Try the off-the-shelf answer first
Before you price out a custom build, rule out the boring option, since it's usually cheaper and ships this week instead of next quarter. Most of what a service business needs (scheduling, invoicing, dispatch, a customer database) already exists as a subscription: Jobber, Housecall Pro, ServiceTitan, and Workiz cover the basics for field service trades, and Airtable or Google Sheets with a form front end covers simpler tracking for close to nothing.
The honest test: open the tool and try your actual workflow in it, not a demo click-through. If it does 90% of what you need, that's the answer. Custom wins when the gap is structural, not cosmetic. A different color scheme is not a reason to build. A workflow that genuinely doesn't exist anywhere, or tools that need to talk to each other in a way no integration supports, is.
Still not sure which bucket you're in? Write down the exact thing no tool will do, and what it costs you to live without it in hours per week or jobs missed. That number tells you whether custom is worth exploring.
Rough ranges by project type
The honest answer to 'what does this cost' is always 'it depends on scope,' but wide ranges beat no ranges. Treat these as the ballpark for a solo developer or small shop, as of mid-2026, not a large agency billing account-management overhead into every hour.
- A simple internal tool, one form, one database, a view or two, no outside logins: often $2,000 to $8,000
- A customer-facing booking or quote tool tied into your existing site: often $5,000 to $20,000
- A multi-user system with logins, roles, and real business logic like dispatch or custom pricing: often $15,000 to $60,000
- Syncing with QuickBooks, a CRM, or a payment processor: add 20-40% on top, since third-party APIs are where timelines slip
- Hosting and maintenance after launch: often $50 to $400 a month, separate from the build price
Why two quotes for the same idea can be 10x apart
You'll describe the same project to three people and get back $4,000, $18,000, and $70,000. All three can be honest. Here's what drives the spread.
Scope definition is the biggest one. A vague ask ('an app to manage my jobs') gets padded, because the person quoting is pricing in unknowns you haven't answered. A tight ask ('a form that takes an address and service type, checks it against these three pricing rules, and emails me the quote') gets priced tight, because there's nothing left to guess at.
Who's building it matters too. A solo freelancer using modern AI-assisted tools builds in days what used to take a small team weeks, and that speed shows up in price. A larger agency quoting the same scope often prices in project management and a design department a small business doesn't need.
Maintenance assumptions hide inside the number. Some quotes include a year of fixes and hosting. Others are build-only, and the real cost shows up as a second invoice the month after launch. Ask what's included past launch before comparing numbers.
Last, rebuild risk: a quote that reuses proven patterns, a login system or payment flow someone's built ten times before, is cheaper and safer than one built from scratch just for you. Ask directly, "have you built something like this before?" A confident yes is worth something.
A quote with no questions attached is a red flag, not a bargain. Anyone pricing seriously asks what happens on the edge cases before naming a number.
The questions that shrink a quote before it's written
You control more of the final price than it feels like, because most of the cost lives in scope, and scope lives in decisions you can make before anyone writes code.
What's the smallest version that would actually help? Most owners describe the full dream version first, every feature handled, and that's the version that costs the most. A version that does the one thing costing you the most time, built well, beats a sprawling one built half-well. Add the second feature after the first one's proven itself.
Who exactly needs to log in? "Just me" is a much smaller build than "me, three techs, and customers checking job status." Every added user role adds real time to the build. Cut the list to who genuinely needs access on day one.
Does this need to talk to anything else? A tool that lives on its own is cheaper than one that stays in sync with QuickBooks or a payment processor in real time. If a manual export-import once a day would be fine, say so.
What does 'done' look like, in writing? A paragraph describing what the finished tool does and doesn't do prevents the scope creep that turns a $6,000 project into a $15,000 one through a dozen "can you also add" requests.
What a template or off-the-shelf tool can't do
To be fair to the custom side: some workflows genuinely don't fit any existing tool, and forcing them into one costs more in wasted hours than building the right thing once. The tell is a workaround you've built yourself, a spreadsheet shadowing the software you already pay for, or a step someone does by hand because the tool can't.
Two or three of those workarounds running at once usually means a $5,000 to $15,000 custom tool pays for itself inside a year from the hours it gets back. Multiply the hours per week on the workaround by what that time is worth, and see how many months it takes to cross the build cost.