How Much Does a Custom Web App Really Cost?
Nobody wants to give you a number, so I will. Here's what custom software actually costs in 2026, what drives the price up, and how to spend less without regretting it.
Every agency website says the same thing: “every project is unique, contact us for a quote.”
That’s technically true and practically useless. You’re trying to work out whether this is a $5,000 decision or a $150,000 one, and you can’t start that conversation without a number. So here are mine — real ranges, with the reasoning behind them.
The short version
| What you’re building | Realistic range | Timeline |
|---|---|---|
| Internal tool replacing a spreadsheet | $4k – $12k | 2–5 weeks |
| Customer portal (login, data, payments) | $12k – $35k | 6–12 weeks |
| Multi-tenant SaaS v1 | $35k – $90k | 3–6 months |
| Automation layer over existing tools | $3k – $15k | 1–4 weeks |
These are what a competent solo developer or small team charges. An agency with account managers and a sales floor will quote two to four times higher for the same output — you’re paying for the overhead, not the software.
What actually drives the price
The number of screens barely matters. Four things move the cost far more than anything else.
1. How many other systems it has to talk to
A form that saves to a database is cheap. A form that saves to a database, creates an invoice in QuickBooks, notifies a Slack channel, updates a HubSpot deal, and syncs to a warehouse system is five integrations — each with its own authentication, rate limits, error handling, and unhappy paths.
Every integration is roughly a week of real work once you account for the edge cases nobody mentions in the kickoff meeting. Integrations, not screens, are where budgets go to die.
2. Who is allowed to see what
“Users can log in” is a day of work. “Admins see everything, managers see their own region, contractors see only assigned jobs, and clients see a read-only view with pricing hidden” is a permissions system — and permissions systems have to be right on every single endpoint, not just the ones in the mockup.
If your app has more than two kinds of user, say so on day one. It changes the architecture.
3. Whether money moves through it
Payments bring compliance, refunds, failed charges, disputes, subscription state, tax, and receipts. Stripe makes the happy path easy and the rest is still yours to handle. Budget real time for it, or start by invoicing outside the app and adding payments in v2.
4. How much of the thinking is already done
This is the big one, and it’s the part you control.
If you arrive with a clear list of what the software must do, a sample of your real data, and one person empowered to make decisions, I can build fast. If the requirements are still being discovered while I’m building, you’re paying developer rates for what is really a strategy exercise. That’s the most expensive way to figure out what you want.
The cheapest project I ever built
A logistics company was paying two staff members to spend most of their week copying delivery confirmations from email into a spreadsheet, then re-typing the totals into their invoicing system.
I didn’t build them a web app. I built a small service that read the inbox, parsed the confirmations, and wrote the invoices directly. Three days of work. It removed about 30 hours a week of typing, permanently.
The instinct is always “we need a platform.” Often you need one small thing that removes the worst part of the week. Ask what the painful hour is, not what the ideal system looks like.
How to spend less without regretting it
Ship the smallest version that a real person uses. Not a demo — something someone’s actual job depends on. You will learn more from two weeks of real usage than from two months of planning, and everything you learn makes the next phase cheaper.
Buy the boring parts. Authentication, email delivery, payments, file storage, analytics. Using established services here isn’t cutting corners; building them yourself is spending your budget on solved problems.
Say no to the “while we’re at it” features. Every project accumulates them. Each one seems small. Together they’re usually 40% of the budget and the least-used part of the finished product.
Own your code and your data. Make sure the repository is in your account and you can export your data without asking anyone. This costs nothing to arrange upfront and is nearly impossible to fix later. If a developer resists this, that tells you something important.
What a fair quote looks like
A quote you can trust has a phased breakdown, not one number. It names what’s explicitly out of scope. It tells you what happens when something takes longer than expected, and who absorbs that. It gives you working software at intervals you can see, not a big reveal at the end.
If a quote is a single figure with a two-line description, you’re not being given a plan — you’re being given a hope.
What I’d tell a friend
Start with the process that annoys you most, not the vision that excites you most. Get something real in front of real users within a month. Keep the first version smaller than feels comfortable. Then let the actual usage tell you what to build next.
Software gets expensive when it’s built on guesses. The fastest way to lower the price is to remove the guessing.
Working out whether your idea is a three-week project or a three-month one? Book a free call — I’ll give you a straight answer, even when the answer is “you don’t need to build this.”