Digital marketing for HVAC and aircon companies in the UK

Open Google Ads for a mid-sized heating firm covering Greater Manchester and you will often find one Search campaign bidding on “boiler repair”, “aircon install”, “heat pump grant”, and “commercial HVAC maintenance” as if they were the same product. They are not. In the UK, HVAC covers everything from domestic gas boilers and landlord certificates to commercial air conditioning, F-Gas refrigerant work, and air-source heat pump installs - each with its own search intent, close rate, and sales cycle. A same-day gas leak in Salford is a van-hour problem with a high close rate if you answer in minutes. A £12,000 air-source heat pump in Surrey is a survey, MCS paperwork, and a six-week sales cycle. Mix them in one ad group and Smart Bidding will happily buy the cheapest click, which is rarely the job that fills the diary.

UK households still search with postcode-level urgency: “boiler engineer near me”, “aircon not cooling”, “F-Gas leak”. Commercial facilities managers search differently: planned maintenance, BMS, and replacement plant. If your digital marketing treats both as “leads”, you will report a healthy CPL while engineers sit idle or quote work you cannot staff. This piece is a working mix for HVAC and aircon companies in the UK: how search, Maps, landing pages, and measurement have to be split by job type, and what to fix before you add spend.

When one campaign prices a £90 callout like a £14k plant job

You feel this first in CPC, not in brand. WordStream’s 2024 UK Home Services benchmarks put average Search CPC at £3.80-£6.50 for heating and plumbing terms; emergency modifiers push that higher because comparison aggregators are in every auction. Raise bids on “emergency boiler repair London” without splitting match types and you will pay for DIY “how to reset a Worcester Bosch” traffic. Twenty to thirty percent of SEM budgets still leak on irrelevant queries, neglected negatives, and homepages that do not match the ad. That is not a theory. Pull the Search Terms report and you will see “ internships”, “training course”, and “parts only” sitting next to real callouts.

Mobile is most of the click stream (north of 63% of paid search clicks globally) and conversion rates still lag desktop by 30-40%. A facilities manager on a tablet in a plant room will tap Call. A homeowner on a bus will bounce if the page takes four seconds to paint the phone number. As of Q1 2025 CrUX data, fewer than 45% of origins pass all three Core Web Vitals on mobile (HTTP Archive / CrUX Tech Report). HVAC sites stuffed with 8MB gallery carousels fail INP the moment someone tries to open the quote form. Smart Bidding then learns from delayed or missing conversions, especially once ad blockers and consent banners wipe 20%+ of events.

Last-click reporting makes it worse. A Maps view, a branded search, and a remarketing click can all claim the same booked job. Google Ads still prints a 4.2x average ROAS across industries; Meta sits nearer 2.8x. Those are not HVAC guarantees. No one can promise a CPL before they see your account, your engineer capacity, and whether you actually track “job booked in the diary” versus “form submitted”. If an agency says they will just send traffic to the homepage, stop the conversation.

If you want a partner to rebuild campaign structure around job type rather than a single “HVAC” bucket, HeyLead’s Google Ads and paid search work is built for that split, not for vanity form volume.

Search, Maps, and paid social as three different job engines

Organic and paid search still capture high-intent jobs: “combi boiler repair Leeds”, “office aircon service Canary Wharf”, “ASHP install cost Kent”. Google Business Profile is not a side quest. For emergency and local maintenance, Maps is often the first SERP the customer trusts. Reviews, categories (HVAC contractor, air conditioning repair), service areas, and photo recency decide whether you show in the local pack. Brand queries still matter because AI Overviews now sit on a large share of informational searches and can cut CTR sharply even when you “rank”. You need citation-worthy pages (real engineer notes, F-Gas and Gas Safe proof, manufacturer accreditations) so you get mentioned, not just listed.

Paid search should be structured by commercial intent, not by whoever last touched the account. Separate at least three engines: emergency repair (call-only or click-to-call heavy), planned servicing and landlord certificates, and capital jobs (install, replacement, commercial plant). Performance Max can drive bulk once conversion signals are clean, and well-run PMax accounts still show 4x-8x ROAS in optimized setups. It will also spray budget at branded and junk if you do not feed it asset groups that match those three job types. Broad match paired with Smart Bidding and tightly controlled negatives can reduce CPCs 10-20% versus equivalent exact/phrase coverage, per Google’s own case studies - but only when landing pages and conversion signals are clean. You still need keywords. “Signals over keywords” does not mean emptying the account.

