What Custom Software Actually Costs

Most agencies will not put a number on a page. That is understandable, because the honest answer depends on what you are building, and a number without context invites the wrong comparison.

It is also unhelpful when you are trying to work out whether to start a conversation at all. So here is roughly where we sit, and more usefully, what moves it.

The ranges we work in

Websites start around five thousand dollars. Most land between ten and fifteen thousand. Complex builds with a lot of custom functionality have gone up to fifty or sixty thousand.

Ongoing operations and maintenance runs from five hundred to five thousand a month, depending on how much is connected and how much is actively running.

Custom applications vary more widely than websites, because the range of what people mean by an app is much wider. We would rather scope yours than quote a range broad enough to be meaningless.

What actually drives the number

How many systems it touches. This is the biggest single factor and the one most often underestimated. A tool that stands alone is straightforward. One that reads from your accounting system, writes to your CRM and sends email through a third service is three integrations, each with its own authentication, its own failure modes and its own vendor who may change something next year.

Whether the process is written down. If nobody has documented how the work currently happens, discovering it is part of the project. That is time, and it is unavoidable, because building the wrong process efficiently is worse than not building it.

How much judgment is involved. Steps with a clear rule are cheap. Steps where an experienced person weighs several things and decides are expensive, because you have to capture what they are weighing. That is often the most valuable part of the work and it is rarely the part people budget for.

Who has to approve what. Approval steps are not free. Each one needs somewhere for a person to see the item, act on it, and have that action stick. Worth it for anything customer-facing, and worth deciding deliberately rather than adding everywhere.

Data you already have. Migrating five years of records out of a system that was never designed to let them go is its own project. Sometimes a large one.

The costs that arrive later

It needs maintaining. Tools change, APIs get deprecated, certificates expire. Something connected to five other systems will need attention several times a year whether or not anything is wrong on your side.

Somebody has to run it. If a workflow drafts things for approval, a person has to approve them. That is far less work than doing it manually, and it is not zero.

Usage costs, if AI is involved. Model usage is metered. For most small business workflows this is modest, but it is a real line item and it should be discussed rather than discovered.

Where the money is usually wasted

Building the whole thing before finding out whether the first part works. This is the most expensive mistake we see, and it is almost always driven by a well-meaning attempt to scope everything up front.

The alternative is to ship one workflow end to end, live with it for a month, and let what you learn shape the next one. It costs slightly more per unit of work and considerably less overall.

What to do with this

If your problem costs you a few hundred dollars a month, a custom build is probably the wrong answer and we will say so. If it costs a few thousand, the arithmetic usually works, and the question becomes which part to do first.

That is the conversation worth having, and it does not require a specification. Get in touch and we can put a real number on your actual project rather than a range on a generic one.