Choosing Field Service Software: Categories, Costs and Rollout

Four demos, four impressive salespeople, a spreadsheet of checkmarks, and eight months later the office is on the software while the technicians are still on paper. That outcome is almost never caused by picking the wrong product — it is caused by evaluating the wrong thing. Feature lists compare software to software. What decides the result is how a system handles the five specific, messy kinds of work your company does every week, what it really costs once the processing rate is included, and how the first thirty days of rollout are run.

Why most software decisions go wrong

The typical evaluation goes like this. The owner watches four demos, each one impressive. He makes a spreadsheet comparing features. He picks the one with the most checkmarks and the friendliest salesperson. Eight months later the office is running the software and the technicians are still using paper, and the company is paying $700 a month to do what it did before, slightly slower.

That failure is not caused by picking the wrong product. It is caused by evaluating the wrong thing. Feature lists compare software to software. What matters is how the software handles your jobs — the specific, messy, five or six kinds of work your company actually does every week.

Two more traps sit alongside it. The first is buying for the company you plan to be in three years, which means paying today for complexity nobody will use until then and losing the team in the meantime. The second is treating the subscription price as the cost, when the payment processing rate buried in the contract usually moves more money than the subscription ever will.

The four jobs a field service system has to do

Strip away the marketing and every platform in this category is competing on the same four jobs. If a system does these four well, the rest is preference.

1. Get the right person to the right address with the right information. Scheduling and dispatch, on a map, with the job history visible on the technician's phone before he knocks. This is the job that saves the most hours and the one most commonly demoed badly.

2. Turn work into money without a gap. Estimate, approval, work order, invoice, payment — in one chain, with no re-typing. Every place where a human copies information from one place to another is a place where money leaks.

3. Remember the customer. Equipment installed, model and serial, what was done last time, what was declined, what is under warranty. This is the difference between a company that sells to its own customer base and one that starts from zero every time.

4. Tell you what happened. Revenue by job type, close rate by technician, average ticket, unbilled work, aging receivables. Not dashboards — the four or five numbers you would actually change a decision over.

The five-job test. Before any demo, write down your five most common jobs — a diagnostic call, a scheduled maintenance visit, a multi-day install, a warranty callback, a recurring contract. In every trial, run all five end to end yourself, from booking to paid invoice. A platform that handles four of them beautifully and turns the fifth into a workaround is the platform that will be abandoned.

Keep reading — free

The rest of this article is worth the 20 seconds

Tell us where to send it and the full piece unlocks right here, along with everything else on this blog.

Rather just talk? Message us on WhatsApp.

When you are ready — and when a spreadsheet is still right

Software is not a milestone. It is a response to a specific kind of pain, and buying it before the pain exists creates a different pain.

You are probably still fine on spreadsheets and a shared calendar if you have one or two people in the field, you invoice from a phone the same day, and nothing has ever fallen through a crack. That is a real business configuration, not a stage to be ashamed of.

Four signals say the spreadsheet has ended:

Work gets done and not billed. The moment you cannot say with certainty that every job last month was invoiced, the paper system has already cost more than the software.
Someone in the office spends more than an hour a day re-typing. That is a salary being spent on data entry.
Estimates go out and nobody follows up. Unfollowed estimates are the single largest pool of free money in most service businesses.
You cannot answer "what did we do for this customer last year" in under a minute.

The three categories, and who each one fits

Product names and prices change constantly, so evaluate categories first and specific vendors second.

All-in-one field service platforms — the group that includes names like Jobber, Housecall Pro and Workiz — bundle scheduling, dispatch, quoting, invoicing and payments in one subscription. They fit companies from roughly two to twenty-five field staff, and their strength is that everything already talks to everything else. Their limit is depth: highly specific trade workflows sometimes need a workaround.

Trade-specific enterprise platforms — ServiceTitan and its peers — go far deeper into pricing books, call center scripting, membership programs and financial reporting. They fit larger operations, typically fifteen field staff and up, and they charge and onboard accordingly. Below that size, most companies buy capability they will not staff.

Assembled stacks — a general CRM plus a scheduling tool plus accounting, connected by integrations. Cheaper on paper and genuinely right for unusual workflows, but you become the integrator. Every update on either side is now your problem.

