What a field service app costs for an HVAC, plumbing, or electrical company

$52,000 – $78,000, over 15–23 weeks. That buys one app that runs on both iPhone and Android without paying to build it twice, sign-in, three kinds of user, work that keeps going with no service and syncs once the phone has a bar, photo capture, an office portal for scheduling and assigning jobs, work order documents, push notifications, and a connection to whatever you use for invoicing.

That range is wide for a reason. At the low end your accounting software is a modern one, somebody at your shop can answer a question inside a week, and your people already agree on how a job should flow. At the high end none of that’s true. How the estimate is put together explains why.

$52,000 – $78,000

Full build

$40,000 – $61,000

The narrower first version, 22% less

Your shop probably isn’t that shop. Fewer roles, no accounting connection, no signatures, and the number moves. Price your own version. The description’s already filled in, and editing it to match how your jobs really run takes about two minutes.

The office portal, at roughly $10,500. That’s about 16% of the build, and the biggest single thing on it. Most contractors don’t expect that, because it’s the part they describe in half a sentence near the end of a meeting: “and the office needs to see the jobs.”

$10,500

The office portal. Without it, this build is $43,000 – $66,000

$7,000

The third kind of user. With two, it’s $46,000 – $70,000

It’s a second app. Scheduling, assignment, customer records, searchable job history, permissions so a dispatcher can’t see payroll, and the reports somebody will ask for eventually. All of it needs building and all of it needs testing. The tech app is the half you picture. The office portal is the half nobody budgets for, and it’s the bigger number.

How it gets described

“The office needs to see the jobs and assign them out.” One sentence, usually with a hand gesture.

What it turns out to be

A scheduling board, a job detail view, customer records, a history you can search, permissions, and reports. Take it out and this build drops from $52,000 – $78,000 to $43,000 – $66,000.

After that it’s your third kind of user. Your tech wants today’s jobs and nothing else. Dispatch wants the whole board and the ability to move work around. You want to know what was profitable last month. Three people, three sets of screens, and the third set costs $7,000.

If your crews work in basements, crawl spaces, mechanical rooms, or out past the last tower, yes. It’s two jobs and you pay for both. Holding the work on the phone runs about $5,500. Getting it back to the office runs another $4,500. An app that drops a job when a tech walks downstairs doesn’t get opened twice.

Holding the work is the easy part

Keeping a job on the device with no service is a solved problem, and it isn’t where the money goes.

Getting it back is not

Your tech edits a job in a crawl space. Dispatch hands that same job to somebody else while he’s down there. Both changes are real. Something has to pick a winner. And it has to do that without losing work either of them believes they saved.

And it gets tested on bad service, not no service

Not airplane mode. One bar. A connection that dies halfway through uploading photos. A phone that grabs a customer’s Wi-Fi with no internet behind it. That’s how your techs actually work, and it’s harder to set up in testing than just switching the signal off.

Probably. Price it before you price a build. Both publish their rates, somebody else maintains them, and a 12-tech shop can be running on either one next week.

$4,788/yr

Jobber Plus, 15 seats, annual billing, August 2026

$3,588/yr

Housecall Pro MAX, annual billing, August 2026

Do the math. Against Jobber Plus at $4,788 a year, a $52,000 – $78,000 build takes about 14 years to break even. Against Housecall Pro MAX it’s nearer 18. Both are before anybody maintains what you built. If something on the shelf does what you need, buy it. For most shops your size that’s the answer.

ServiceTitan is a different animal. It doesn’t publish pricing, it’s quoted per tech, and reported figures put a crew of 12 somewhere around $36,000 – $58,000 a year, before an implementation fee that can run five figures. Against that a $52,000 – $78,000 build pays back in roughly 17 months, and that’s still before anybody maintains it.

Buy it when

Your dispatch problem is a normal dispatch problem. Schedule the job, send whoever’s free and closest, invoice what got done. That’s what these products were built for and they do it well.

Build when

The thing that makes you money doesn’t fit. Dispatch that has to match certifications instead of availability, because only two of your guys can pull a permit for that work. Backflow test certificates that have to file themselves. A rooftop unit’s service history that follows the equipment and not the customer. Panel schedules nobody else models.

Either way, test the signal problem first

All three publish an offline feature and the depth varies a lot. The only way to see yours is to put a trial on a tech’s phone and send him to the worst job site you’ve got. One afternoon tells you more than the feature list will.

The connection itself is about $2,000 of this build, one of the smaller pieces on it. It’s also the thing most likely to blow your schedule, and what you’re running decides which one you get.

If it’s a modern cloud product with a documented way for other software to talk to it, it’s cheap. If it’s a copy running on a machine in the back office, with no safe way to test against it and one guy left who remembers how it was set up, it isn’t. Same feature either way, and you pay the difference in weeks.

You can settle this before anybody quotes you. Which product do you run, which version, is it in the cloud or on a computer in your office, and has anyone ever connected anything to it? Your bookkeeper can answer all four in one phone call.

If nobody knows, say so. It gets priced as a risk instead of a certainty, and that costs less than finding out in month three.

The job record and the photos. That’s $40,000 – $61,000, about 22% off the full build. Those two settle disputes and get invoices out faster, so they start paying you back while everything else is still being built.

• 01

  The office portal: one simple internal screen instead of the full thing

• 02

  Push notifications: add them once people are using the app every day

• 03

  Work order documents: one fixed layout instead of one you can change

• 04

  The third kind of user: launch with techs and dispatch, add owner views after

It takes 22% off the price and roughly a week off the calendar, 14–21 weeks against 15–23 weeks. The foundation underneath takes as long either way. Speed isn’t the reason to do it. The reason is that you find out what your techs actually do with the thing before you pay to build the office half around a guess.

Running dispatch off the spreadsheet you already use for another two months is annoying and completely survivable. It also means you build the office tool once, after you know what dispatch needs, instead of twice.

You can measure the payback on this, which you can’t say for most software. Take three days out of your invoicing cycle, stop two disputes a month, and inside a quarter you know whether it paid. Then you’re spending the rest of the money on something you’ve watched work.