What are the most important KPIs every HVAC company should track in 2027?

A shop that grew revenue and lost money

Picture a residential HVAC company running eleven trucks in a Sun Belt market. Last year it billed roughly $9.4 million. This year it is pacing toward $11.2 million, and the owner is quietly panicked, because the operating account is thinner in August than it was the previous August. The revenue line went up nineteen percent and the cash went down. Nothing in the monthly P&L explains it, because the monthly P&L is the wrong instrument for this question.

Here is what the underlying numbers actually said once someone pulled them apart. Call volume rose about twenty-two percent, driven by a bigger paid-search budget. Booked-call conversion — the share of inbound calls that turned into a scheduled appointment — fell from roughly eighty-four percent to seventy-one percent, because the CSR team stayed at three people while the call load grew. That thirteen-point drop meant the company paid full price for demand and then dropped one call in seven at the phone. The marketing spend showed up in the expense column immediately; the revenue it should have produced never arrived.

Meanwhile the mix shifted. Replacement jobs, which had been thirty-one percent of revenue, dropped to twenty-four percent, because technicians were being pushed through more calls per day and stopped taking the fifteen extra minutes to walk an aging system's condition with a homeowner. Service revenue grew, replacement revenue stalled, and blended gross margin fell about four points — invisible on a single blended margin line, obvious the moment you split it.

And the fleet quietly got less productive. Two new trucks were added mid-season. Revenue per truck-day slipped from about $2,150 to roughly $1,780. The company was carrying two more vehicle payments, two more insurance lines, two more sets of stocked parts, and two more technician salaries, against a fleet that was individually less productive than before. Total revenue is a sum; it goes up when you add capacity even if every unit of capacity gets worse. That is exactly what happened.

This is the failure a good HVAC scorecard is designed to catch. Every number that mattered was a ratio, not a total, and none of the ratios were on the wall. The owner was reading the one number that could not tell him anything — the top line — and the four numbers that would have flagged the problem in April were sitting unqueried inside the field service platform the whole time. The rest of this page is about which ratios earn a place on that wall, what they should read, and what to do when they read wrong.

How the measurement chain actually works

An HVAC scorecard is not a list of independent numbers. It is a chain, and each link multiplies the one before it. That structure is the whole reason ratio metrics beat totals here: a five percent slip in three consecutive links compounds into a fifteen percent revenue loss that no single metric flags.

The chain starts with demand arriving — a phone call, a form fill, a membership renewal reminder. Booked-call conversion rate governs how much of that demand becomes a scheduled job. This is the cheapest link to fix and the most commonly broken. Every point of conversion you recover is revenue you have already paid for. Measure it as booked appointments divided by opportunity calls, and be strict about the denominator: exclude vendor calls, wrong numbers, and existing-customer scheduling changes, or the number flatters you.

Booked jobs then hit dispatch, where two things get decided at once — which technician goes, and how far they drive. Dispatch quality shows up downstream as revenue per truck-day and as first-time fix rate, which is why dispatch is the single highest-leverage seat in the building and almost never measured directly. A dispatcher who matches an older-system diagnostic to your strongest comfort advisor rather than to whoever is geographically closest is trading twenty minutes of drive time for a materially better close rate.

At the job, three things get determined. First-time fix rate decides whether this truck-day produces one billable job or half of one plus a callback. Average ticket captures what the visit was worth. And the repair-versus-replace conversation decides whether a failing eighteen-year-old condenser becomes a $600 patch this month or a $9,000 system this week — a decision with an order-of-magnitude revenue difference and a completely different margin profile.

Downstream of the job sits the part most operators underweight: whether the customer leaves the interaction on a maintenance agreement. Membership penetration is the metric that converts a transactional business into a recurring one. It smooths seasonality, it gives you first call when the system finally dies, and it is the single largest driver of what an acquirer will pay for the company. Everything upstream is a flow; membership is the reservoir.

Read the chain left to right and the diagnostic logic falls out. If revenue is soft, ask which link moved. Conversion soft means a front-office problem. Truck-day revenue soft with conversion healthy means a dispatch or capacity problem. Ticket soft with both healthy means a pricing or presentation problem. Margin soft with ticket healthy means a cost or mix problem. You are not guessing; you are walking the chain until you find the link that moved.

Real numbers, ranges, and what good looks like

Benchmarks in home services vary by market, mix, and how a given platform defines a field, so treat every range below as a starting frame you calibrate against your own trailing twelve months rather than as a verdict. The direction of travel in your own data matters more than hitting someone else's number.

