HVAC Software for 2 to 5 Techs: Getting the Owner Out of the Truck

At one truck the constraint is your hands — you cannot be on a call and on the phone at the same time. At two to five trucks the constraint moves to your attention, and it is much harder to see because it does not look like a shortage. It looks like a busy owner, which every owner assumes is normal.

Put a number on it. Every truck in the field generates somewhere around 30 to 60 minutes a day of office work: dispatching the day, answering the customer who wants a window, sourcing a part, following up a quote, chasing an invoice, asking for a review that nobody ends up asking for. At four techs that is two to four hours a day landing on somebody. At this size that somebody is you, or your spouse, and it lands after the workday rather than during it.

The clearest tell that you are in this window is a set of quotes sitting in a folder that nobody has chased. At two to five techs, unfollowed quotes are the most reliable pile of found money in the building — the work is already sold at least halfway, the customer already met you, and the only thing standing between you and the revenue is a phone call that nobody has time to make. The second tell is that you cannot say, without looking, what your close rate was last month.

Be clear about what is and is not a software problem here. Software does not replace a person at this size and it does not supply discipline you do not have. What it does is make it possible for one non-technical office person — often part-time, often related to you — to hold four trucks without dropping anything. That is the actual buying criterion at this stage, and it is not a feature list. If the person running your office cannot use it confidently in a week, it is the wrong product no matter how it demos.

Every owner at this size asks some version of "when can I stop running calls?" and almost everybody answers it by feel, usually during a bad week. Run it as two numbers instead.

The first number is what you produce in the field. A residential service technician in a decent shop generates somewhere between $180,000 and $250,000 of revenue a year, with top-quartile producers pushing $250,000 to $400,000 and up. Say you personally produce $200,000 at a 45% gross margin. That is roughly $90,000 of gross profit a year that walks out the door the day you stop running calls.

The second number is what you would produce off the truck, and this is where owners undersell themselves. There are three jobs at this size that only get done if somebody is not holding a wrench: chasing quotes that were never closed, selling maintenance agreements, and making sure the phone converts. Take quote follow-up alone. A shop running 60 replacement quotes a year at a $9,000 average ticket and closing 30% closes 18 of them. Systematic follow-up — a call at 48 hours, a call at one week, a call at 30 days — routinely moves that to 40%. That is six more sales, $54,000 of revenue, and roughly $24,000 of gross profit from making phone calls. Add memberships and booking rate and the office side typically clears what you were producing in the field.

The catch, and it is a real one: that only holds if you actually do those three things. The common failure is an owner who comes off the truck and spends the recovered hours on the same coordination that was already eating his evenings, just now during the day. Coming off the truck without a system underneath it converts a $90,000 producer into an expensive dispatcher.

Then price the replacement. A competent service technician runs $28 to $38 an hour, and burdened with payroll taxes, insurance, a truck, fuel, a phone, and benefits, that is roughly $75,000 to $110,000 a year all in — plus 90 days where he is slower than you are. Which is why the honest sequence at this size is partial rather than binary: cut yourself to three field days, then two, then one, and build the office system in the gaps. Buy the software that makes the office side possible before you hire the person, not after. The shops that do it the other way end up paying a salary to somebody with no system to run.

The most commonly mis-sequenced purchase at this size is a part-time CSR. At $18 to $22 an hour that is $1,500 to $2,000 a month part-time and $3,000 to $4,000 fully burdened for a full-timer. That hire is legitimate at two to five trucks, and it is still usually not the first thing to buy, for a structural reason: a person covers 40 hours and your phone rings for 168.

Shops in this trade miss 25% to 40% of inbound calls, and the misses are not evenly distributed — they cluster exactly where the money is. The first 95-degree afternoon, Saturday morning, 4:45 on a Friday when all three techs are on jobs and you are under a condenser. The homeowner with no cooling does not leave a voicemail. He calls the next result.

Run it against your own numbers. A three-tech shop taking 250 inbound calls a month that misses 30% is missing 75 calls. Say a third of those were real bookable jobs: 25 jobs a month, at a $450 average service ticket, is $11,250 a month of revenue that went to whoever answered second. Against that, an AI receptionist runs $49 to $500 a month depending on tier. This is not a close call, and it is the single clearest arithmetic in the whole stack.

Pick the tier honestly. Goodcall and Rosie at $49 to $199 a month are the right fit at two to five techs — they answer, qualify, book into Google Calendar, and text the customer a confirmation, with integration through Calendar and Zapier rather than native CRM writeback. Numa at $249 and up is the step up when you want the booking written directly into your CRM and text-back on every missed call. Avoca at $300 to $800 is built for 5 to 25 tech shops with a real call center and call-review workflows, and under five trucks you would be paying for capability you have nobody to operate.

What matters more than which vendor you pick is what happens after hours, and this is the part that gets skipped. Write down the escalation rule before you go live: what counts as a true emergency, who gets woken up for it, what the system is allowed to promise about arrival time, and what it should do with a call that is neither an emergency nor bookable. Almost all of the disappointment with AI receptionists at this size traces back to nobody having written that down. Then read the transcripts weekly for the first month — you will find two or three phrasings that need fixing, and after that it mostly runs itself.