The honest advice for most companies with three to fifteen technicians: start in the middle category, and only move up when you have a specific, named limitation you have already hit — not because a demo showed you something impressive.

What it actually costs

The subscription is the smallest line. Four costs are larger and less visible.

Per-user pricing that grows with you. At $89 per user per month, eight seats is $8,544 a year. Adding two techs adds $2,136. Check whether office staff, dispatchers and read-only users each consume a seat — the answer varies wildly and changes the total.

Payment processing. This is the one to read carefully. On $1.42 million of card volume, the difference between 2.6% and 2.9% is $4,260 a year — half the subscription. Some platforms discount the software and make it back on the rate.

Onboarding and data migration. Budget $1,500 to $5,000 in fees, and more importantly 40 to 80 hours of someone internal. This is not optional work and it cannot be delegated to the vendor, because only your people know which of your customer records are real.

The productivity dip. Plan for four to eight weeks of running slower. Companies that do not plan for it conclude in week three that the software is bad and revert.

The evaluation that takes one afternoon

Skip the feature spreadsheet. Do this instead, with two or three finalists.

Run your five jobs end to end in the trial. Book, dispatch, complete, invoice, take payment. Yourself, not the salesperson.
Put it in a technician's hands for one real day. Your least tech-comfortable technician, not your best. If he can complete a job on it without calling the office, adoption will work.
Test the exit before the entrance. Ask exactly how you export your customer list, job history and invoices if you leave, and in what format. A vague answer is an answer.
Call support twice during the trial — once with an easy question, once with a hard one — and time both responses. Support quality during a sales trial is the best it will ever be.
Ask for two references at your size in your trade, and ask them the only question that matters: what does it still not do well?

One year, measured

A six-technician mechanical company doing $1.42 million a year ran on paper work orders, a shared calendar and accounting software. Before the change, three numbers were measured deliberately for a quarter.

Work completed but never invoiced ran at 3.2% of revenue — about $45,440 a year. Invoices went out an average of nine days after the job, with days-sales-outstanding at 41. And of 214 estimates issued in the year, 81 were never followed up even once; the close rate sat at 41% on an average estimate of $2,850.

Twelve months after implementation, with invoicing done on site and automated estimate follow-up running:

Unbilled work fell to 0.4% — $39,760 recovered.
Invoices went out same-day and DSO fell from 41 to 23 days, freeing roughly $70,000 of working capital — money that had been financing customers.
Estimate close rate rose from 41% to 47%, worth about $36,594 in additional revenue on the same estimate volume.

Total cost for the year: $8,544 in subscription, $2,500 in onboarding, and roughly 60 internal hours. Against $76,354 in recovered and added revenue plus the cash freed, it paid for itself in the first two months — and the largest single item was the most boring one: billing work that had already been done.

The migration nobody plans

Data migration is where implementations die, and it dies for a predictable reason: companies try to bring everything.

Bring three things and nothing else. Active customers with correct address, phone and email — not every name from the last twelve years. Open balances, reconciled to the penny before the first day. And equipment records for customers under service agreements, which is the data you cannot reconstruct later.

Everything else — old invoices, closed jobs, dead leads — stays in the old system as an archive. It is available if you ever need it and it is not slowing down your launch.

One rule saves weeks: clean the data before it moves, not after. Duplicates and bad phone numbers are ten times harder to fix once they are inside the new system and attached to records.

Getting technicians to actually use it

Adoption is not a training problem; it is a design-of-the-rollout problem. Four things decide it.

Pick one champion in the field — the technician other technicians ask questions of — and involve him during the trial, not after the purchase. Turn off the old system on a date, because running both in parallel guarantees that neither is trusted. Cut what you require in the app to the minimum for the first month: arrive, complete, invoice. Photos, checklists and custom fields get added in month two. And fix the friction they report in the first two weeks immediately, even the small stuff, because the first two weeks decide whether the crew believes this is permanent.

Five mistakes in software selection

Buying for the company you will be in three years. You pay now and lose the team to complexity today.
Comparing subscription prices and ignoring the processing rate. The rate often moves more money.
Letting the office choose alone. The office is not the user who can abandon it.
Migrating twelve years of data. It delays launch by weeks and nobody ever looks at it.
Switching again within eighteen months. The second migration costs more than the first, and the crew stops believing in any system.

