"It's hard" is usually the wrong reason
57 of the 173 apps on this register land on a JUST PAY verdict. It'd be easy to assume that's always because the code is hard — but that's true for maybe half of them. The other half are apps where the individual features are genuinely buildable in a weekend, and the real reason to just pay has nothing to do with difficulty at all. Knowing which kind of moat you're up against is the actual useful skill here, more useful than any single verdict.
Moat 1: network effects
Notion's individual features — a block-based editor, a database view, real-time collaboration — are all things this register scores as buildable, several in the 80s and 90s out of 100. Its verdict is still JUST PAY, because a wiki or a workspace's value comes almost entirely from other people already using it. A perfectly-coded clone with zero users has zero of the value the original has. This is the one moat category where a higher score and a JUST PAY verdict genuinely coexist without contradiction — the score measures the parts, not the network, and the network is the actual product.
Moat 2: licensed content
Spotify's audio player is simple software — a file list and a play button. The catalog of millions of licensed tracks is the entire product, and licensing music from labels is a legal and business relationship, not a coding problem. The same shape shows up with Audible (audiobooks) and Splice (sample packs): the moat isn't technical at all, it's who has the rights to distribute what.
Moat 3: regulated infrastructure
This is the one category where 'build it yourself' was never actually on the table. Processing a card payment without a licensed processor, extending real credit, or executing trades on a stock exchange all require licenses individuals can't obtain by writing good code — this isn't a difficulty question, it's a legal one. Stripe, Robinhood, and Ramp are all reviewed on this register for exactly this reason: the honest verdict for genuinely regulated financial infrastructure is always JUST PAY, regardless of how simple the surrounding dashboard would be to build.
Moat 4: specialist rendering engines
A precise, real-time canvas that handles vector shapes, layers, and smooth zooming — the kind Figma, Sketch, and the Adobe design tools all need — is genuinely deep graphics engineering, refined over years by teams that specialize in exactly that. This is the closest this register gets to a pure 'it's actually hard' moat, and it's real: basic shape tools are achievable, a performant professional-grade canvas engine is not a weekend project.
Moat 5: the integration ecosystem
HubSpot and Salesforce's individual CRM features — contacts, deals, a pipeline — score well on their own. Their actual moat is being deeply integrated with hundreds of other tools a company already relies on, built up over years of partnerships. A perfectly-built CRM that only talks to itself is missing most of what makes the paid version worth using for a company already invested in that ecosystem.
Recognizing which of these five you're actually facing is the real value of reading past the stamp. A network-effect moat means build a personal version and accept it won't have the network. A rendering-engine moat means genuinely reconsider. A regulated-infrastructure moat means don't even try — that one was never a difficulty question.