The Complete HVAC Marketing Guide for 2026

TLDR: The complete national guide to HVAC marketing in 2026: the three demand modes and why they need different channels, a climate-aware seasonal plan, local SEO and LSA setup, lead handling, and the measurement chain from spend to collected revenue. Updated August 2026; every statistic is sourced and dated.

HVAC Marketing in 2026: The Short Answer

HVAC marketing is the system that puts your company in front of a local property owner at the moment their equipment fails, ages out, or needs service, and turns that moment into a booked job your crew can actually run. If a tactic does not end in a booked job you have the capacity to complete, it is not working, whatever the dashboard says.

This guide is for owner-operators, general managers, and in-house marketers at residential and light-commercial HVAC contractors anywhere in the United States, from one truck to a multi-crew shop. It assumes no marketing training. It does assume you know your own book of business: what you sell, and how many jobs a day you can complete before the schedule breaks.

Demand in this trade arrives in three modes, and most wasted spend comes from treating them as one:

• Emergency. No cool, no heat, water on the floor. The decision window is minutes. They call whoever answers, looks available, and looks safe to let into the house.
• Planned replacement. An aging system the owner has decided to replace on their own schedule. The window runs weeks to months, usually with two or three quotes and a financing question.
• Maintenance. Tune-ups, agreements, seasonal service. No urgency at all, which means you create the timing instead of waiting for it.

The three modes want different messages, pages, response speeds, and budgets, and they behave differently by season and climate zone. That is why no honest national guide hands you one month-by-month calendar.

Pick your starting point by naming your real constraint:

1. The phone does not ring enough in peak season. Start with local visibility and paid demand capture.
2. The phone rings and jobs do not get booked. Fix answering, response time, and your website conversion path first; more traffic only widens the leak.
3. Peak is fully booked and the shoulder months are empty. Your problem is maintenance and replacement demand, not emergency demand.
4. You cannot tell which is true. Start with measurement before changing any spend.

No channel is universally correct. The right mix follows your market, season, service mix, crew capacity, and unit economics, roughly in that order.

Why HVAC Marketing Is Different From Ordinary Local Marketing

Most local marketing advice assumes a steady stream of buyers who can be persuaded over time. HVAC does not work that way. Your demand is triggered by weather and equipment failure, capped by how many trucks you can dispatch, and often decided in a single phone call about letting a stranger into a house. Four mechanics drive every recommendation in this guide.

The decision window varies enormously. An emergency buyer decides in minutes; a replacement buyer takes weeks and compares. A campaign built for one is useless for the other: the emergency buyer will not read a financing explainer, and the replacement buyer does not care that you answer at 2 a.m.

Trust is a safety judgment, not a brand preference. The buyer is deciding whether to let your technician into their home, usually while stressed. License and insurance status, named technicians with faces, recent reviews, and upfront pricing convert better here than clever creative.

Your service radius is a hard boundary. Reach outside your drive time is not merely wasted; it produces calls you decline, burning answer time in the hours that matter most (see our local SEO services page).

Capacity is a ceiling, and leads spoil. An emergency inquiry you do not reach within the hour is usually somebody else’s job. Once the board is full, extra leads produce unanswered calls and disappointed reviews rather than revenue.

The consequence: run the demand modes as separate campaigns, each with its own message, channel, and follow-up rule. This table is the reference the rest of the guide builds on.

What Actually Changed Going Into 2026

What the primary data shows

The Bureau of Labor Statistics counted 425,200 HVACR mechanic and installer jobs in 2024, projects 8% growth from 2024 to 2034, and expects about 40,100 openings a year; 70% of those workers were employed by plumbing, heating, and air-conditioning contractors in 2024. Read it as a capacity signal: for many shops labor is the binding constraint, not demand, and a plan that outruns hiring becomes missed calls.

