How much should a small business pay for custom software?
A small business should expect custom software to cost somewhere between a few thousand and a few tens of thousands, and the honest answer is that anyone quoting a precise figure before understanding your process is guessing.
Anyone who gives you a firm price in the first meeting is either guessing or planning to change it later. That sounds evasive, so here is the useful version: the number depends almost entirely on how much of your process already exists in writing, and you can find that out yourself before you talk to anybody.
Why do quotes for the same job vary by ten times?
Because you are not asking ten people to build the same thing. You are asking ten people to interpret a description, and they will each imagine something different. One assumes a single-user tool with a spreadsheet import. Another assumes multi-user permissions, an audit trail and a mobile view, because they have been burned before by someone who wanted those things in month three.
Both are quoting honestly. They are quoting different products.
The way to collapse that variance is to be specific about the parts that actually drive cost:
- How many kinds of user? One person doing everything is a fraction of the work of three roles with different permissions.
- Does it need to talk to anything else? Integrations are where estimates go to die. Two systems that “just need to sync” can be a week or a quarter.
- What happens when it breaks at 6pm on a Friday? If the answer is “nothing, we do it on paper for a day”, that is a very different build from one that needs redundancy.
- How much of the current process is written down? If it lives in one person’s head, someone has to get it out of there first, and that is real work.
What are you actually paying for?
Not lines of code. You are paying for someone to make a hundred small decisions correctly, most of which you will never see. Whether a field is optional. What happens to half-finished records. Whether deleting a customer deletes their history.
Those decisions are the difference between software you still use in three years and software you quietly abandon.
Fixed price, hourly, or a subscription?
Each is a way of deciding who carries the risk of being wrong about scope, and scope is always wrong to some degree.
Fixed price puts the risk on the builder, so they price for the worst case and defend the boundary. You will meet the phrase “that is out of scope”. It works when the thing is genuinely well understood: a known integration, a defined report.
Hourly puts the risk on you. It is honest and it is flexible, and it makes budgeting impossible because nobody can tell you when it ends. It suits ongoing work with a trusted partner and suits nothing else.
A subscription takes scope out of the argument. You pay a known amount per month, the queue is yours to reorder, and if priorities change in week two nobody has to renegotiate. It suits a continuous stream of work, and it also suits a single well-defined job: a month is often enough to get one built and shipped, and you can stop there.
We charge a flat $2,800 a month for that last reason. Whether that is one month or twelve is up to you.
What should you do before asking for a price?
Write two pages. Not a specification: two pages of plain prose covering what happens today, where it goes wrong, and what you would want to be true instead. Include the parts you think are boring. The boring parts are usually where the cost is.
Bring that to three conversations. The quotes will come back much closer together, and the differences will be about approach rather than guesswork.
If you would rather just talk it through, that is what the first call is for.