• 목록
• 답변
• 글쓰기
• 게시판 리스트 옵션
  • 수정
  • 삭제

The dominant factor is rarely the choice of framework — it is almost always uncertainty. Every ambiguity in the brief becomes a buffer somewhere in the quote. A vendor that has no visibility into the exceptions and edge cases has to assume a pessimistic case. Putting two weeks into a discovery phase frequently cuts the total far more than negotiating the rate.

Connections to other systems are another reliable source of cost. A screen that writes to your own database is low risk; the same functionality wired into an old accounting system is another matter entirely. The effort sits in the third party: poor documentation, long certification processes, inconsistent data. Ask any vendor to list every external system, as this is the usual source of overruns.

Quality attributes quietly rewrite the estimate. A tool used by twenty people is a very different build an mvp from the same functionality serving a hundred thousand users. Security reviews, availability guarantees, performance under load, traceability and localisation all add real engineering time. State them early or else expect them priced as extras.

The mix of people behind the number changes the arithmetic. A day rate tells you almost nothing on its own: an experienced engineer at twice the price is often cheaper per delivered feature than a pair of junior developers who require heavy code review. Ask as well who else is billed: project management, quality assurance, release engineering and UX design are legitimate costs, but they should be itemised.

The quoted figure is never what you will actually spend. Expect infrastructure, paid APIs, logging and alerting and a change budget annually. A common working assumption holds that enterprise software development with java in active use requires a recurring percentage of the original budget annually in fixes, updates and small changes. Treating the launch as the finish line remains the most common budgeting mistake.