Booked-call conversion rate. Well-run residential shops with a trained, scripted, incentivized CSR team commonly run in the low-to-mid eighties. Under seventy percent, you have a front-office problem that is more expensive than any marketing channel decision you could make this quarter. The fix sequence is cheap and ordered: record and score calls weekly, staff to peak-hour call volume rather than average, remove price-quoting from the phone script, and make booking the only acceptable outcome of an opportunity call. Recovering ten points of conversion on 800 monthly opportunity calls at a $520 average service ticket is roughly $41,000 of monthly revenue against zero incremental ad spend.

Average ticket, split service and replacement. Track these as two separate numbers forever; a blended average ticket is a meaningless statistic because it moves whenever mix moves. Residential service tickets commonly land somewhere in the $400–$700 band depending on market and whether diagnostic fees are captured; replacement tickets in the high four figures to low five figures depending on equipment tier, ductwork, and efficiency requirements. Watch the trend and, more importantly, watch *why* it moved. A rising service ticket driven by accessory attach — surge protection, float switches, media filtration, membership enrollment — is durable. A rising ticket driven by pressure is a review-score problem and a warranty problem arriving on a delay.

Membership penetration. This is the metric to fight for. Strong residential operators run somewhere in the thirty-five to fifty percent range of their active customer base on a recurring agreement; many shops sit under fifteen percent and do not realize how much stability they are giving up. The mechanics that move it are unglamorous: every technician offers on every call, the offer is scripted and role-played, enrollment is a tracked technician metric reviewed weekly, and the agreement is priced monthly rather than annually so it competes with a streaming subscription rather than with a repair bill. Two hundred new members at $22 monthly is roughly $53,000 of annual recurring revenue that also generates two hundred guaranteed maintenance visits — visits that produce repair findings, replacement leads, and referrals.

Revenue per truck-day. Total revenue divided by truck-days actually run, not trucks owned. A healthy residential truck in a normal market generates somewhere in the $1,800–$2,500 range per day; premium replacement-heavy operations run considerably higher. This is your capacity metric, and it is the one that catches the growth-without-profit failure described above. If you are adding trucks, watch this number weekly during the ramp; a new truck should reach seventy percent of fleet-average truck-day revenue within about ninety days, and if it has not, the problem is dispatch feeding it or training, not the truck.

Gross margin by job type. Split service, maintenance, and replacement, and load each honestly — technician wages, burden, vehicle cost, and parts, not just parts. Service and replacement behave nothing alike: service carries high labor content and higher percentage margin on smaller tickets; replacement carries heavy equipment cost with lower percentage margin on much larger tickets and far more gross profit dollars per truck-day. Managing them as one blended number is the single most common way an HVAC company mis-prices itself into a slow bleed.

First-time fix rate. The share of jobs closed on the first visit. Strong operators run above ninety percent. Every callback consumes a truck-slot you cannot resell, generates zero revenue, and lands during the exact window when the customer is deciding whether to recommend you. The three causes, in the order they usually apply: truck stock is wrong for your actual failure mix, diagnostic skill is uneven across the team, or dispatch is sending the wrong skill level to the call. Track fix rate *by technician* — a fleet average of eighty-eight percent can be one tech at sixty-eight and the rest at ninety-four, which is a training problem wearing a fleet-metric costume.

Cost per booked call and customer acquisition cost. Measure spend per *booked* call, not per lead, or you will optimize a channel that generates cheap calls nobody books. Fully loaded CAC only means something read against membership penetration and replacement conversion, because HVAC lifetime value is back-loaded — the acquisition pays off on the agreement renewals and the eventual system replacement, not on the first ticket.

Net revenue retention on service agreements. Year-over-year revenue retained and expanded from the existing membership base. Above one hundred percent means members are renewing and upgrading. Healthy renewal rates on residential agreements typically run in the high eighties to mid nineties; below eighty-five percent, the value proposition needs work — usually because the maintenance visit itself is being rushed and the member cannot articulate what they are paying for.

Capacity and cash supporting metrics. Billable-hour utilization — technician shift hours that produce revenue divided by hours paid — is the truest efficiency read; most residential shops land in the sixty-five to eighty-five percent band, and sustained sub-sixty-five means dispatch inefficiency or overstaffing for current demand. Route density, expressed as calls per technician per day with average drive time between stops, is what makes utilization move: three to four and a half residential service calls per day with drive times under about twenty minutes is a workable target, and density collapse is what quietly raises cost per call. On the cash side, days sales outstanding matters mainly on the commercial side of the book, where net-30 to net-60 terms are normal and a DSO drifting past seventy-five days means collections, not sales, is your constraint. Residential collects at the point of service and should barely register.

Trade-offs, tensions, and what to leave off the wall

Every metric on this list can be gamed, and several of them pull against each other. A scorecard that ignores those tensions produces a shop that optimizes one number into the ground.

