Is ServiceTitan Worth It for a Small HVAC Company?

The honest answer is that it depends almost nothing on your revenue and almost everything on your shape. A full field service platform is a dispatch and operations system built for a company that already has people in an office doing office jobs. If you have a dispatcher, a pricebook your techs quote from, a real book of maintenance agreements, and somebody whose job it would be to own the setup, it usually pays. If you are three trucks, you answer the phone yourself between jobs, and there is no spare hour in your week, you will be buying capability you cannot physically reach, and you will feel that every month when the invoice clears.

That is the whole answer. The rest of this page is how to tell which one you are without booking a demo to find out, because a demo is a sales conversation and you should arrive already knowing your own numbers.

There is a six-question tap test below that sorts you in forty seconds, a slider that turns whatever you pay for software into a cost per completed job, and seven questions to send any rep before you get on a call.

What the thing actually is

ServiceTitan is field service management software for the trades: HVAC, plumbing, electrical, roofing. Scheduling and dispatch on a board, a customer record with equipment history attached, a pricebook the tech presents from in the home, invoicing and payment, marketing attribution that ties a booked call back to its source, owner reporting, and a technician app in the truck. It is a real product, it is well built, and shops that fit it will tell you it changed their business. It is also now a public company, which matters less for the software than for where it points: public software companies grow by expanding inside bigger accounts, and roadmaps follow the money.

Two structural facts are worth having before anything else. It does not publish pricing: you get a quote after a demo, shaped to your company, and what you are quoted is not what your neighbor is quoted. And it is an implementation, not a download. Neither of those is a criticism. They are the normal shape of software sold to companies with operations teams, and they become a problem only when the buyer is not that company, which is most of the people typing this question into a search bar at ten at night.

The number that explains the mismatch

Here is the structural thing nobody selling software will open with. According to the U.S. Census Bureau's Statistics of U.S. Businesses, as summarized by the SBA Office of Advocacy, firms with fewer than twenty employees make up about 89 percent of all employer firms in the United States. The trades are not an exception to that. They are one of the places it is most true.

So the market has a shape problem that has nothing to do with any one vendor. The flagship products are built for the visible top slice: the hundred-truck shop, the roll-up, the company with a call center. The overwhelming majority of actual contractors are a handful of people who know each other's kids. A tool designed for the first group is not a bad tool. It is a correct tool pointed at somebody else, and the reason it stings is that the marketing does not say so.

Which means the real question is not "is this good software." It is: am I the company this was built for yet. That one you can answer yourself, today, for free.

The five jobs you are actually trying to hire software for

Nobody wakes up wanting a platform. They wake up with one of five pains, and the pain decides whether a big platform is the answer or an expensive detour.

1. Dispatch chaos. Two techs sent to the same call, a customer standing in a driveway at four o'clock. You are trying to buy a schedule everyone can see.
2. Price inconsistency. The same job goes out at three prices depending on who quoted it, and the cheap one was somebody's memory. You are trying to buy a pricebook.
3. Money arriving slowly. Invoices go out Sunday night and half of them chase for three weeks. You are trying to buy payment speed.
4. Calls falling on the floor. The phone rings while everybody is under a house. You are trying to buy an answered phone, which is often a person and not software.
5. Not knowing what anything made. Busy year, thin bank account. You are trying to buy the truth about margin per job, a different animal from all four above.

Only the first two are really platform-shaped. Payment speed can be solved with far less. An answered phone is a hiring decision. And margin per job, the one that decides whether you survive a bad quarter, is regularly the last thing a shop configures and the first thing it needed, a trap we wrote up in Why Am I Booked Out but Not Making More Money?

Before you shop, name your one. Not three. The shop that walks into a demo with one named pain and its monthly cost buys well. The shop that walks in with a general feeling of chaos buys a platform and still has the feeling.

Tap test: are you the company this was built for?

Six questions, no typing, forty seconds. It never asks your revenue, because revenue is the least useful number here. It asks about shape.

Nothing there is sent anywhere, and the scoring is not a secret: every answer that pushes you toward a platform is one about somebody in an office doing an office job. That is the threshold.

The honest comparison

Four ways to run a shop, compared on the things that actually decide it. Vendors change their packaging, so confirm the details yourself before you sign. The structural columns are the ones that hold.

One more option belongs on that list even though it does not fit a table: change nothing about your software and hire a part-time office person instead. For a lot of five-truck shops that is the higher-return purchase, and nobody markets it to you, because no salesperson gets paid when you hire a human.

