HVAC Seasonal Marketing: The Shoulder-Season Playbook

Two peaks and two dips produce four distinct marketing postures across the year, not two:

1. Cooling peak (roughly May-August). Demand exists; the job is capture speed — answer rate, ad coverage, capacity messaging.
2. Fall shoulder (roughly September-October). Heating changeover queries appear before the cold does. This is the one shoulder window where search itself carries pre-season demand.
3. Heating peak (roughly November-February). Capture again, on heating terms, with no-heat urgency.
4. Spring shoulder (roughly March-April). Cooling changeover queries are thin. Demand has to be manufactured out of your own customer list.

Those month labels are national shorthand and your climate division moves them. EIA notes that "the West North Central Census Division generally has the most HDDs, and the West South Central Census Division generally has the most CDDs each year" — so a Minneapolis shoulder window and a San Antonio shoulder window are not the same weeks, and a national calendar applied literally will launch your fall campaign a month late in one market and a month early in the other.

The demand shape also decides which asset does the work. Paid capture pays peak-month prices in the peak month. Ranked pages and an owned customer list are what carry a shoulder month, because they cost the same in April as in July.

Two of those deserve the detail generalist agencies skip.

Planned replacement is a shoulder-season product, not a peak-season one. A homeowner whose system dies in July buys under duress and buys fast. A homeowner in April buys on comparison, financing and efficiency — which is a longer, better-margin conversation your techs have time to have. The refrigerant transition gives that conversation a real, citable hook: EPA's Technology Transitions restrictions set a 700 GWP limit for residential and light commercial air conditioning and heat pump systems with a January 1, 2025 compliance date, and the agency's footnote records "No installation compliance date for systems where all specified components were manufactured or imported prior to January 1, 2025." That is a genuine change in what is on the truck and what a replacement quote looks like. Explain it; do not weaponise it into a deadline that does not exist.

Maintenance has a published standard, and almost nobody markets with it. ACCA 4 QM – 2019 (R2024), Maintenance of Residential HVAC Systems, is the designation ACCA publishes, and ANSI approved its reaffirmation on August 7, 2024. ACCA states the standard "provides a nationally recognized, manufacturer-endorsed set of minimum tasks that should be performed for HVACR equipment maintenance inspections." A shoulder-season tune-up offer that names the standard and lists the tasks it covers is differentiated against a competitor advertising a discounted tune-up with no scope attached at all. HomeGuide's own hiring advice points the same way: it tells homeowners to "Ask for a checklist of items the AC tech will inspect as part of the tune-up" and to "Beware of companies that offer low-cost AC tune-ups." The scope is the differentiator the buyer is already being told to ask for.

One discounting caution belongs here. HomeGuide's AC service cost guide, written by Kristen Cramer and dated September 14, 2023, lists season among its cost factors and records that "AC companies may offer discounts during the off-season" — alongside plans that typically bundle two tune-ups plus 5%-15% discounts on repairs or parts. Off-season discounting is therefore the market's default expectation, not an edge. Discounting the agreement rather than the visit keeps the recurring relationship intact while still giving the shoulder month a reason to act.

The changeover questions confirm the asymmetry at a much smaller scale. when to turn on heat runs 12,100 in October 2025 against 40 in July 2026 — the widest peak-to-trough ratio of any keyword in our 150-term pull that returned a full 12-month series. Its cooling-side mirror, when to turn on ac, peaks at only 880 in April 2026. Same behaviour, same calendar logic, roughly one-fourteenth the volume.

That produces two different pre-season plays rather than one repeated twice:

The sequencing rule that falls out of this: fund the fall campaign against a search curve and the spring campaign against a customer list. A spring campaign built on cooling-search capture is bidding into demand that has not arrived yet; the same budget spent on your own lapsed-customer file is buying an audience whose contact cost is already sunk.

Weather triggering is the one tactic the ranking pages on this query push hardest, and it deserves an honest description. Google documents weather as a scripting integration rather than a campaign setting. Its Google Ads help page on targeting your ads covers two families — audience targeting, which reaches people by "Who they are", "Their interests and habits", "What they're actively researching" and prior interaction with your ads, website or app, and content targeting (topics, placements and content keywords) — and no weather condition appears among them. Where weather does appear in Google's own documentation is the scripting layer: "A powerful feature of Google Ads scripts is the ability to integrate with data and services from third-party APIs," with UrlFetchApp as the mechanism, and the documented example fetches weather data from OpenWeatherMap. That page also cautions that "In writing or using a script that makes use of a third-party API, it is recommended to review the Terms and Conditions associated with the API to ensure your intended use complies with them." So weather-triggered bidding is real, it is a build, and it lives in the peak-season capture layer — a first heat wave or first freeze, not a shoulder month where the whole point is that the weather is doing nothing.