Average ticket versus trust. Push average ticket as a standalone technician-level target with commission attached and you will get a higher average ticket within a month — partly from genuine accessory attach and partly from work the customer did not need. The damage arrives on a lag, in review scores, callback rates, and membership churn. The counterweight is to pair ticket with a customer-satisfaction or review metric and with membership retention, and to review outlier tickets by hand. Any single metric with money attached and no counterweight will be gamed; that is not a character claim about technicians, it is what incentives do.

Speed versus first-time fix. Calls per truck per day and first-time fix rate pull in opposite directions. Compress every call and diagnostic quality degrades, callbacks rise, and the truck-days you thought you gained come back as unbillable return visits. The right target is not maximum calls; it is maximum *completed, non-callback* calls. Track fix rate alongside call count and the tension resolves itself.

Replacement conversion versus reputation. A replacement-to-repair ratio that is too low means technicians are patching systems that should be replaced, leaving revenue on the table and leaving customers to fail again in August. Too high, and you have a sales-pressure problem with warranty and review consequences. There is no universally correct ratio because it depends on the installed-base age in your service area; what matters is that the ratio is stable and that replacement recommendations survive a spot audit of the diagnostic notes.

Membership growth versus membership quality. Discounting an agreement to hit a penetration target buys members who churn at renewal and drag NRR down a year later. Penetration and net revenue retention have to be read as a pair. Penetration climbing while NRR falls is a sign you are buying enrollments rather than earning them.

Lead volume versus booked-call conversion. The reflex when revenue softens is to buy more leads. If conversion is the broken link, more leads make the P&L worse — you pay for the demand and drop the same share of it. Fix conversion first; it is nearly always cheaper than the next increment of ad spend.

There is also a trade-off in the scorecard itself. Every metric you add costs attention, and attention is the scarcest resource in a fifteen-person office. A daily wall with four numbers that people actually look at beats a thirty-metric dashboard nobody opens. The metrics that earn daily placement are the ones where a bad day is unrecoverable: booked-call conversion and revenue per truck-day. Weekly gets average ticket, first-time fix rate, membership enrollments, and replacement conversion. Monthly gets margin by job type, cost per booked call, and penetration trend. Quarterly gets NRR, lifetime value by segment, and the view an acquirer would underwrite. If a metric cannot be assigned to a cadence and an owner, it does not belong on the board.

Common pitfalls and how to avoid them

Reading totals instead of ratios. The scenario at the top of this page is the canonical version: revenue grew, every ratio underneath it degraded, and the totals concealed all of it. Totals answer "how big"; only ratios answer "how well." Put ratios on the daily board and keep totals for the monthly close.

Blending margin across job types. Blended gross margin is the most reliably misleading number in home services, because it moves whenever the service-to-replacement mix moves and tells you nothing about whether either side got better. Split it in the first week of any measurement project. Everything else is easier once it is split.

Trusting the platform's default field definitions. Whatever system runs your dispatch — ServiceTitan and Housecall Pro are the two most common — has opinions about what counts as an opportunity call, a completed job, or a membership. Those opinions may not match yours. Before you set a single target, write down your own definition of each metric on one page, confirm what the report is actually counting, and freeze the definition. Half of the "our numbers are wrong" arguments in HVAC are definition arguments in disguise, and a target set against a misunderstood denominator produces confident, wrong decisions.

Measuring at the fleet level when the variance is at the person level. Fleet averages hide the individual. Fix rate, average ticket, membership enrollment, and replacement conversion should all be visible per technician, because that is where the coaching happens. The eight-point spread between your best and worst technician on membership enrollment is usually worth more than any marketing change available to you.

Setting targets without staffing the constraint. Publishing an eighty-five percent conversion target while the CSR team is understaffed for peak-hour volume just creates a number people learn to explain away. Check whether the constraint is skill, staffing, or process before you set the target, and staff the phones to peak rather than to average — the calls you miss cluster in exactly the hours you are thinnest.

Ignoring callbacks because they are not invoiced. A callback generates no revenue, so it can vanish from a revenue-oriented report entirely. It still consumed a truck-slot, fuel, and a technician's day. Count callbacks explicitly, tie them back to the original job and technician, and treat the truck-day they consumed as a real cost.

Chasing seasonal noise. July and October are different businesses. Comparing them tells you about the weather, not about your operation. Compare against the same period last year and against a trailing twelve-month view; that is what separates structural change from seasonality, which is precisely why membership and margin metrics deserve more weight than monthly revenue in a business this seasonal.

Building the scorecard nobody reads. The most important discipline is subtraction. Every company that tries to track everything ends up tracking nothing, because a board with thirty numbers has no signal. Pick the handful that map to the chain — conversion, truck-day revenue, fix rate, ticket, margin split, penetration, NRR — assign each a cadence and an owner, and delete the rest until one of them stops earning its place.

Related questions

How many KPIs should a small HVAC shop actually track?