The numbers to watch after you switch

Four numbers tell you whether the system is earning its keep, and all four should be checked at 30, 90 and 365 days. Percentage of jobs invoiced within 24 hours, which is the fastest-moving indicator and the one that pays for the software. Unbilled work as a percentage of revenue, which should approach zero. Estimate follow-up rate, where the target is 100% and the gap is pure lost revenue. And days sales outstanding, which converts directly into cash you no longer have to borrow. If none of these has moved by day 90, the problem is the rollout, not the product — and switching again will not fix it.

Find out what your paper system is costing you

Send us your annual revenue, your technician count and how long invoices take to go out. We will run the unbilled work, DSO and estimate follow-up math with you, and show you whether software is your bottleneck or something else is.

Talk on WhatsApp See how it works

Frequently asked questions

How do I choose field service software?

Evaluate against your own work rather than against feature lists. Before any demo, write down your five most common jobs — a diagnostic call, a scheduled maintenance visit, a multi-day install, a warranty callback and a recurring contract — then run all five end to end yourself in each trial, from booking to paid invoice. A platform that handles four beautifully and turns the fifth into a workaround is the one that gets abandoned. Four more steps decide the rest. Put the trial in the hands of your least tech-comfortable technician for one real day, because if he can finish a job without calling the office, adoption will work. Ask exactly how you export customers, job history and invoices if you leave, and in what format. Call support twice during the trial and time both answers, since support is at its best during a sales process. And ask two references at your size in your trade what it still does not do well.

What does field service software actually cost?

The subscription is the smallest of four costs. Per-user pricing compounds: at $89 per user per month, eight seats is $8,544 a year and two more technicians add $2,136, so check whether office staff, dispatchers and read-only users each consume a seat. Payment processing is the line to read carefully — on $1.42 million of card volume, the difference between 2.6% and 2.9% is $4,260 a year, roughly half the subscription, and some platforms discount the software and recover it on the rate. Onboarding and data migration run $1,500 to $5,000 in fees plus 40 to 80 hours of internal time that cannot be delegated, because only your people know which customer records are real. And there is a productivity dip of four to eight weeks; companies that do not plan for it decide in week three that the software is bad and revert.

When is a service business too small for field service software?

Software is a response to specific pain, not a milestone. With one or two people in the field, same-day invoicing from a phone and nothing falling through the cracks, a spreadsheet and a shared calendar are a legitimate configuration rather than a stage to be embarrassed about. Four signals say that has ended. Work gets completed and never billed — the moment you cannot say with certainty that every job last month was invoiced, the paper system already costs more than the software. Someone in the office spends more than an hour a day re-typing information, which is salary spent on data entry. Estimates go out and nobody follows up, which is usually the largest pool of free money in the business. And you cannot answer what you did for a given customer last year in under a minute.

What is the difference between Jobber, Housecall Pro and ServiceTitan?

They sit in different categories, and category matters more than any individual feature because product names and prices change constantly. All-in-one field service platforms, the group that includes Jobber, Housecall Pro and Workiz, bundle scheduling, dispatch, quoting, invoicing and payments in a single subscription; they fit roughly two to twenty-five field staff and their strength is that everything already talks to everything else, with depth as their limit. Trade-specific enterprise platforms such as ServiceTitan go much deeper into price books, call center scripting, membership programs and financial reporting, fitting operations of about fifteen field staff and up, with pricing and onboarding to match — below that size most companies buy capability they will not staff. The third option is an assembled stack of CRM plus scheduling plus accounting, cheaper on paper and right for unusual workflows, at the cost of becoming your own integrator.

How do I get technicians to actually use new software?

Treat it as a rollout design problem rather than a training problem, and four decisions carry most of the result. Pick one champion in the field — the technician the others already ask questions of — and involve him during the trial rather than after the purchase. Turn the old system off on a set date, because running both in parallel guarantees neither is trusted. Require the minimum in the app for the first month, meaning arrive, complete and invoice, with photos, checklists and custom fields added in month two. And fix every piece of friction reported in the first two weeks immediately, including the small ones, because those two weeks decide whether the crew believes the change is permanent. Data migration deserves the same restraint: bring active customers, open balances and equipment records under service agreements, and leave everything else in the old system as an archive.