System mix is regional. The EIA reports 88% of U.S. households used air conditioning in 2020, about two-thirds with central AC or a central heat pump. For heating in 2024, natural gas was the main fuel for 47% of households and electricity for 42%, and about 13% used a heat pump as their main heating equipment in 2020. Write for the systems in your service area, not the national average.

Repair mix is shifting. In Housecall Pro’s platform data, repair work rose from 21.6% of full-year HVAC revenue in 2021 to 31.3% in 2025, drawn from roughly 2 million tagged jobs representing about 11% of that platform’s HVAC job volume. That is one software platform’s dataset, not a census of the U.S. market. Use it to check your own repair-to-replacement ratio, not as a benchmark.

EPA’s Technology Transitions rule sets a 700 global-warming-potential limit for new residential and light-commercial air-conditioning and heat-pump products and systems, with a 2025-01-01 manufacture and import compliance date. Homeowners meet unfamiliar terms while researching replacement. Use EPA’s own wording, and never claim one named refrigerant is universally required.

Shifts with no clean number

• Answer engines summarize before anyone clicks. Put the question in the heading and the answer in the first two sentences; pages that build slowly to a conclusion are harder to quote — our guide to AI search visibility covers how assistants choose what to cite.
• Review recency outweighs review count. A steady trickle of recent reviews reads as an operating business; a pile of three-year-old ones reads as a business that stopped asking.
• Speed-to-lead keeps compressing. Homeowners contact several contractors in one sitting, so the first useful response wins a real share of emergency work.

Update the year on your own site too: a stale trends heading signals a business that stopped maintaining its storefront.

Segment by the Job in Front of You, Not by Demographic

Age, income, and household bands tell you almost nothing useful in this trade. What tells you everything is what the person is deciding right now and whether they have bought from you before. Five segments cover most residential and light-commercial shops, and each needs its own offer and its own route to reach it.

Two of them are worth more than the attention they usually get. Maintenance members are your seasonal shock absorber: they fill shoulder months, they call you first when something fails, and they replace through you. Past customers without an agreement are the cheapest replacement pipeline you own, because you already know the install date of their equipment and can time the conversation instead of buying the click.

The decision this resolves: build one landing destination and one follow-up rule per segment before you add another channel — purpose-built landing pages are what make the segmentation real. A single generic contact page asked to serve an emergency caller and a replacement researcher underperforms for both, and you will blame the traffic source rather than the page.

Goals, Capacity, and Unit Economics

Set your goal in booked jobs, not leads. Leads are a vanity unit here because they are cheap to generate and easy to lose in the answering step. Booked jobs are the first number that touches revenue, and cost per booked job is the only acquisition figure worth arguing about with a vendor.

Five numbers run the system, and all five already exist inside your CRM or field-service software:

• Booked jobs by service line, so replacement is not hidden inside a repair total.
• Cost per booked job — marketing spend divided by attributable booked jobs, not by leads.
• Book rate — the share of answered inquiries that become scheduled appointments. This is a phone-handling metric, not a marketing one.
• Close rate on quoted replacement work.
• Average ticket by service line, which decides how much a booked job is allowed to cost you.

Then set your capacity ceiling before you set budget: jobs per day per crew, multiplied by working days, minus the share of the schedule already committed to members and callbacks. That is your monthly job capacity. If a campaign is forecast to book past it, cap the campaign rather than discovering the ceiling through missed calls in the second week of a heat wave.

A hypothetical example, not a client result: a shop logs 100 tracked inquiries in a month. Eighty-two are answered live or returned within fifteen minutes; eighteen go to voicemail, and twelve of those never call back. Of the 82 answered, 45 become booked appointments — a 55% book rate on answered inquiries, but only 45% across everything the marketing paid for. Twelve of those visits produce replacement quotes and five close, a 42% close rate. Cost per booked job is that month’s marketing spend divided by 45.