While you are in there, start measuring booking rate: of the calls that were genuine service opportunities, what share became a scheduled job? Most shops at this size have never measured it, and it usually comes back between 40% and 60%. On 180 bookable calls a month, moving 50% to 70% is 36 additional jobs, roughly $16,000 at a $450 ticket, with no additional marketing spend. That number is why the phone deserves your attention before your dispatch board does.

Entry tiers on every major platform are priced for a single user. At two to five techs you are buying the second tier, and the jump is real. Jobber goes Core $49, Connect $149, Grow $349. Housecall Pro goes Basic $65 for one user, Essentials $169, Max around $450. FieldPulse starts at $79 and lands near $159 at this size. Workiz is flat-tier starting around $225 with a phone system built in rather than bolted on.

The genuine choice is narrower than the internet makes it sound. Housecall Pro Essentials at $169 and Jobber Connect at $149 are the two default answers for a residential shop at this size, and the difference between them is mostly ecosystem versus setup speed. Housecall Pro has the deeper integration list, which starts mattering the moment you bolt on a receptionist, a review tool, and QuickBooks. Jobber is faster to get running and slightly friendlier if you also do some plumbing or electrical work. FieldPulse at roughly $159 is the budget play and is genuinely competent for 2 to 10 techs — the trade-off is a thinner third-party integration list, which is the exact thing you are about to start relying on.

Workiz deserves a specific mention for one situation. If your real problem is the phone — no call recording, no idea which ad produced which job, techs handing out their cell numbers — then Workiz at $225 to $295 buys the phone system and the field service platform as one thing. At two to five trucks, wiring a separate VoIP, an AI receptionist, and a CRM together is more integration work than a shop your size has hours for, and the bundled version often wins on effort even when it loses on features.

Two costs are larger than the plan price and both get missed. The first is per-seat charges above the users included in your tier — check the exact included-user count before you sign, because that is the number that surprises shops at the fourth truck. The second is card processing at 2.6% to 3.5%. On $1.2 million of revenue with 70% collected on cards, that is roughly $25,000 a year, several times your entire software bill. Compare effective processing rates with more energy than you compare monthly prices; a half point at your volume outweighs the whole subscription.

Whatever you choose, connect it to QuickBooks Online at $30 to $90 a month on day one rather than month nine. Retrofitting a chart of accounts onto a year of already-recorded transactions is a weekend nobody enjoys, and until the sync exists you do not have job costing — you have a bank balance and a feeling.

At one truck, pricing consistency is automatic because there is only one of you. At three trucks it is gone, and the variance is wider than owners expect. The same repair, in the same neighborhood, priced by two different technicians, commonly differs by 30% to 40% — and close rates differ by more than that. Right now, which technician gets the call is a pricing decision, and you are not making it on purpose.

The fix is a flat-rate price book with good-better-best options, not a pep talk at the Monday meeting. Three written prices for every common repair and replacement, presented the same way at every kitchen table. What that buys you is not mainly a higher ticket, though it usually produces one. It is that average ticket and close rate stop being personality traits and become numbers you can coach. You cannot manage a close rate that depends on who was on the truck.

Sequence matters here: the price book comes before any quoting software. DinoQuote at $199 to $599 a month puts an instant replacement quote on your website and syncs the lead into your CRM as a full customer record, and it earns its keep at this size by capturing the homeowner comparison shopping at ten at night. But it is only as accurate as the price book behind it. Installing it first just automates your inconsistency and does it faster.

Diagnostic consistency is the other half of the same problem and it does more reputational damage. MeasureQuick at $30 to $120 per tech per month puts the same measurement standard on every truck, so a system passes or fails on data rather than on which technician looked at it. The practical effect at three trucks is that you stop getting the call where your guy said one thing in March and your other guy said the opposite in July. That conversation costs you a customer faster than a bad review does, and it is entirely self-inflicted.

Two supporting pieces worth their cost at this size. CompanyCam at $24 to $36 per user per month gives you a photo record attached to the job, which settles disputes, makes a replacement quote credible, and gives you something to show the homeowner who is not standing in the attic with you. And Interplay Learning at $35 to $150 per user per month is not urgent at four trucks, but it is what eventually lets you develop a green apprentice instead of competing for experienced technicians who are scarce and expensive. That shift is the real staffing constraint at your next stage, and shops that start it early are the ones that can add the sixth and seventh truck at all.

At two to five techs you are completing somewhere around 60 to 120 jobs a month. That is enough volume for review generation to be the highest-return marketing available to you, and almost nobody at this size runs it systematically.

The failure is never the customers — homeowners who had a good experience will leave a review at a decent rate if asked well and asked immediately. The failure is the ask. Reviews happen when the request goes out automatically within an hour of job completion, by text, with a direct link that takes one tap. Anything that depends on a technician remembering, or on somebody in the office working a list on Friday, gets done in January and abandoned in July, which is exactly backwards from when you need it.

