The dominant factor is not technology — it is uncertainty. Every ambiguity in the brief becomes padding in the estimate. A supplier that does not know the exceptions and edge cases will assume the more expensive option. Putting two weeks into a discovery phase can cut the final cost by far more than negotiating the rate.
Connections to other systems tend to be the second big multiplier. A screen that writes to your own database is low risk; the same screen talking to an old accounting system is not. The unknown lives in the counterparty: rate limits and sandbox access, waiting on someone else’s team, inconsistent data. Ask each bidder to price integrations separately, since this is the usual source of overruns.
The requirements nobody writes down quietly rewrite the number. A tool used by a handful of staff is a very different build from the same feature set handling public traffic. Audit and compliance requirements, high availability, scalability, which is better vue or react traceability and localisation add weeks of work. Write them down at the start or you can expect them to arrive later as change requests.
The team you are quoted matters. An hourly rate reveals very little on its own: an experienced engineer at a premium rate can be less expensive in the end than two inexperienced developers who require heavy code review. Check too who else is billed: coordination, testing, infrastructure work and analysis have to be done by someone, but these should be named rather than hidden inside a blended rate.
The number in the proposal is rarely the total cost. Expect infrastructure, subscriptions and licences, logging and fintech app development alerting and .net core cross-platform development workload an ongoing support budget each year. A common working assumption is that software in active use consumes a meaningful share of the initial investment annually in fixes, updates and small changes. Ignoring this is the classic mistake.