Read the hypothetical for the leak rather than the ratios. The twelve inquiries lost to voicemail were already paid for. Recovering half of them raises booked jobs without adding budget, and moves cost per booked job further than any bidding change would. Measure your own version of that number first — our measurement methodology shows how tracking gets tied to booked revenue; until answering is fixed, extra budget buys extra waste.

The Seasonal Demand Engine, Adapted to Your Climate

Seasonality is not a calendar; it is a demand-and-capacity cycle that repeats around your local peaks. The four phases are the same everywhere — only the dates move.

The four phases

1. Pre-season, six to eight weeks before demand arrives. The highest-leverage window in the cycle. Update problem and service pages, verify call and form tracking, request reviews, reactivate past customers by email and SMS, and confirm answering coverage before the phones move.
2. Peak. Prioritize high-intent capture and live answering: emergency pages, click-to-call, honest availability, fast dispatch. Once the board fills, cap spend rather than buy calls you cannot serve.
3. Shoulder. Change the offer, not the effort: maintenance agreements, tune-ups, indoor air quality, planned replacement, financing education. This is also when you produce content and fix the conversion problems peak crowded out.
4. Post-peak. Reconcile spend against booked and completed jobs, collect reviews while the work is fresh, audit missed calls and lost quotes, and build next cycle’s audiences.

Setting the clock by climate zone

Cooling-dominant markets need their own note, because the two-peak model misleads them. Where cooling load runs most of the year and humidity never lets go, there is no quiet stretch to schedule maintenance into and no heating peak to balance it; coastal air adds its own wear. So maintenance gets booked deliberately across all twelve months, and pre-season outreach is timed to the hottest stretch, not a change of season.

Set triggers from two sources you already own: weekly booked-job history for two or three years, and local weather records. Count back six to eight weeks from the week your call volume actually climbs — that is the week your paid search campaigns should already be tuned and funded. A month borrowed from a national calendar will be wrong by weeks in most of the country.

Website conversion foundation

Your website has one job during a no-cool or no-heat search: start a phone call in under a minute, one-handed, from someone standing next to equipment that just quit. Every other design goal ranks below that. In a 2016 Google/Ipsos study — dated, and not HVAC-specific — 76% of people who searched on a smartphone for something nearby visited a business within a day. The window is short; a slow mobile page spends it for you.

Speed is the first gate. Check your homepage and top service pages in Google PageSpeed Insights on the mobile profile, serve images as WebP or AVIF, and delete third-party scripts nobody can justify. Then give the page two exits and no more: call now and book a time. A sticky click-to-call control belongs in the thumb zone on every page. The form beside it asks name, phone, ZIP, and what is wrong — four fields. Every extra field is a filter, and you do not want to filter an emergency.

Trust proof belongs where the decision happens, not on an About page: license number, rating and review count, financing, and the workmanship guarantee, within sight of the call button. Emergency traffic needs its own path: an after-hours page with honest hours, not a contact form implying someone is awake at 11pm. Paid traffic gets its own landing pages, one per service line, matching the ad.

QA the tracking before you judge any channel — everything in measurement, budget, and ROI depends on it. Call taps, form submits, and completed bookings should each fire as their own event; the usual failure is a tracked number that quietly stopped tracking after a redesign.

Website conversion checklist

• Sticky click-to-call visible on every page on mobile, without scrolling.
• Top service pages load usably on a mid-range phone over cellular.
• Lead form four fields or fewer; booking supplements the phone, never replaces it.
• License number, rating, review count, financing, and guarantee beside the primary CTA.
• A real emergency path with honest after-hours availability.
• One dedicated landing page per paid service line, matching its ad.
• Call taps, form submits, and bookings firing as separate tracked events.
• A monthly test call and test form submission from an outside phone.

HVAC SEO and content architecture

Organize the site by page type, not by keyword list. Four page types carry almost all HVAC organic demand, each winning a different query with different content. Build one page per service you actually sell, one page per market you actually serve, and problem and cost content for the research window before the call.

