TL;DR: Getting your product into users’ hands matters more than building it. When development costs collapse to zero, distribution becomes the only moat that still works. Most founders get the ratio backwards and build for two years before noticing nobody cares.

What it means

Distribution is not marketing. Marketing is one input. Distribution is the totality of choices between the finished product and the working product in the user’s hands — pricing, packaging, partnerships, channels, virality mechanics, sequencing of launches, narrative, who gets the first invite, what they see on the landing page, how they share it, why they would (distribution-is-king). Get any link wrong and the chain doesn’t close.

The reason distribution has gotten more important, not less, is the AI moment. When anyone can build the same software in a weekend, the surrounding ecosystem — brand, channel, narrative, who-already-trusts-you — becomes the only durable asset (software-dev-costs-moats). The code is commoditized. Distribution is the moat that’s left.

The argument

Distribution is a moat when code isn’t. Traditional switching costs from technology lock-in are eroding fast. Database migrations that used to take months now happen in an afternoon. What still defends value: brand awareness, channel relationships, partnerships, utility-based pricing, and proprietary data pipelines (moats). All distribution. Almost no engineering.

Lightning strikes beat drip campaigns. Sacks’s argument: startups should release news in concentrated bursts that punch through clutter, not a steady stream of feature announcements nobody reads. Combine product, customer, and milestone news into events that focus attention (startup-is-a-movement). The asymmetric attention spike beats the steady trickle every time.

Build marketing into the product. The best distribution is built-in — the YouTube effect (users share links because they want to) and the Facebook effect (friends have to join). Products with viral mechanics in the core loop have a structural advantage over products that have to buy attention every quarter (hook-model, network-effects).

Design for users, not VCs. Graham’s line: “If you start a startup, don’t design your product to please VCs or potential acquirers. Design your product to please the users. If you win the users, everything else will follow” (hackers-and-painters). Most VC-pleasing design choices actively hurt distribution because they optimize for slide-deck legibility over user delight.

Product management and marketing are one job. Tony Fadell insists there should be no separation between what the product is and how it will be explained — the story has to be cohesive from day one (build-book). The artificial PM/PMM split creates products that are hard to explain because nobody owns the full narrative. The narrative is part of the product.

The cleanest modern proof

ChatGPT is the AI-era poster child for distribution beating capability. Two and a half years before ChatGPT, GPT-3 had been available via API to anyone with a credit card. The model under ChatGPT (GPT-3.5) was barely improved over what was already public. What changed was packaging: a chat UI, no signup gate, and a shareable URL. Same model. Different distribution. 30,000 developers became 100 million normies in sixty days. Every time you hear “we have a better model” without “and here’s our distribution play,” remember that story.

Loose threads

  • Distribution moats are not all equal. Brand-as-distribution (ChatGPT, Apple) is extremely durable. Channel-as-distribution (Amazon SEO, App Store ranking) can be yanked out from under you overnight when the platform changes the rules.
  • The hardest part of distribution is admitting when you don’t have it. Many founders confuse “we got featured on Product Hunt” with a distribution strategy.
  • See google-no-moat for the inverse: a frontier-research lab that had everything except distribution and got eaten by a smaller team that did.

What links here