The constraint is which competencies are present, not how many people. A launch needs someone accountable for the regulatory relationship, someone who owns the ledger and its invariants, someone who owns payment rail integrations, a compliance operator who will run the controls rather than only design them, and product and engineering leadership that can decide quickly. Programmes fail from missing one of those entirely far more often than from being understaffed across all of them.
The most commonly missing competency is ledger ownership. It gets treated as an integration task assigned to whoever is available, because the core is bought and the work looks like configuration. Postings, limits and reconciliation encode the operating model, and having nobody whose job is to be right about them is how a programme discovers in month nine that its ledger cannot express a product it already sold.
The second is a compliance operator as distinct from a compliance adviser. Someone must actually run screening, review cases and own the queue on a Tuesday. Advisers write the policy; a programme that has only advisers arrives at launch with an excellent policy and nobody to execute it.
Adding people rarely accelerates the critical path, because the critical path is usually the licence or a partner's timeline instead of engineering throughput. Programmes that respond to schedule pressure by hiring typically get slower, since the new capacity lands on work that was not the constraint.
The composition changes sharply at launch. The team that builds is not the team that runs. Operations, support and monitoring become the bulk of the work, and planning that transition late is a common and avoidable failure.