Do not buy a full platform if

Do not buy it if you have fewer than about four techs and no office staff, because there is nobody to run the system and it becomes a second job for the owner. Do not buy it if you cannot name, right now, the one problem you are solving and roughly what it costs you a month, because you will end up evaluating features instead of outcomes and features always demo well. Do not buy it if the implementation would land in your busiest season, which in most of the country means do not go live in July. Do not buy it if your crew still argues about whether jobs were assigned, because a platform will faithfully record that argument in a nicer interface. And do not buy it because a competitor across town has it. You cannot see their bank account, whether they use a quarter of it, or what the renewal did to them.

Do buy it if

Buy it if somebody in your office spends a real part of the day just moving techs around, and that person could be doing something worth more. Buy it if your price varies by who is standing in the kitchen. Buy it if maintenance agreements are a serious line of business and you are tracking them somewhere that could be lost. Buy it if you can name the person who owns the setup, and that person is either not you or is you with the hours genuinely blocked. And buy it if the number from the calculator below is small next to what the pain costs you.

The monthly number is not the cost

Owners compare software the way they compare tool prices, which is the wrong instinct here. Drag your own numbers below. The useful outputs are not the annual total: they are the cost per completed job, and the extra jobs a month it has to produce to be free.

Drag those honestly and the cost per job usually turns out smaller than the monthly bill made it feel, and the break-even in extra jobs usually turns out small too. The number that should worry you is not the price. It is a break-even you cannot clearly picture yourself hitting, because that is the arithmetic of a tool you already suspect you will not use.

Ninety days on a big platform, told honestly

This is the part that decides outcomes and the part nobody schedules. Here is the shape of a real implementation in a shop with eight techs and one office manager, written as the calendar it actually is.

Weeks one and two: the pricebook

Everything waits on this. Your prices live in three places right now: a spreadsheet, the senior tech's head, and a decade of habit. Somebody has to turn that into a structured list with tasks, materials and labor. It is not hard work, it is long work, and it is the most underestimated part of the project. A platform with an empty pricebook is a calendar with an invoice attached. Decide who owns this and protect their hours in writing. If the answer is "the office manager, in between everything else," you have scheduled the failure.

Weeks three and four: customers and equipment

Importing customers is mechanical. Attaching equipment history is the part with judgment in it, and it is where most of the value lives, because a tech who can see that this same condenser was serviced fourteen months ago walks in already knowing something. Expect your existing records to be worse than you thought. Everybody's are.

Week five: the office goes live, the trucks do not

Run dispatch in the new system while the techs stay on whatever they know. It feels slow and redundant, and it is the best decision in this timeline, because it finds the configuration mistakes with one person's workflow instead of nine.

Weeks six and seven: the trucks

Now the techs. Expect a real dip in output and expect a senior guy to hate it, usually about the interface rather than the idea. Two rules save this phase: train in small groups, never on a Monday, and pick the one tech who is naturally curious and make him the person everyone asks. A crew will ask a peer nine times before asking the owner once.

Weeks eight through twelve: the part everyone skips

Reporting. Getting jobs into the system is not the win. The win is the first month you can see which job types actually make money, which techs close, where calls come from, and what an agreement is worth over a year. Shops that stop at week seven have bought a very expensive schedule, and that is the most common bad ending: nothing forces you to finish, so the last twenty percent quietly does not happen.

None of this is a reason not to do it. It is the actual price, it is payable, and it is just not on the quote.

Send this to the rep before you book anything

Send these in an email before a call and the whole conversation changes, because you are asking for numbers instead of receiving a tour. Any vendor worth buying from answers them without flinching.

Question five predicts the outcome. A vendor who answers it specifically is describing a real implementation. One who waves it away has told you the hours will land on you anyway, unplanned.

Six mistakes contractors make buying this

1. Buying the biggest thing you can afford

Software is not a truck. A bigger one does not carry more. Capability you never reach is not saved for later, it is paid for monthly and forgotten. Buy for the company you are this year, and change tools when you outgrow them.

2. Starting the rollout in peak season

In July an HVAC shop has no spare attention, and a half-learned system during the busiest weeks teaches your crew that the new thing makes their day worse. That lesson does not wear off in the fall. Go live in a shoulder season.

3. Not naming the owner of the project

"We are all going to use it" is how nobody configures it. One name, hours actually protected. It is the same rule that governs work in a crew chat, which we laid out in How to Assign Jobs to Your Crew in a Group Chat Without Things Falling Through. A job owned by everyone is owned by no one, and an implementation is just a very large job.

4. Keeping the old system running beside the new one

Parallel running is correct for two weeks and fatal at two months. If the whiteboard still has the real schedule on it in week nine, the whiteboard is your system and the platform is a bill. Pick a date and let the old thing die.