First, the demand curve for maintenance is nearly flat. In the same 2026-08-17 pull, hvac maintenance moves between 33,100 (November 2025) and 74,000 (July 2026) — a 2.2-fold band against the 17-fold swing on furnace tune up. duct cleaning is flatter still, sitting between 27,100 and 49,500 every month of the window. Flat demand is what a shoulder month needs: a product whose audience does not disappear when the temperature is comfortable.

Second, shoulder-season calls convert better. ServiceTitan's aggregated platform data from more than 3,000 trade businesses, June 2022, records HVAC booking 38% of calls, and states plainly that "HVAC booking rates varied with the seasons, with higher booking rates in shoulder seasons and lower booking rates in summer." Fewer calls, each worth more of your CSR's attention. A shoulder month is the cheapest month to fix your call handling, and the most expensive month to leave it broken.

The agreement math is arithmetic, and it runs in one direction: shoulder capacity first, agreement count second.

Agreements needed = (techs × visits per tech per day × working days in the shoulder window) ÷ visits per agreement per window

A standard agreement schedules two visits a year — HomeGuide records "AC tune-ups twice per year" as a typical plan inclusion — so one visit per shoulder window, which makes the divisor 1 and the arithmetic direct:

Every figure in the italicised rows is a placeholder, not a benchmark — the visits-per-day and working-days numbers are modelling assumptions you replace with your own dispatch data, and the revenue column uses HomeGuide's published national plan range rather than your pricing.

The table's real output is not the revenue line. It is the agreement count — the number that tells you whether your shoulder problem is a marketing problem or a base problem. A four-truck shop needing 640 shoulder visits and holding 200 agreements has a 440-visit gap that no April email blast closes; it closes over two or three peak seasons of disciplined attachment at the point of repair. That gap is the honest scope of what it takes to build shoulder-season demand you own, and it is why the fix starts in July, not in March.

Three readings carry the playbook.

The fall shoulder owns October. furnace tune up at 33,100 in October 2025 is not merely high for the shoulder — it is 1.8× its own January value of 18,100. Heating maintenance demand arrives before heating repair demand, which means the campaign that captures it has to be live in September.

The spring shoulder has no equivalent. ac tune up reads 14,800 in March and April 2026 against 27,100 in July — 55% of its own peak. Cooling maintenance demand arrives with the heat, not ahead of it. Any playbook that promises a symmetrical spring version of the October furnace campaign is selling a curve that is not in the data.

The flat lines are the hedge. hvac maintenance never leaves the 33,100-74,000 band and duct cleaning never leaves 27,100-49,500. Those are the two products whose audience is present in every month of the year.

Four caveats bound all of it, and they matter more than the numbers:

1. The only cross-check we hold does not cover these keywords. A three-tool reconciliation run on August 13, 2026 covered nine contractor-facing terms from this site's own set — hvac seo, hvac seo services, hvac seo company, hvac marketing agency, hvac marketing, hvac digital marketing, hvac website design, hvac answering service and hvac lead generation — and on those nine DataForSEO understated volume by roughly 2-4× against Ahrefs. Not one consumer term in the table above has been checked against a second tool, so that ratio does not transfer and nothing here should be read as a floor on the strength of it. Read the shape of each column rather than its absolute level, and never multiply a figure to "correct" it.
2. Values are rounded buckets (…12,100 / 14,800 / 18,100 / 22,200 / 27,100…), so small month-over-month moves may be quantisation rather than movement.
3. Two vendor endpoints disagree about the heating inversion. A Google Trends pull for ac repair, furnace repair, hvac maintenance and ac installation, run on the same day through the same vendor, returned 12-month average relative-interest values of 51, 28, 9 and 16 — and in its weekly series ac repair sits above furnace repair in all 52 complete weeks, never inverting. The Google Ads monthly series does invert. Trends is a 0-100 relative index scaled across four terms in one query, a different metric that cannot be converted to volume, so the two are not arithmetically comparable — but the disagreement is the point: do not build a heating calendar from a single relative-interest chart. The monthly volume series is the one that matches the degree-day mechanism.
4. This is one national snapshot. National curves cannot locate your shoulder window. Your own Search Console impressions by month and your own dispatch board can, and they are the only versions of this table that should drive spend.

The full 150-keyword pull — CPC, difficulty and intent for every term quoted here, plus the nine-term three-tool reconciliation behind caveat one — is published as the HVAC keyword dataset behind these curves.

So what does a contractor actually do with a calendar this uneven?

• The twelve-month plan this calendar plugs into — targets, budget and the annual planning sequence that sits above the seasonal layer.
• How the spend split shifts between peak and shoulder months — the published percentage-of-revenue benchmarks and channel allocations, with the front-loading guidance.
• Automating the reactivation sequence a spring campaign depends on — the email, SMS and follow-up mechanics behind a database-driven shoulder month.