:37

«what is going to be really important with specs is showing people all the incredible new ways they can hang out and connect together with their friends wearing specs»

— 17:22

«if you want to have a good idea, you have to have lots of ideas.»

— 25:04

«hey if you need PM support why don't you just do it yourself? What's the big deal?»

— 31:38

«Well, it sounds highly dysfunctional if folks are having that standoff.»

— 35:11

«design actually has always operated as like a bottleneck at the company, which is incredibly important»

— 36:33

«staying close to your customer, staying close to the product and the way that it's serving your customer is the fundamentally most important thing you could possibly do»

— 38:41

«when you're creating a thousand new ideas, who cares that most of them are not great ideas?»

— 43:00

«I think we're going to enter a period of time where there's going to be a huge amount of societal push back on a lot of the changes that are coming with AI.»

— 64:53

«You have two ears and one mouth. Use them in that proportion.»

— 68:23

«I mean, all-time favorite is probably The Vomiting Rainbow. I think like that just brought so much joy uh to so many people, which is great.»

— 68:40

«I don't think people realize like Bobby created our Gen AI lab like 10 years ago, a decade ago or something like that.»

— 69:09

«I think there's so much to be optimistic about in terms of the way that computing can become more human and that's really what we're focused on.»

— 69:46

Reception

Reception is polarized: several viewers praise Lenny's podcast quality and admire Evan Spiegel, but many others sharply criticize his CEO performance, Snap's collapsing stock, and call the interview a 'softball' that dodged tough questions.

A dense, wide-ranging founder interview that pairs high-level distribution/moat strategy with granular design-process and org-management detail rarely shared publicly by Spiegel, though it remains an unchallenged founder narrative without independent scrutiny of Snap's business performance.

70:25

↳ Lenny's Podcast · YouTube

Watch original