In almost every licensed programme the critical path is regulatory approval or a partner's onboarding process, not engineering. That single fact should reshape the plan: accelerating a path that is not the constraint produces a finished product waiting for a licence, which costs the same as being late while feeling like progress. The work worth accelerating is whatever the regulator or sponsor is waiting on.
The practical implication is that the regulatory workstream deserves the strongest people and the most attention early, which is the opposite of how programmes usually staff. Applications are slowed far more by incomplete or inconsistent submissions than by regulator throughput, and quality of submission is something you control entirely.
The second determinant is coherence between what was described to the regulator and what is being built. Where they diverge, the divergence surfaces at an assurance review and costs either rework or a regulatory conversation. Both of which land squarely on the critical path at the worst possible moment.
Partner onboarding is underestimated for the same reason. Sponsor banks, card processors and payment providers all run their own diligence on their own timelines, and several of those run in sequence rather than parallel because each wants to see the others in place first.
The useful discipline is to maintain a single view of what is genuinely blocking first customer, and to be honest when that is not engineering. Programmes that track engineering velocity as the headline metric consistently discover they optimised the wrong number.