Any firm that quotes a web application before understanding your workflow is guessing, and the guess will be wrong in whichever direction wins the deal. What follows is the shape of the real numbers, so you can tell a serious estimate from a hopeful one.
Ranges by size
Simple internal tool — $15,000 to $40,000
One workflow, a handful of screens, login, a database, basic reporting. Replaces a spreadsheet that three people fight over. Six to ten weeks.
Customer portal or line-of-business app — $40,000 to $120,000
Multiple user roles, permissions, document handling, notifications, and integration with one or two systems you already run. Three to six months. This is where most serious business applications land.
Multi-role platform — $120,000 to $350,000+
Several distinct user types, complex rules, multiple integrations, real availability requirements, and possibly a compliance regime. Six to twelve months, and it should be phased.
These assume a competent domestic team building something maintainable. You can pay less. What you generally buy with "less" is code nobody else can pick up, which becomes expensive at exactly the wrong moment.
The five things that actually move the number
- Number of user roles. Each distinct role multiplies screens, permissions, and testing. An app with one kind of user is dramatically cheaper than one with four.
- Integrations. A modern API is a day or two. An accounting package with a CSV export and no API is weeks. Ask early what you are integrating with — this is the most common estimate-breaker.
- Data migration. Getting fifteen years of records out of the old system, cleaned, and in without losing history is frequently the largest single line item, and it is almost always underestimated.
- Compliance. HIPAA, PCI, CJIS, or Section 508 accessibility change architecture, hosting, testing, and documentation. Not a surcharge — a different project.
- Design ambition. A clean, conventional interface built on a proven component library is efficient. Bespoke visual design and custom interaction patterns cost real money, and for internal tools they rarely pay back.
The recurring costs that belong in the budget
Applications are not a capital purchase you stop paying for.
- Hosting and database: $50–$600/month for most business applications.
- Third-party services: email delivery, SMS, payments, error monitoring — $30–$300/month.
- Maintenance: 15–20% of build cost annually. Dependency patching, browser changes, small fixes. Not optional; unpatched dependencies are how applications get breached.
- Enhancements: whatever you choose. Real usage always generates a list.
A $60,000 application realistically carries $12,000–$15,000 a year afterward. A proposal that does not mention this is not cheaper — it is incomplete.
Fixed price or time and materials?
Fixed price works when scope is genuinely nailed down, which in practice means after a paid discovery phase. It transfers risk to the vendor, and they price that risk in — expect a 20–40% premium and firm change control.
Time and materials is honest when scope will evolve, but only with a real budget cap, a demo every two weeks, and the right to stop. Without those it is an open tab.
The arrangement that works most often: fixed-price discovery producing a specification and a firm estimate, then a phased build with a cap per phase.
How to make an estimate trustworthy
- Pay for discovery separately and own the output. If the relationship ends, you can take the specification elsewhere.
- Ask what is explicitly excluded. The exclusions tell you more than the inclusions.
- Ask who owns the code and where it is hosted. Get it in writing before work starts.
- Ask what happens after launch, and what it costs. Silence here is the warning sign.
- Require a working demo every two weeks. Nothing exposes a drifting project faster.
The same procurement discipline that applies to choosing an IT support model applies here: get the total cost of ownership on the table before you sign, not in month seven.
We scope this work as application development engagements, and if the finished system needs monitoring and backup afterward, that is managed IT territory.
Describe what you are trying to build and we will give you a range and the assumptions behind it.