Custom software usually carries two sorts of cost. One is the build. The other is keeping it useful, safe and running once it is live. Plenty of buyers plan for the first and get caught out by the second.
The build: a one-off project
A sound build begins with a brief discovery covering goals, users, risks and the features you cannot do without. That produces a written scope and a fixed quote. Payment is divided into milestones, so you pay as working software appears, not everything in advance.
Changes will come up. What separates a calm project from a stressful one is that each change is written down, priced and agreed before work on it begins.
Once live, a warranty period covers bugs in what was agreed. A new feature is not a bug; it belongs on the roadmap.
Once live: a monthly care plan
Software never stands still. Libraries, phones and browsers change every month, security patches come out, and your team will ask for improvements. A care plan takes in hosting, monitoring, backups, updates and a fixed amount of development time, with written response times.
The other route, ringing a developer only when something breaks, tends to cost more: problems surface late, nobody recalls how the system works, and fixes are hurried.
Questions for any supplier
Who owns the accounts and the code? What happens should you wish to leave? What exactly falls under the warranty? What are the response times, and does the contract state them? Clear answers here count for more than the day rate.