The compounding is the point. At 80 jobs a month, a manual process converting 5% produces 4 reviews a month. An automated request converting 20% to 30% produces 16 to 24. Over two years that is the difference between a Google profile with 40 reviews and one with 300 — in a market where a homeowner with no heat is comparing three profiles on a phone and deciding in under a minute. The shop with 300 recent reviews wins that comparison before you ever speak to the customer, and the gap cannot be closed quickly once a competitor has it.

By tier: NiceJob at $75 to $149 a month is the right answer for most shops at two to five techs — it integrates with Housecall Pro and Jobber and does one job well without adding a platform to learn. Podium at around $399 a month bundles review generation with a shared business texting inbox, which is worth it only if you were going to buy a texting platform anyway. Birdeye at $299 to $1,200 is built for multi-location operations managing dozens of Google Business Profiles and is more platform than a single-location shop needs. Before buying any of them, check whether the tier you already pay for includes review requests — several do, and a meaningful share of shops at this size are paying for the feature and have never switched it on.

Do not buy an agency yet. Scorpion at $2,000 to $10,000 and up a month is a real full-service option for 16-plus tech companies with the volume to feed it, and at four trucks that monthly spend is a technician's payroll being traded for leads you cannot yet staff. The marketing sequence at your size is short and it is mostly free: a Google Business Profile filled out completely with real job photos, reviews flowing automatically, and your existing customer list contacted twice a year — once before cooling season, once before heating season. That list is the cheapest revenue you own, it is already in the CRM you are paying for, and most shops at this size have never once marketed to it.

Skip enterprise field service software. ServiceTitan runs $350 to $500 per tech per month plus implementation, which at four techs is $1,400 to $2,500 a month before onboarding and before the modules. The reporting and pricebook depth you would be paying for assumes a dispatcher, a CSR team, and somebody whose actual job is running the system. At two to five trucks you have none of the three, so you would be buying capability with nobody to operate it — the same mistake as the Avoca tier, just an order of magnitude more expensive.

Skip route optimization. Routific at $49 to $149 per vehicle solves multi-stop sequencing, which is a real problem at ten trucks across a spread metro and not yet a real problem at four. A zoning rule and a 30-minute cap on drive time between stops will get you most of the benefit for nothing. Revisit it somewhere around eight trucks.

Skip a dedicated dispatcher. At this size dispatch is two to three hours a day, not a role. What you need is for those hours to belong to one named person with the board open, not to be split between whoever is free — split dispatch is how double-bookings happen.

Then the useful part: the triggers that change each answer. Move up to a CRM-native receptionist like Numa when you pass roughly 300 inbound calls a month, because at that volume the manual re-entry of bookings becomes its own part-time job. Hire the second office person when your one office person is regularly past 45 hours, or sooner if quote follow-up is the thing being dropped — that is the most expensive corner you can cut. Add instant web quoting when you are running more than about five replacement quotes a month and losing them to competitors who answered faster. Add training software when you hire your first apprentice. Revisit the platform tier when you add the sixth truck or the second office user, whichever comes first.

The general rule for this whole stage: buy the thing that removes the constraint you can name this month. Owners at two to five trucks get in trouble by buying the stack they will need at twelve trucks on four trucks of revenue, and then spending a season configuring software instead of selling work.

Realistic all-in for a four-truck residential shop: platform $149 to $225, AI receptionist $49 to $199, review automation $75 to $149, QuickBooks Online $30 to $90, photo documentation $24 to $36 per user, diagnostics $30 to $120 per tech. Call the whole stack $400 to $900 a month, or roughly $5,000 to $11,000 a year. On $1 million of revenue that is half a percent to one percent — and your card processing alone will be two to three times that number. Software is not where your money is going, which is worth remembering the next time you spend an evening comparing $20 differences between plans.

Roll it out one thing at a time, and not in July or January. Two migrations at once in peak season is how shops end up back on paper.

Days 1 to 14, measure and buy nothing. Three numbers: how many inbound calls you took last month and how many went unanswered (your phone bill will tell you), how many quotes you sent in the last 90 days versus how many closed, and how many Google reviews you received in the last 12 months. Most owners are surprised by at least two of the three, and the surprises tell you which section of this guide applies to you first.

Days 15 to 30, fix the free things. Switch on review requests if your platform already has them. Write the after-hours escalation rule. Then pull every unclosed quote from the last 90 days onto a list and call every one of them. That exercise alone routinely pays for the entire year of software, and it costs you nothing but two evenings.

Days 31 to 45, buy the phone coverage. Pick the tier that matches your call volume, route it, then test it yourself by calling your own number at 7pm on a Sunday. Read the transcripts weekly for a month and fix the two or three responses that are wrong.

Days 46 to 60, standardize pricing. Build the flat-rate book with three options per job, roll it out with every technician in one room on one day, and then hold it for a full season before you judge it. Price books get abandoned in week three because one customer pushed back; that is not data.

After that, revisit the platform tier — and expect an anticlimax. Most shops at this size find they do not need a different platform. They need to use the two-thirds of the one they are already paying for that nobody ever turned on.