The decision most contractors get wrong is the city page. Build one only if you can put three things on it that exist nowhere else on your site: real completed jobs in that city, reviews from customers who live there, and named neighborhoods or landmarks you actually drive to. If you cannot, skip it. A page that swaps a city name into the same paragraph is thin, and enough of them drag down the pages that were working.

Internal links carry the structure. Problem and cost articles link up to the matching service page; the service page links down to the city pages that serve it; every page carries one path to the phone. That mirrors the demand modes in why HVAC marketing is different and keeps a homeowner researching replacement from landing on an emergency page.

Treat Core Web Vitals and clean indexing as hygiene, not strategy. Fix them because slow pages lose calls; do not expect a ranking jump for passing. Formatting for answer engines is the same discipline as formatting for a homeowner scanning on a phone: answer the heading’s question in the first two sentences, put comparisons in a table, define terms where they appear, and keep the specifics — ranges, timelines, conditions — in text rather than in an image. Get these four page types right and the rest of HVAC SEO is maintenance HVAC SEO service page.

Local SEO, GBP, citations, and reviews

Proximity is the one Map Pack factor you cannot change, so spend your effort on the three you can: category accuracy, service-area truth, and review velocity. Set your Google Business Profile primary category to the service you most want more of, not the broadest label available, and add secondary categories only for work you genuinely perform. Define service areas by real drive time, not by every county you would technically accept a job in — an inflated radius dilutes relevance in the towns you actually cover.

Keep the profile alive with evidence rather than announcements: job photos with dates, equipment you installed, the team, and posts tied to what is happening in your market this month. Answer questions in the Q&A yourself before someone else does. Across every listing, your name, address, and phone must match one canonical format character for character — including the suffix and suite line. If you use call tracking, keep the real number as the primary on the profile and add the tracking number as a secondary, so attribution never costs you NAP consistency.

Reviews are where most contractors leave the most on the table. BrightLocal’s 2026 Local Consumer Review Survey found 94% of consumers open to writing a review, 69% had written one in the past 12 months, and 45% had written one on Google. That is general local-consumer research, not HVAC-specific behavior — but it settles one argument: the customer whose AC you just fixed is probably willing. Most are simply never asked. Ask at the moment of relief, from the technician, and respond to every review inside a week, including the bad ones.

For citations, a listing earns your time if a homeowner in your market would plausibly use it or if the membership is real: Yelp, BBB, Angi, and trade associations such as ACCA and PHCC. Stop there. Do not stuff keywords into your business name, register addresses you do not staff, gate review requests to happy customers only, buy reviews, or blast hundreds of low-quality directories. Each of those is either a policy violation or wasted hours, and none of them outperform a correctly configured profile with fresh reviews local SEO and Google Business Profile service page.

Content, social, and trust-building

Content earns its place in HVAC only when it maps to a specific moment in the decision, so write for the four moments that produce revenue. No-cool and no-heat troubleshooting catches the panic search minutes before the call and should end by telling the reader exactly when to stop and dial. Repair-versus-replace and cost content owns the research window, where the ticket is largest and the decision takes days. Indoor air quality reaches customers who are not broken but are uncomfortable. Rebates and financing remove the reason a quote sits unsigned.

Your strongest trust asset is already on your payroll. Publish the technicians: names, faces, certifications, and two sentences each about the person who will walk into a stranger’s house. Pair that with before-and-after media shot on a phone with the customer’s permission — the corroded coil, the duct repair, the finished install with the site left clean. That material is credible in a way stock photography never is, and it works identically on a service page, a Google Business Profile post, and a social feed.

On cadence, be realistic. One substantial piece a month, produced properly and updated when facts change, beats a posting schedule you abandon in July. Cut each piece down rather than starting over: the article becomes a profile post, a short social post, an email to past customers, and a talking point for technicians in the field. Reuse is not laziness; it is how a small team covers the moments above without hiring.

Ignore platform hype. You do not need to be everywhere — you need to be where your customers already are, which for most residential contractors is Facebook, a neighborhood app, and the profile itself. A dormant account costs you little; an active one you cannot sustain costs you credibility. If capacity collapses mid-season, protect the cost and problem pages and the review responses first, and let the posting schedule lapse. Content also feeds the sequences in retention, maintenance, referrals, email, and SMS, so write it once with reuse in mind.

Paid demand capture: LSA and Google Ads

Local Services Ads and Google Search Ads are not two versions of the same buy. LSA buys leads and sits above everything on emergency queries; Search buys clicks and covers the far wider range of research and replacement intent. Run them for different jobs.

Local Services Ads and Google Guaranteed

Google Guaranteed requires license and insurance verification and a background check, and ranking inside LSA is influenced by review quality and how reliably you answer. That makes call handling an ad setting, not an operations detail. Charged leads that are out of your service area, for work you do not do, or plainly spam can be disputed — but the window is short, so make disputes a fixed weekly task rather than a quarterly cleanup, and record what you dispute so you can spot a pattern in the categories you enabled.

Search campaign structure

Separate campaigns by service line — AC repair, heating repair, installation, maintenance — because intent, average ticket, and allowable cost per booked job differ per line, and a shared budget quietly funds the cheapest clicks instead of the best jobs. Keep branded and non-branded apart so your name’s low cost never disguises what new demand actually costs. Build negatives before you build ads: jobs, salary, training, DIY, parts, manuals, rental, free, and the neighboring cities you will not drive to. Point each ad group at its own landing page. Then import booked and completed jobs back as offline conversions so bidding optimizes toward revenue rather than raw calls, using the allowable numbers from goals, capacity, and unit economics. Finally, tie budgets to the board: when the schedule is full for three days, cap spend instead of buying leads you will answer late.

Paid and organic social

Social does not capture urgent demand — it creates and reminds. Somebody scrolling a feed did not wake up looking for you, so judge these campaigns by maintenance signups, replacement quotes, and audiences you own, never by emergency calls. Confusing interruption demand with urgent search demand is the single most expensive mistake here: it leads contractors to shift budget out of the channels catching a live no-heat call and into the channel catching a bored evening.

Start with retargeting, because it is the cheapest honest win available. Build audiences from people who read a cost or repair-versus-replace page and did not call, then show them the objection they were stuck on — financing terms, what a quote includes, how long an install takes. Only then run prospecting, and give it a real reason to exist: a tune-up before the season turns, an indoor air quality offer, a rebate deadline. Prospecting without an offer is a brand ad, and most contractors cannot afford to pay for one.

Neighborhood proof is the creative that works. The install two streets over, the truck the viewer has already seen parked locally, a technician’s name and face, a review from someone whose town the viewer recognizes. Keep it visibly local and visibly real; polished production reads as a national brand and gets scrolled past.

The lead form versus landing page choice is a follow-up capacity question, not a creative one. Instant forms produce more leads at a lower cost, and they produce them from people who tapped twice with little intent; landing pages produce fewer and better leads because the visitor had to arrive somewhere and read. If you can call inside five minutes during business hours, forms will out-earn the landing page. If leads sit until the next morning, that cheap volume becomes wasted spend and a wall of unanswered names — use the landing page and let the page do the qualifying. Either way, route social leads into the same speed-to-lead workflow you use for search, with a text sent the moment the form clears.

Retention: maintenance plans, referrals, email, and SMS

Your cheapest booked job next season is a customer you already served — but only if someone contacts them on a schedule. Sell it at the truck, automate the reminders: a maintenance agreement converts best in the minutes after the repair is finished and the house is comfortable again. Enroll on site, book the first tune-up before the technician leaves, and let marketing own the reminders rather than the pitch.