Four to six on a visible daily and weekly board, with a slightly wider monthly review. A five-truck shop should live on booked-call conversion, revenue per truck-day, first-time fix rate, and membership enrollments. Add margin split and NRR at the monthly and quarterly cadence.

What KPI matters most if we can only track one?

Revenue per truck-day, because it is the only single number that catches capacity problems, dispatch problems, and pricing problems at once. It is also the number that exposes growth-by-adding-trucks that is actually shrinking per-unit productivity underneath the top line.

How does commercial HVAC change the scorecard?

Add days sales outstanding and contract renewal rate, since commercial invoices on net-30 to net-60 terms and revenue concentrates in fewer, larger accounts. Membership penetration becomes contract coverage of the installed base, and backlog value matters more than daily call volume.

Do these KPIs change what a buyer pays for the company?

Substantially. Acquirers underwrite recurring revenue, so membership base size, penetration, and net revenue retention drive the multiple more than raw revenue does. A shop that cannot document its agreement retention gets valued closer to a labor business than a platform.

How long before a new scorecard changes results?

Expect definitions and clean baselines in the first month, the first real movement — usually booked-call conversion — in the second, and margin and penetration trends to become readable by the end of a quarter. NRR needs a full renewal cycle before it means anything.

FAQ

How often should each of these be reviewed?

Match the cadence to how fast a bad reading becomes unrecoverable. Booked-call conversion and revenue per truck-day are daily, reviewed in a short morning huddle, because a lost booking in peak season is capacity you can never resell. Average ticket, first-time fix rate, replacement conversion, and membership enrollments are weekly. Gross margin by job type, cost per booked call, and penetration trend are monthly. Net revenue retention and lifetime value are quarterly. Anything reviewed on a cadence slower than the decision it informs is decoration.

Do we need ServiceTitan, or is a simpler system enough?

The platform matters less than the discipline. ServiceTitan and Housecall Pro are the two most common operating systems in residential home services and both can produce every metric on this page. A five-truck shop can run a perfectly good scorecard out of a simpler system plus a spreadsheet; a forty-truck shop generally cannot, because the reconciliation work grows faster than the office. Pick the system you will actually configure correctly and audit, rather than the one with the longest feature list.

What is the fastest KPI to improve from a standing start?

Booked-call conversion, almost always. It requires no new spend, no new hires in most cases, and no equipment. Record calls, score a sample weekly against a simple rubric, coach the two or three specific behaviors that lose bookings — quoting prices on the phone, failing to offer a same-day slot, not asking for the appointment — and staff the phones to peak-hour volume. Shops that start under seventy percent frequently move several points within a month.

How do we set targets without industry benchmarks we trust?

Use your own trailing twelve months as the baseline and target directional improvement. Pull the last four quarters for each metric, look at the distribution rather than the average, and set the target near the top quartile of your own history — you have already proven you can hit it. External benchmarks are useful for sanity-checking that you are not wildly off, but definitions vary enough between sources that treating them as targets imports someone else's measurement conventions along with their number.

Should technician pay be tied to these metrics?

Tie pay to metrics only in balanced pairs, never singly. Average ticket alone produces over-selling; membership enrollment alone produces discounted enrollments that churn; call count alone produces rushed diagnostics and callbacks. If you attach compensation, attach it to a small combination — ticket plus review score, or enrollment plus retention at renewal — so that gaming one side costs the other. And audit outliers by hand, because the distribution tells you more about incentive damage than the average does.

What is the first thing to do if the numbers look wrong?

Check the definition before you check the performance. Confirm what the report counts as an opportunity call, a completed job, a truck-day, and an active member, and confirm it matches what you meant. Then check whether the metric is measured at the level where the variance lives — fleet averages routinely hide a single outlier technician or a single broken dispatch rule. Only after both checks should you conclude that performance actually moved.

Sources

• https://www.servicetitan.com/blog
• https://www.housecallpro.com/resources/
• https://www.acca.org/
• https://www.ashrae.org/
• https://www.achrnews.com/
• https://www.energy.gov/energysaver/heat-and-cool
• https://www.contractingbusiness.com/
• https://www.bls.gov/ooh/installation-maintenance-and-repair/heating-air-conditioning-and-refrigeration-mechanics-and-installers.htm

Related on PULSE

• [What are the most important KPIs every catering company should track in 2027?](/knowledge/ik460)
• [What are the most important KPIs every property management company should track in 2027?](/knowledge/ik457)
• [What are the most important KPIs every dermatology practice should track in 2027?](/knowledge/ik468)
• [What are the most important KPIs every escape room should track in 2027?](/knowledge/ik467)
• [What are the most important KPIs every laundromat should track in 2027?](/knowledge/ik466)
• [What are the most important KPIs every dog boarding and daycare business should track in 2027?](/knowledge/ik465)