Ask what a custom app costs and you'll get a range so wide it's useless. Five thousand dollars. Two hundred thousand dollars. Both answers are true, which is exactly the problem.
Here's a more useful way to think about it — not a price list, but the things that actually move the number.
What You're Really Paying For
Almost nobody's estimate is driven by the screens. It's driven by four things:
How many distinct things it does. Not features in a list — genuinely different capabilities. An app that logs one kind of record is a fraction of the cost of one that logs records, schedules jobs, invoices, and reports.
Whether it talks to anything else. A self-contained app is straightforward. One that syncs with your accounting software, your calendar, and a supplier's system is a different project, because you don't control those systems and they change without warning.
How many kinds of user it has. One type of user is simple. Customers plus staff plus admins means three sets of permissions, three interfaces, and considerably more that can go wrong.
Where it runs. A web app is one build. Native iOS and Android is closer to two, plus app store review, plus device testing.
If you want a cheaper app, the lever is almost always the first one. Fewer distinct capabilities. Not cheaper developers.
Rough Shapes
Without pretending these are quotes, projects tend to cluster:
A focused internal tool — one job, one type of user, replaces a spreadsheet. The smallest real project worth doing. Weeks, not months.
A customer-facing app with accounts — logins, profiles, data that must stay private and correct, probably payments. Several times the first. Months.
A platform — multiple user types, integrations with systems you don't own, reporting, an admin side. Substantially more again, and the timeline is measured in quarters.
The gap between the first and the third isn't a gap in quality. It's a gap in how many things can go wrong.
The Costs People Forget
The build is not the whole number. Budget for:
- Hosting and services. Modest, but ongoing forever.
- App store fees, if you go native.
- Maintenance. Phones update, browsers change, dependencies get security patches. Something in the range of 15–20% of the build cost per year is a realistic planning figure. Skipping it doesn't save money; it defers it and adds interest.
- The second version. Real users will want things you didn't predict. This is a sign of success, and it costs money.
- Your own time. Someone at your company has to answer questions, test, and make decisions. That's real hours from someone who already has a job.
An estimate that covers only the build is describing part of the cost.
How to Make It Cheaper Without Making It Worse
Cut scope, not corners. Half the features built properly beats all of them built badly. You can always add.
Start with the internal version. Build the tool your team uses before the one your customers see. Smaller, lower risk, and you learn what actually matters.
Use boring, standard technology. Novel choices cost more to build and far more to maintain, and make you harder to hand over to somebody else.
Don't design by committee. Every additional decision-maker adds meetings, revisions, and compromise features nobody asked for.
Be honest in the first conversation. The most expensive projects we've seen are the ones where the real requirements surfaced in month three.
The Question Worth Asking Instead
"What does an app cost?" has no answer. But this does:
What is this problem costing us right now, every year, and what would we pay once to stop it?
If somebody spends ten hours a week on a task software could do, that's roughly $13,000 a year at $25 an hour. Against that, a focused tool that eliminates most of it pays for itself quickly and keeps paying.
If you can't put a number on the problem, you're not ready to price the solution — and any developer who quotes you anyway is guessing.
We work through this with businesses in Bethlehem, Allentown, Easton and across the Lehigh Valley before anyone talks about budget. If you'd like to know what your problem is actually worth solving, start with a conversation.