5. Comparing monthly prices instead of total first year

The monthly number is the one designed to be compared. Setup, training, migration and processing rates are where the real difference lives. Ask for one number: the total first year.

6. Solving a people problem with a purchase

If your techs do not communicate, if nobody wants to own dispatch, if the office and the field do not trust each other, software makes that visible and does not make it better. Make sure you are buying a tool problem. You will know, because you can describe a good week and point at the exact step that breaks.

The smaller version that works

If the tap test put you in the "not yet" band, here is what to do instead. None of it is a stopgap: plenty of very profitable shops run this way permanently.

• One schedule everyone can see on a phone. A shared calendar is fine. The requirement is not features, it is that nobody has to ask where they are going tomorrow.
• One name and one date on every job. The most common cause of a missed appointment is a job addressed to everybody. A name at the front fixes more than any purchase will.
• A pricebook outside one person's head. Even a shared sheet on a phone. The point is that two techs quoting the same job land within a few dollars of each other.
• Invoice from the driveway, not the couch on Sunday. If the invoice leaves before the truck does, your cash cycle shortens by a week with no new software at all.
• One number a week. Pick margin per job and look at it every Friday. It is the number that would have warned you a year early.

That list is free. Do all five for one quarter, and if you can still feel the pain, you now have a very well-defined problem and you are ready to spend money on it. That is the right order, and almost nobody uses it.

Where we fit, said plainly

We are Class Company US. We started in New Orleans as a web design studio, grew through Houston and across Texas, and we work with brands nationwide. We also built our own AI agent, Tarvis, and we are not going to pretend it is a dispatch platform.

Two honest paths. If your real problem is that customers cannot find you, cannot reach you, or bounce off your site before they call, that is our actual trade and the consultation is free. Plenty of contractors are losing more at the front door than anywhere in operations, which is covered in My Website Gets Visitors but No Phone Calls. What Is Wrong?

If your problem is that the crew lives in a group chat and nothing is tracked, Tarvis AI is a chat where the AI sits in the thread with the crew. A job assigned in conversation becomes a real task with an owner and a date, and assigning it is also how that person gets invited. Projects carry dates and countdowns, money is tracked per project so you can see margin, and the weekly status report gets drafted from what actually happened. Two free teammates come with every brand, it is free while in beta, and it installs from the browser to your home screen.

And the limits, since this page is not a brochure. Tarvis does not dispatch, does not hold a pricebook, does not run payroll, and does not do call attribution. If those are the words in your search history, a field service platform is the right category, and you should buy the one that fits your shape.

Frequently asked questions

Is ServiceTitan worth it for a small HVAC company?

It depends far less on your revenue than on your shape. If you have somebody whose actual job is dispatching, a pricebook the whole crew quotes from, a real book of maintenance agreements, and a person in the office who would own the setup, a full field service platform usually earns its keep. If you are three trucks, you answer the phone yourself, and nobody has spare hours for an implementation, you will be paying for capability you cannot reach yet.

How much does ServiceTitan cost for a small company?

ServiceTitan does not publish its pricing, so the only honest answer is that you have to get a quote. Ask for the total first year in writing, including setup, training, payment processing rates and anything billed per user, and ask in the same breath what the price does at renewal.

What is the cheapest thing that actually fixes dispatch chaos?

A schedule everybody can see on a phone, one named owner per job, and a real date on every job. That costs nothing and it removes the most common cause of a missed appointment, which is a job that belonged to everybody and therefore to nobody.

Will switching field software make my techs faster?

Not on its own. Software changes where information lives, not how long a coil takes. The gains that are real come from fewer repeat trips, faster invoicing, and fewer jobs quoted wrong, and every one of those depends on the crew actually using the thing.

How long does a field software implementation take?

Ask the vendor for their own median, in weeks, for a company your size, and ask how many hours of your people it assumes. The failure mode is rarely that setup is long. It is that nobody was given the hours, so the platform goes live half configured and the crew learns it as a worse version of what they already had.

Can I run an HVAC company on a group chat and a spreadsheet?

Up to a point, yes, and plenty of profitable shops do. The limit shows up when nobody can say what happened last Thursday without scrolling, or what a job made without rebuilding it from memory. That is the moment to add something, and not before.

What should I do before I book a demo?

Write down the one problem you want gone and what it costs you in a month. Walk into the demo holding that number. A demo with no number on it turns into a tour of features, and a tour of features sells every time.

Test yourself

Five questions on what is above. We email the answer key with a short explanation for each, written for owners.

Help us make this better

This is written for people who run trucks, and the best corrections come from people who run trucks. What was missing? Thirty seconds, and we read every one: