What Custom Field Service Software Actually Costs in 2026

Custom field service software costs $5,000–$50,000 for a scoped, block-based build in 2026. Dev-shop lump-sum quotes for the same work commonly hit $75,000+. Per-seat SaaS looks cheaper — until you multiply seats by years: one 14-tech shop was quoted $57,000 per year, forever. Under ~10 seats, SaaS wins. Past that, owning usually pays back in two to three years.

Those aren’t ranges we pulled from a pricing survey. They’re three real quotes one real buyer collected while shopping — and this post walks through his math, because it’s the same math you’re doing right now.

One Buyer, Three Prices

The buyer: a commercial facility-maintenance vendor in the Northeast doing about 450 work orders a month at a $400 average ticket — roughly $1.8M a year with 10 people on payroll. Work orders come in through facility-management platforms; CompanyCam handles photos and work orders on the front end (here’s what CompanyCam costs and where it stops); the KPIs he actually runs the business on are tracked by hand.

He went shopping for software that could measure tech performance. Three prices came back:

Custom-quote figures from a July 2026 evaluation; last verified 2026-07-25.

The SaaS quote died on a whiteboard:

“She gives us a price that’s $57,000 a year… That’s only with 14 techs. What happens when we go to 50 techs?”

He’s targeting $30–50M through acquisitions. At 50 techs, that per-seat contract extrapolates past $200,000 a year — for software he still wouldn’t own. And on today’s revenue, $57K is 3.2% of gross coming straight off the top:

“If I implement that today, I wouldn’t make one nickel today.”

The $75K lump-sum quote is the more interesting failure. He didn’t reject the price. He rejected writing one big check to a shop he didn’t trust yet, for software he hadn’t seen run. Which brings us to the actual build-vs-buy question.

What Per-Seat SaaS Really Costs at 10–15 Techs

Before the crossover math, anchor the “buy” side with real numbers. These are annual-billing rates:

Jobber and Housecall Pro pricing last verified 2026-07-25; ServiceTitan third-party estimates last verified 2026-05-13. Full breakdowns: Jobber pricing · Housecall Pro pricing.

Notice the spread. General-purpose platforms like Jobber and Housecall Pro are genuinely cheap per seat — that matters for the honest half of this argument below. It’s the vertical, quote-only, per-tech platforms where the number quietly climbs to ServiceTitan territory: that $57K quote works out to roughly $340 per tech per month, right in line with ServiceTitan’s estimated mid-tier.

For the wider landscape — who owns what, what each platform does well — see the complete guide to home-services software.

The Crossover Math: When SaaS Wins, When Owning Wins

We build custom software for a living, so you’d expect us to say “always build.” We won’t, because it’s not true.

SaaS wins when:

• You’re under ~10 seats. Jobber Grow’s 10-user config is $210/mo — about $2,520 a year. No custom build on earth pays back against that.
• Your workflows are standard. Schedule, quote, invoice, collect. These platforms have poured a decade of development into those loops. Rebuilding them custom is buying a hand-made hammer.
• You need it running next week. SaaS onboards in days. A custom build ships its first block in weeks.

Owning wins when:

• Seats × rate × years beats the build cost. The 14-tech shop above: $57K/yr rented versus a build in the $30K–$50K range. The build is break-even inside year one and roughly $120K ahead by year three — before any headcount growth.
• You measure things no platform measures. Tech performance KPIs, per-region leaderboards, role-based dashboards for the owner, dispatchers, and techs. This buyer looked at every out-of-the-box option and found the same gap: plenty of features, not his fields.
• Growth multiplies the rent. Per-seat pricing means every hire raises your software bill. If your plan is 14 techs today and 50 through acquisitions, you’re not comparing $57K to a build cost — you’re comparing $200K+/yr to it.
• The seat contract outlives the seats. More on that next.

The honest summary: the crossover sits around 10 seats and one non-standard workflow. Below both thresholds, buy. Above either, run the math — it’s closer than the SaaS pricing page suggests.

The Honest Fine Print on Owning

Custom software isn’t free after launch, and anyone who quotes it that way is hiding the ball. Budget for three things:

• Hosting and infrastructure. Real, but small — for a system this size it’s a rounding error next to any per-seat bill above.
• Fixes. Software breaks. The question isn’t whether, it’s what happens when it does. This buyer had lived the SaaS version: “I know that softwares aren’t perfect. But I need to be able to get it fixed, worked out right away… When I’m paying these companies $3,500 a month, I would call and they got corporate.” With an owned system and the shop that built it, the person who wrote the code is the person who picks up.
• Changes. Your business will outgrow version one — new client platform, new territory, new report. Those are new blocks, priced as blocks. The difference from SaaS is you buy changes when you need them, not features a roadmap committee shipped for someone else’s business.

None of that erases the crossover math. It moves the break-even by months, not years — and it’s the part a lump-sum shop won’t bring up until after you’ve signed.

The Risk Nobody Puts in the Spreadsheet

Our buyer had run a 34-employee mortgage company across five states, and his software scar tissue is why he shops the way he does.

His platform bill there had crept to $3,500 a month — on three- and five-year per-seat contracts. Then the market turned and he cut staff 20–40%:

“I’m still tied into these user seats, and I wanted to puke.”

That’s the part of per-seat SaaS the demo never covers: the contract is sized to your best year and enforced through your worst one.

The second risk is who owns your vendor next year. Private-equity roll-ups are working through field-service software the way they worked through mortgage tech:

“As soon as they were bought out by the equity firms — it happened to me three times — I was lost, done, forget it, history.”

Post-buyout, support gets tiered, roadmaps refocus on enterprise accounts, and prices step up at renewal. You can’t control it and you can’t price it — except by owning the software instead of renting it. His four-word version of the whole argument: “I’d rather it be mine.”

Software risk isn’t the only invisible line item in this industry, either — install businesses routinely eat a $50K-a-year scheduler problem without ever seeing it on a P&L.

Why the $75K Lump-Sum Model Fails (Even at a Fair Price)

Here’s the thing about that $75K quote: for a multi-workflow custom system, it isn’t outrageous. Dev-shop content mills will tell you custom CRM builds run anywhere from $20K to $200K, which is true and useless. The real problem is the shape of the deal, and the buyer named it exactly:

“I don’t want to be signing a contract for 60 grand and then go in that route. What I’d like to be able to do is take one or two hours on three parts of the business… and you say, here’s what it would cost to do that, here’s my rough estimate on how many hours.”

A lump-sum contract asks you to bet the whole number on a vendor you haven’t worked with, against a spec you can’t fully write yet. His alternative — the one that kept the conversation alive — is block-based:

1. Scope three workflows. The two or three parts of the business that hurt most — say, work-order intake, tech KPI tracking, and an owner dashboard.
2. Estimate hours per block. Each block gets its own number, in the open.
3. Build one block at a time. You pay for a block, it ships, it runs in production, and then you decide whether the next one gets built.
4. Under-budget hours roll forward. As he put it: “You might say, we’re a little bit under budget. We’ll use that time for the next build-out.”

No single check big enough to hurt. No paying for workflow three before workflow one has proven itself.

One more thing that call taught us: this audience doesn’t speak developer. When “MVP” came up, his response was “what’s the acronym for that?” What landed was a car analogy, so that’s the one we use — the first version is deliberately minimal because the engine runs before the turbo goes in. Your work-order tracker doesn’t need the AI data-extraction layer on day one. It needs to start, run, and get you home. Turbo comes later, funded by a system that’s already working.

This is how our own custom-build engagements are priced: $5,000–$50,000 depending on scope, block by block, with a 10% buffer built into every quote so the estimate you approve is the invoice you get. A three-workflow system of the size discussed above lands around four months end to end — with the first block live well before that.

How to Get a Real Number for Your Business

Generic cost ranges — including ours — only narrow to a real quote after scoping. The fastest scoping method we’ve found came out of this same deal:

Record your workflow instead of describing it. Open a screen recorder (Loom, or whatever you have), spend 30 minutes doing the work the way you actually do it — pulling work orders off the client portal, hand-entering them, chasing the numbers your office team writes down by hand — and narrate what you wish happened instead.

A recording beats a discovery questionnaire every time. It shows the fields you actually touch, the double entry you actually do, and the reports you actually wait a week for. The buyer’s reaction when we proposed it: “I can break down what we want individually, and it’s very clear.”

Then the estimate writes itself: workflows → blocks → hours per block → a number you can veto one block at a time.

If you’re at the stage where you’re comparing a per-seat quote against owning the thing, book a discovery call — bring your current software bill and the workflow that annoys you most. We build these systems for a living, and if your math says SaaS wins, we’ll tell you that too. It’s a shorter call, and it costs us nothing to be right.