How Much Does a Custom Business App Cost? Price the Exceptions
By Mike Evan — Founder, Social Media Strategy HQ•Updated November 2026
A focused internal web app commonly costs $10,000 to $40,000. Customer-facing apps with logins, payments and integrations run $40,000 to $150,000; native iPhone and Android apps often start near $75,000. Plan 15 to 25 percent a year for upkeep. The price is set less by screens than by exceptions, integrations and the data you already have.
Ask three developers to price "an app for our crews to log jobs" and you can get quotes of $8,000, $35,000 and $120,000 in the same week. None of them is necessarily wrong. Each one pictured a different app, because the sentence describes the easy part: a form, a list and a login. What it leaves out is everything that makes your business yours — the job that gets split across two crews, the customer who pays half up front, the estimate that changes after the site visit, the manager who needs to approve anything over a certain amount.
This guide explains what actually drives the number, so you can tell which quote matches the app you need. If the thing you want is mostly a website with a booking or quote form, start with our small business website cost guide instead; if you want existing tools wired together rather than something new built, see how much AI automation costs. This post is about software your team or your customers log in to and use every day.
The Ranges, and Who Lands in Each
No-code and low-code platforms: a per-user monthly bill
Database-and-form builders let a capable office manager assemble a working tracker in a few days. For a process that looks like most companies' version of the same process, this is the right first step and nobody should talk you out of it. The bill grows with every seat you add, and the ceiling arrives the day you need a rule the platform cannot express.
Internal web app: about $10,000 to $40,000
One job done well for your own staff: a quoting tool that applies your real pricing rules, a dispatch board, an intake system that replaces a shared inbox, an inspection checklist that produces a report. A handful of user roles, one or two integrations, runs in any browser. Most small businesses that need custom software need this tier and nothing more.
Customer-facing web app: about $40,000 to $150,000
Your customers log in. They see their own orders, documents or appointments, they pay, they upload, they get notified. Now the app needs account recovery, permissions that keep one customer from ever seeing another's data, payment handling, a support path and a design that strangers can use without training. The jump in price is mostly the cost of being trusted by people who do not work for you.
Native mobile apps: often $75,000 to $250,000 and up
Separate iPhone and Android apps, a back end they both talk to, app store review, and a release process for every update. Worth it when the job demands it, which is less often than owners assume; the next section covers how to tell.
Price the Exceptions, Not the Screens
Count the screens in an app and you have counted maybe a quarter of the work. The happy path — a job is created, assigned, completed and invoiced — is quick to build in any tier. The cost lives in what happens when the path breaks: the customer reschedules after the crew is dispatched, a part is back-ordered, an invoice is disputed, an employee leaves mid-job. Every one of those is a decision your team currently makes in someone's head, and the app has to make it in writing.
That is why the cheapest way to lower an app quote is not cutting features. It is writing down the exceptions before anyone starts building. Try this before you call a single developer:
- List the last twenty times the process went sideways. Not the rules — the actual incidents. A cancelled order, a duplicate customer, a refund after a partial delivery. Each one is a requirement.
- Find every "ask Dana" step. Anywhere the answer is "check with the person who knows," there is a rule nobody has written down. The app cannot ask Dana.
- Mark which exceptions are rare enough to handle by hand. An app does not need to automate the situation that happens twice a year. Deciding that up front can take a third off the build.
Vendors who ask for this list are pricing your app. Vendors who quote from a one-paragraph description are pricing an app, and the gap between the two shows up later as change orders. Our guide to writing a website brief covers the same discipline for websites; for software, the exceptions list is the brief.
Want it done for you?
Websites, SEO, and AEO — built with Claude Code in days, not months.
Get a Custom QuoteWeb App or Phone App?
The single biggest lever on the price is whether you need native mobile apps at all. A modern web app runs in the browser on every phone, tablet and laptop, can be saved to a phone's home screen like an app, updates for every user the moment you publish, and never waits on an app store reviewer. For office staff, managers and customers checking an order status, it is almost always enough.
When native earns its cost
- Crews work where there is no signal — basements, rural sites, warehouses — and need to keep entering data offline.
- The camera, barcode scanner, GPS tracking or Bluetooth hardware is central to the job, not occasional.
- Push notifications are the product, such as dispatch alerts that must reach a phone in a pocket.
- Your customers genuinely expect to find you in an app store, which is common for memberships and repeat ordering and rare for most service businesses.
If none of those apply, build the web version first. If one of them appears later, the back end you built is reused and only the phone front end is new money.
The Two Costs Hiding Behind the Quote
Integrations: every connection is a small project
Accounting software, a CRM, payment processing, a scheduling tool, text messaging, a supplier's price list. Each connection has to be built, tested against the other system's quirks, and kept working when that system changes. A well-documented integration might be a few days of work; a poorly documented one, or a system that only exports spreadsheets, can cost more than the screens that use it. List every system the app must read from or write to, and ask the vendor to price each one separately. If the CRM itself is still undecided, settle that first — our CRM cost guide covers the choice.
Your old data: the spreadsheet has opinions
Most custom apps replace a spreadsheet, a shared drive or a stack of paper, and the history in it has to come along. Years of customer records with three spellings of the same company, typed notes where a date should be, and columns whose meaning changed in 2021. Cleaning and importing that data is real work, and it is the line item most often missing from low quotes. Ask whether migration is included, how much of your history is coming over, and who decides what to do with records that do not fit.
What AI-Assisted Development Changed, and What It Did Not
Writing code is markedly faster than it was two years ago. We build with Claude Code, and the effect on a project is real: screens, data models and integrations that used to take weeks can be working in days, and a change requested on Tuesday can be in front of your team on Wednesday. That pulls the lower tiers down and makes it practical to build small, single-purpose internal tools that would not have been worth the money before. Our custom AI apps page shows the kinds of tools that now make sense at small-business scale.
What AI did not speed up is deciding. Your team still has to agree on how the exceptions work, test the app against real days, and change habits. A vendor promising a full customer-facing app in a week is either building the happy path only or skipping the conversation that makes the app fit. Faster coding should buy you more rounds of testing with real users, not fewer.
Ownership and Upkeep
An app is not finished at launch. It needs hosting, a database, security and framework updates, and a support allowance for fixes and the small changes your team will ask for in the first months. Many owners plan for 15 to 25 percent of the original build cost per year, then adjust once real usage is known. Paid services the app relies on, such as text messaging, mapping or payment processing, bill separately and scale with use.
Ownership matters more for software than for almost anything else you buy. The source code, the database, the hosting account and any app store listings should belong to your business, with the code in a repository you control and a written handoff explaining how to deploy it. The reasoning is the same as in do I own my website, with higher stakes: a website can be rebuilt in weeks, while an app full of your operating history cannot.
Seven Questions Before You Sign
- What did you assume about our exceptions, and can we see that list?
- Is this a web app, native apps, or both — and why?
- Which systems does it connect to, and what does each connection cost on its own?
- Is moving our existing data included, and how much of our history comes over?
- When will our team first use a working version, and how many rounds of feedback are included?
- What will hosting, paid services and support cost per year after launch?
- Do we own the code, the database and the accounts, and where does the repository live?
The first question is the one that separates quotes. A vendor who can show you the exceptions they priced has priced your business; a vendor who cannot has priced a template, and you will pay the difference later. If you are still deciding whether to bring on a developer at all, our page on hiring an AI development team walks through that choice.