Split members onto two tracks. The default is tune-up reminders keyed to your local pre-peak window; the second is replacement nurture, entered on a trigger — system age, a second repair on one unit in a season, or a quote crossing your repair-versus-replace threshold. Repair keeps growing as a share of revenue: in Housecall Pro’s platform data (about 2 million tagged jobs, roughly 11% of its volume, not a market census) it went from 21.6% of full-year HVAC revenue in 2021 to 31.3% in 2025 (Housecall Pro). One more tune-up reminder to a replacement candidate wastes the bigger job.

Ask for the review on the job; keep consent clean. Willingness is not the constraint: BrightLocal’s 2026 survey found 94% of consumers open to writing a review, 69% having written one in the past year (BrightLocal; general local-consumer research, not HVAC-only). Being asked — once, by the person who did the work, the day it went well — is. Keep transactional messages (arrival windows, missed-call text-back) separate from marketing ones, log consent, honor STOP permanently, and never move a dispatch list into a promotional send. Lapsed customers get one short pre-peak reactivation sequence, not a newsletter.

Sample member sequence:

1. Same day, on site: SMS confirming coverage and the next visit date, to kill buyer’s remorse. Owner: technician.
2. Next morning: one review request naming the technician; one reminder if unopened.
3. Day 7: email with the agreement, booking link, and after-hours number.
4. Six to eight weeks before local peak: tune-up offer with two date choices; same link by SMS three days later if unbooked.
5. Day after the visit: findings summary; if a trigger fired, switch tracks.
6. 45 days before renewal: notice with the cancellation path stated plainly.

Traditional and neighborhood marketing

Offline still works in HVAC because your market is a physical radius, not an audience segment. It stops working the moment you cannot tell which channel produced the call. One rule makes the category defensible: every offline channel gets its own tracked number or QR code and landing URL, and none may use your main number. A channel that cannot carry a unique identifier is brand spend — fund it from profit and stop calling it measurable.

Work the job radius, not the city. Door hangers dropped on the street where a crew just finished carry proof the neighbors can see: the truck was there this morning. Yard signs work on replacement jobs in walkable neighborhoods, with the homeowner’s permission and a code unique to that run. Direct mail is strongest aimed at your lapsed list or a build-year pocket where systems age out together, not a rented city-wide list.

The wrap and the technician are one channel. The wrap earns attention; the clean uniform and named technician convert it. Give the truck its own tracked number so it stops being invisible in reporting.

Referral partners beat sponsorships on measurability. Plumbers, electricians, home inspectors, and property managers each get one tracked number and a standing agreement about what you send back. Sponsorship is worth doing for reasons that are not attribution. Log every offline code at intake, in the same system your calls land in.

Lead handling and operations

This is where marketing dies. Every dollar in the sections above converts into one physical event — a phone ringing in your office — and if it rings out at 4:40 on a Friday, the money is gone while the ad report still counts it as a lead. Fix the answering before you buy more ringing.

Live answer is the first conversion rate. Target a human answering during published hours, with an overflow service that can actually book into your calendar rather than take a message. A message is not a booking; it is a second chance you have to earn. Emergency demand does not wait: in a 2016 Google/Ipsos study — dated, and about retail visits rather than service calls — 76% of people who searched on a smartphone for something nearby visited a business within a day (Google/Ipsos, 2016). Treat it as directional, not as your benchmark.

Qualify in the first ninety seconds, then book. Address inside the service area, homeowner or property manager, system type, symptom, urgency, and membership status — members skip the queue. Never end a first call with “someone will call you back to schedule.” Book the window, name the technician if you can, and hand off to dispatch with the call recording attached.

Give every lead a status and an outcome. A workable minimum: new, answered, qualified, booked, completed, collected — plus lost, with a reason code. Then close the loop:

• Speed-to-lead: form and lead-form submissions answered in minutes; after-hours rules written down before the season, not improvised during it.
• Missed calls: automatic text-back inside a minute, and one named person clearing the missed-call queue at fixed times daily.
• No-shows: confirm the day before and the morning of; feed cancellations straight into a same-day standby list.
• Feedback to marketing: review a sample of recorded calls weekly against a short rubric. Unqualified calls are a campaign problem; unbooked qualified calls are a training problem. They get different fixes.

