The honest answer to almost every cost question is that the number depends on three decisions, and two of them are yours. This post walks through those decisions with real ranges attached, so you can arrive at a vendor conversation already knowing which band you are in.
Decision one: are you buying a system or a rebuild?
Replacing a working process is cheaper than replacing a broken system, because a broken system comes with data you still have to migrate and users who already distrust the idea. Budget for the migration, not just the build.
Rule of thumb
If your data lives in more than three spreadsheets maintained by more than two people, migration will be 20–30% of the project, not a rounding error.
Decision two: how much of your process is genuinely unique?
Every business believes its process is unusual. Usually two or three parts genuinely are, and those parts are where custom software earns its money. The rest should look boring on purpose.
What we ask on the first call
- Which step costs you money when it goes wrong?
- Who currently fixes it, and how long does that take?
- How many times a day does it happen?
- What would you do with that time back?
Those four answers give a defensible budget faster than any feature list, because they turn the project into arithmetic instead of ambition.
Want the number for your specific case?
Thirty minutes, free. You will leave with a scope and a band, whether or not you hire us — book a consultation.
Decision three: who maintains it in year two?
The cheapest quote usually assumes nobody ever changes anything. Ask every vendor what handover includes, who owns the repository, and what a change request costs after launch. If those answers are vague, the low price is not a price — it is an option they can reprice later.