Before you ask what custom software costs, what’s the workaround costing you?
Someone’s keeping a spreadsheet alive. Someone’s re-keying the same order into a second system. You’ve learned to live with it, and now you’re wondering what it’d take to stop. That’s a better place to start than a price.
Four questions worth answering first.
Nobody can price your project from a web page, and we won’t try. These are the questions we’d ask you on a call, because the answers decide what’s worth building at all.
There’s usually one person who knows how the spreadsheet really works. What happens the week they’re out?
How many evenings go to work a system should be doing? Who’s getting tired of it?
The product you bought, or the vendor who went quiet after the kickoff. What did that cost you, beyond the invoice?
A year from now, same process, more customers. Does it still hold?
Why we don’t give you a number yet.
Any price given before someone’s seen how you work is a guess, and guesses come back as change requests. So we start with a conversation, then usually a workshop to define the scope. You get a price agreed before we start, and it holds unless you change the scope.
Why are the market’s numbers so big?
You’ll see some large figures further down. Before you do, here’s what sits behind most of them, in our experience.
The price was a guess, and the real scope turned up halfway through. Every surprise came back as a change request.
There was no retainer after launch. Small fixes waited, then piled up into a second project.
A rebuild two or three years in is where a lot of the biggest numbers come from.
What we do the other way round.
A workshop works out what’s in, what’s out and what each system really returns. Then you get a price agreed before we start, and it holds unless you change the scope.
A retainer keeps the software patched and improving a little every month, so it never turns into a second big project.
If you still want a reference point.
This is what the market publishes. It won’t tell you what yours costs. It shows what projects end up costing when they’re built before the scope is agreed and left without a retainer after launch.
| Complexity | Cost | Timeline |
|---|---|---|
| Simple or MVP | $40,000 to $120,000 | 2 to 4 months |
| Medium | $120,000 to $300,000 | 4 to 8 months |
| High complexity | $300,000 to $1,000,000+ | 8 to 18+ months |
Keyhole Software, custom software cost guide, updated August 2026, including its Clutch figures. We read Clutch’s number through Keyhole. Figures updated .
What moves the price.
Every kind of user adds screens, rules and tests.
What each system’s API actually returns is the largest unknown in most builds.
Software for a moving process costs more to build and to keep.
Audit trails, retention and access rules are real work.
Keyhole puts senior rates between $30 and $250+ an hour, depending on region.
The cheapest way to cut a price is to cut scope, which is what discovery is for.
What if nobody plans for after launch?
Keyhole estimates that keeping software running costs 15% to 25% of the build every year, and improving it another 10% to 15%. Pick a build cost to see three years when that work arrives as surprises.
With a maintenance retainer in place from launch, the same work is a small monthly cost you planned for, not a bill you didn’t see coming. So ask anyone quoting you who’ll keep it running.
So where do you start?
Not with a number. With three steps, in this order.
You tell us what’s getting in the way. We tell you honestly if we’re the right fit.
One to two days with us, and you leave with a plan, an estimate and what to leave out.
A proposal with a fixed price for the agreed scope, and a retainer for after launch.
Questions buyers ask first.
What do you charge?
That depends on what you need, and we won’t guess before we’ve seen how you work. We start with a conversation, then usually a workshop to define the scope. You get a price agreed before we start, and it holds unless you change the scope. The smallest project we take is $5,000.
Why such wide ranges?
Because the scope isn’t known yet, and most projects in those figures had no plan for after launch. A number given before scoping is a guess, and it usually comes back as change requests.
How do I compare two quotes?
Check that they cover the same scope, who’s on the team, and what happens after launch. Then compare three-year cost, not the build.
How much does custom software cost on average?
Clutch’s 2026 data puts the average project at $132,480 and 13 months. Keyhole Software puts a simple build at $40,000 to $120,000, a medium one at $120,000 to $300,000, and a highly complex one at $300,000 to $1,000,000 or more. Those are market figures, not a quote. In our experience, most of that size comes from building before the scope was agreed and having no retainer after launch. Your scope decides yours.