Measurement, budget, and ROI

Measure the whole chain or you will optimize the wrong end of it: spend to lead to answered to booked to completed to collected. Four systems own segments of that chain — ad platforms, call tracking, your CRM or field-service software, and accounting — and the only version that matters is the one where a campaign name survives all the way to a paid invoice.

Three costs, three different jobs. Cost per lead audits the ad account and nothing else; it rewards cheap, unqualified volume. Cost per booked job audits the ad account plus your phone, and it is the number to judge channels on. Customer acquisition cost divides total marketing spend by new customers, not jobs — a member who books three times is one acquisition — and it is the number that tells you whether the business can afford the marketing at all. Compare seasons like for like: this July against last July, this pre-peak cohort against last year’s, never against the month before.

A hypothetical worked example. These are placeholder numbers chosen for clean arithmetic, not benchmarks. Use your own.

Revenue there assumes a $500 average completed ticket — again, a placeholder. Now change exactly one input. Lift the answer rate from 80% to 95% and hold everything else: 95 answered, 47 booked, about $85 per booked job, roughly 40 completed jobs, 5x on the same spend. To reach those same 47 booked jobs by buying instead, you would have to cut cost per lead by 15%. Answering the phone is the cheaper project. Build the reporting behind it — the measurement methodology we publish shows the full chain.

The 90-day launch plan and the 12-month rhythm

Sequence beats effort here. Traffic bought before tracking and answering are fixed produces an expensive month with nothing to learn from, so the first thirty days buy no new demand at all.

Days 1-30: baseline and tracking

1. Count the last twelve months: leads, booked jobs, completed jobs, revenue by service line.
2. Install call tracking, verify form and call conversions end to end, and set lead statuses in the CRM.
3. Measure your real answer rate and after-hours behavior for two full weeks.
4. Set the capacity ceiling: jobs per day per crew, and the lead time you will accept.

Days 31-60: conversion and local foundation

1. Fix the website’s mobile speed, click-to-call, and short form; add the emergency path.
2. Complete the Google Business Profile — categories, service areas, hours, photos — and start the review ask on every job — this is the local SEO foundation everything later leans on.
3. Publish or rewrite the core service pages and the repair-versus-replace content.
4. Stand up missed-call text-back and the written after-hours rules.

Days 61-90: demand capture and retention

1. Launch paid capture on your highest-intent service lines only, with capacity caps — LSA and Google Ads come after the foundation, not before.
2. Turn on the maintenance-member and reactivation sequences from the retention section.
3. Start the weekly scorecard; hold the first monthly channel review at day 90.
4. Begin pre-peak preparation six to eight weeks ahead of your local demand curve.

Then the rhythm. Weekly, fifteen minutes: leads, answer rate, booked jobs, missed-call queue, spend pacing, new reviews. Monthly: cost per booked job by channel, service-line mix, organic and Map Pack trend, membership net gain. Quarterly: reallocate budget and prepare the next peak. Stop and scale rules, decided in advance: scale a channel only after two consecutive review periods at or under your allowable cost per booked job and with capacity to absorb the work; pause it when answer rate slips below target — fix operations first — or when it misses target twice running after one structural fix. Full boards get budget cuts, not more leads.

Common mistakes and what to do first

Most stalled HVAC marketing fails in a small number of recognizable ways.

