Building a Multi-State Licensing Product Without Drowning in Complexity
State-by-state, line-by-line licensing requirements are a content nightmare. Here's the architecture I used to make ProfPrep's multi-state insurance prep scalable instead of unmanageable.
Building a Multi-State Licensing Product Without Drowning in Complexity
Insurance licensing is a brutal content problem disguised as a simple one. The process looks the same everywhere — education, exam, background check, application, appointment — but the details fragment by state and by line of authority. Different hour requirements, different fees, different exam splits, different state codes, multiplied across life, health, property, and casualty. I laid out the candidate-facing sequence on ProfPrep. This is about how you build a product on top of that fragmentation without it becoming unmanageable.
The trap: treating every state as bespoke
The naive build treats each state-and-line combination as a one-off — write it from scratch, maintain it separately, and watch the maintenance burden explode as you add coverage. Dozens of states times multiple lines is hundreds of bespoke products, and the moment a state changes its code, you're hunting through hundreds of places to update.
That doesn't scale, and a product that doesn't scale in a fragmented market just dies slowly under its own maintenance cost.
The architecture: separate the universal from the specific
The thing that makes it tractable is recognizing that the structure is universal and only the content is state-specific. The two-part exam shape, the general-knowledge half, the active-recall study method, the weak-category tracking — those are identical everywhere. What varies is the body of state law and the line-specific product knowledge.
So you build the universal structure once and treat state-and-line content as data that plugs into it. Adding a new state isn't building a new product; it's populating the existing structure with that state's content. The maintenance problem shrinks too — a state code change updates one content set, not the whole machine.
Why this is the whole game in fragmented markets
A lot of valuable markets are fragmented exactly like this — regulated, state-by-state, detail-heavy. They look unattractive because the complexity is real. But the complexity is also the moat: most competitors either go shallow (a generic national product that ignores the specifics) or drown (a bespoke build that can't scale). The opening is to be specific and scalable, which you only get by separating the universal structure from the specific content.
That separation is a pattern I reuse across the 2057 portfolio — build the reusable spine once, treat the specifics as data.
I'm Jesse Myers — I run 2057 Holdings and build its companies, including ProfPrep.