What custom software actually costs a small business

The honest answer is that nobody can quote your project from a paragraph. But “it depends” is a useless answer, so here is what it actually depends on — and roughly where the numbers land.

The range is real, and it is mostly about scope

When you ask three companies what a booking system costs, you might get $4,000, $40,000 and $180,000. None of them is necessarily lying. They are quoting three different things:

• $4,000 is one screen that does one job well, for one team, with your rules baked in.
• $40,000 is that plus accounts, permissions, payments, notifications, a mobile view, and integration with the software you already run.
• $180,000 is a product: multi-tenant, audited, on-call support, ready to sell to others.

Small businesses almost always want the first one and get quoted the second, because agencies are used to being asked for the second. The single most valuable thing you can do before asking for a quote is to decide which one you are actually buying.

What actually drives the price

How many different people use it. One team with one workflow is cheap. Customers, staff and administrators each seeing a different version of the truth is not. Every new kind of user roughly adds a wing to the building.

How much it has to talk to. A tool that stands alone is straightforward. A tool that must stay in sync with your accounting software, your inventory, and a supplier’s twenty-year-old system is a different job, and most of the cost is in the third one.

Whether money moves through it. Payments, invoices and anything a tax authority cares about carry a weight of correctness that a scheduling screen does not. Not harder to write — harder to be confident about.

How wrong it can be. Software that inconveniences someone when it fails is cheap. Software that loses an order, double-books a room, or misprices a quote needs a level of checking that shows up directly on the invoice.

How much of your process is undecided. This is the big hidden one. If your team does the same thing three different ways depending on who is working, someone has to decide which way is correct. That decision is genuinely the hard part, and it belongs to you — paying a developer to make it by guessing is the most expensive way to get it wrong.

Rough shapes, for orientation only

For a small business in the US or Colombia, a focused first version — one real workflow, one type of user, no payments — is generally a matter of weeks, not quarters. Adding customer-facing accounts, payments or integration with existing systems tends to multiply that, not add to it. Anything described as “a platform” is a different category of spending and usually a signal that the scope has not been narrowed enough yet.

The useful mental model: you are not buying software by the hour, you are buying decisions made and risk removed. A narrow scope is cheap because there is little of either.

How to get a number you can trust

1. Describe the day, not the system. “Every morning Ana exports orders and retypes them into the invoicing system, and twice a month something gets missed.” That paragraph is worth more than a feature list.
2. Name the one thing that must work. If the software did only that, would it be worth paying for? If yes, you have found your first version.
3. Ask for a fixed price on that. Any developer who understands the problem should be able to commit to a number for a narrow scope. If they cannot, either the scope is still too vague or they have not understood it.
4. Refuse phase one of five. A first phase that is useless without phase two is not a first phase, it is a deposit.

When the answer is “do not build it”

Sometimes the right recommendation is a setting in software you already pay for, a better-organised spreadsheet, or hiring someone instead. A developer who tells you that has just saved you the entire budget, and is also the one worth calling when you genuinely do need something built.

If you want a number for your own situation, book a call. Bring the messy version — it is the useful one.