• Thin city pages. Twenty near-identical pages for towns you do not staff are a liability, not coverage. One page per market you genuinely serve, with real proof from that market.
• Untracked calls. If the phone is not tracked, you are optimizing on form fills while most of your revenue arrives by voice.
• Buying traffic before fixing response. Paid demand amplifies whatever your phone process already is, including the bad version.
• Scaling past capacity. Hiring is the binding constraint: BLS counted 425,200 HVACR mechanic and installer jobs in 2024 and projects about 40,100 openings a year on average through 2034 (BLS; workforce context, not a market forecast). Leads you cannot serve become bad reviews.
• Treating all climates alike. A national calendar copied from a blog does not match your demand curve. Even the installed base splits: in 2024, natural gas was the main space-heating fuel for 47% of US households and electricity for 42% (EIA, 2024 data). Build your calendar from local weather and your own booked-job history.
• Judging channels on cost per lead. It is the metric most likely to make you buy worse jobs.
• Asking for reviews only when things go wrong. Unasked-for reviews skew negative; a routine ask on every completed job is the fix.

The action ladder, in order. Measure the baseline. Fix answering and tracking. Build the local and organic foundation. Then buy demand capture. Then automate retention. Then, and only then, scale — one channel at a time, against capacity. Each rung depends on the one below it, which is why the guide is ordered this way and why skipping to paid ads is the most common expensive mistake in the category.

Frequently asked questions

How do I set a marketing budget before I know my own numbers?

Work backwards from one job instead of forwards from a percentage. Take your average completed ticket, subtract direct costs, and decide what share of the remaining margin you will pay to win the job. That is your allowable cost per booked job. Multiply it by the jobs your crews can absorb this month and you have a ceiling that is true for your business, not an industry rule of thumb.

Should I hire an agency, hire in-house, or do it myself?

Split by function, not by loyalty. Answering, booking, and the review ask stay in-house — they are operations wearing a marketing hat. Content about your own market needs an owner or senior technician’s input. Paid account management, tracking implementation, and technical site work are specialist trades where a competent outsider beats a part-time attempt. Keep whoever owns the numbers close enough to question weekly.

How long before local SEO produces booked jobs, and what do I watch meanwhile?

Months, not weeks, and how many depends on how competitive your market already is. Track leading indicators instead of refreshing rankings: Google Business Profile calls and direction requests, impressions and clicks on your service and city pages in Search Console, review count and recency, and the share of booked jobs sourced to organic or the Map Pack. Those move before revenue does.

Do I need a separate page for every city I serve?

Only for markets you actually staff and want more work in. A useful city page names local landmarks or neighborhoods truthfully, shows jobs and reviews from that area, states response times you can honor there, and reflects the systems common in that area’s housing. If you cannot write those things without inventing them, the market does not deserve a page yet.

Are purchased or shared leads worth buying?

They can be, but only under two conditions. First, judge them on cost per booked job like any other channel, never on the price per lead the seller quotes. Second, be honest about speed: shared leads go to several contractors at once, so whoever calls first usually wins, and a business that cannot answer inside minutes is buying leads for its competitors.

What do I do with marketing when the schedule is already full?

Throttle demand capture, not the whole program. Pause paid budgets on the service lines that are booked out, keep retention and review generation running, and shift messaging toward work that can be scheduled later — maintenance agreements, replacement consultations, duct and indoor air quality. Protect members’ access to the calendar. A full board is a moment to raise price, not to buy leads.

How should I respond to a negative review?

Publicly, quickly, once. Acknowledge the specific problem, avoid arguing facts or discussing the customer’s home in public, name a direct contact, and move the resolution offline. Future readers are the real audience, and they are judging your temperament more than the original complaint. After it is resolved, ask — do not pressure — whether the customer would update it. Never buy, incentivize, or write reviews.

We get plenty of calls but the jobs are too small. How do I change the mix?

The mix follows the offer and the page, not the ad budget. If most of your spend points at repair and no-cool queries, you will get repair. Build separate paths for planned replacement — cost and efficiency content, financing information, a consultation booking rather than a dispatch — and give technicians a documented trigger for flagging replacement candidates. Then measure revenue by service line, not lead count.