Ask five developers to quote the same idea and you'll receive numbers spanning two orders of magnitude. Owners usually conclude that someone is being dishonest. Usually, nobody is, the quotes describe different products wearing the same name. Here's how to read them.
The first cost driver is what counts as done. A page that presents information is one kind of project. Accounts, payments, bookings, and an administration screen are each a step up, not because developers pad the bill, but because each one adds failure cases that must be handled: a declined payment, a double booking, a password reset at midnight. The $2,000 quote and the $60,000 quote often differ mainly in how many of those steps they include. When you compare quotes, compare their lists, not their totals.
The second driver is who has to use it. A tool used only by your own staff can be plain and forgiving. A product used by paying strangers has to withstand them, every confused tap, every outdated browser. Customer-facing costs more than internal, and it should.
The third driver is what happens after launch, and it's the one owners forget to ask about. Hosting, upkeep, small fixes, and improvements are ongoing, like rent and utilities on a premises. A fair project quote should come with a rough monthly figure beside it. If it doesn't, ask: "what does this cost per month once it's live?" is one of the most revealing questions you can put to a developer.
What should raise your eyebrow: a large price with no written scope (you can't hold anyone to a number without a list of what it buys); a price far below the others without an explanation of what was cut; and hourly billing with no cap on a project you described in one paragraph. What shouldn't concern you: a developer asking many questions before quoting. The one who quotes instantly is the one guessing.
The pattern I recommend to every owner: start with the smallest version that changes your business, one that does the single most valuable thing well, and budget in stages, where each stage has to earn the next. It keeps risk small, and the product grows in the direction real customers pull it rather than the direction everyone guessed at the start.