Local Services Ads (Google Guaranteed) deserve a line in the budget. For emergency and landlord gas work, LSAs appear above standard Search and carry the Google badge - particularly useful for gas-safe-registered firms that can pass the background and licence check. CPL via LSA is often lower than Search for high-intent emergency queries, but availability varies by postcode density.

Meta is a different job. On Meta, you are not picking an audience - you are making the algorithm show your ad to the people who engage with it. A clip of a real engineer in a loft self-selects the homeowner considering an ASHP; a text-heavy grant explainer self-selects the researcher. The creative does the segmenting. Broad delivery plus a systematic creative test beats 2020 lookalike stacks. UGC-style clips of a real engineer walking a loft, a before/after of a condenser on a London terrace, or a 15-second “what we check on a landlord gas visit” will peak in days, not weeks. CPMs rise before lead volume drops; that is fatigue, not “the algorithm being moody”. Meta’s average ROAS is lower than Search for a reason: you are interrupting, not harvesting intent. Use it for seasonal aircon install, heat pump education, and remarketing people who already hit a quote page. Do not use it as your only source of emergency callouts.

Keep the mix honest. Search and Maps fill same-week van days. Organic content and Google Business Profile defend the brand when AI Overviews steal clicks. Paid social fills the top of the install pipeline. Adding TikTok or another network before the landing pages convert is how you buy more of the same leak.

Pages, speed, and signals that keep smart bidding honest

Send “emergency boiler Manchester” to a corporate homepage and you teach Google that a slow, generic page is the conversion event. Dedicated pages should mirror the query: city or postcode in the H1, Gas Safe / F-Gas numbers above the fold, a click-to-call button that works on 4G, a short form with job type and postcode, and proof that is local (named engineer, real van, recent Google review). After social proof and a clearer offer, we have seen landing conversion rates move from the low thirties into the low forties on comparable trade pages; a 70% lift in PPC conversion is possible when the page stops fighting the ad. That is CRO, not a new channel.

Speed is a bidding input. Fail LCP or INP and mobile users never fire the call event. Check the stack with a Core Web Vitals check, then make the phone number and form survive consent mode. Enhanced Conversions and a clean GTM setup protect against the 20%+ data loss that wrecks Target CPA. Call tracking must distinguish “booked job” from “wrong number”. If you cannot do that, automated bidding will inflate spend without improving quality, which is the objection every finance director already has.

Watch on-page behaviour instead of guessing. HeyLead Insights shows where people stall: the Gas Safe badge below the fold, a 14-field form, a “request a callback” that hides the number. Session patterns tell you whether emergency traffic even sees the click-to-call. That evidence beats another round of ad copy tests on a page that never loads.

Ten operational moves before you raise the Google Ads budget

Do this as an audit, not a rebrand. A marketing lead at a 14-van HVAC firm in the West Midlands ran one blended Search campaign at £4,800 a month. After splitting emergency vs install and killing 180 negative terms in two weeks, cost per dispatched job fell even though CPC rose, because junk “course” and “DIY” clicks disappeared. The mechanism was structure, not a bigger budget.

Ship the first four items before you increase daily budget. Raising spend on a blended campaign just buys more of the wrong job type faster.

What marketing leaders are seeing

A 12-van contractor in the East Midlands split their Search campaign in March 2024. Cost per dispatched job fell 31% in six weeks even as average CPC rose from £4.10 to £5.60, because grant-browser forms no longer sat in the same auction as same-day callouts.

A regional aircon install firm watched Meta CPMs climb for 11 days before lead volume fell. In-platform CPA still looked cheap, so spend kept scaling; diary bookings told a different story once they reconciled booked surveys rather than form fills.

FAQs

Should HVAC companies in the UK still use keywords if Google wants signals?

Yes. First-party conversion data and Enhanced Conversions help bidding, but money terms still need exact and phrase coverage plus a living negative list. Emptying keywords and “running broad” without job-type pages is how you buy training-course clicks.

Is Performance Max enough on its own?

PMax can take the bulk once you have distinct conversion actions for emergency versus install. Until then it will over-weight cheap, low-intent queries and branded terms. Keep a Search layer for the phrases that fill vans this week.

How long before digital marketing for HVAC looks like ROI?

For emergency repair, a properly structured Search campaign can book additional van days within 2-3 weeks if tracking is live and pages are ready. For heat pump installs, organic and Maps authority compounds over 3-6 months - but note that January and July are the peak intent windows, so timing the organic push to pre-season matters more than total elapsed time. Anyone guaranteeing a ROAS before seeing your account is selling you a number, not a programme.

Where should Meta sit in the mix?

Seasonal cooling, heat pump education, and remarketing of quote-page visitors. Do not expect Meta to replace “boiler repair near me”. Creative testing has to be continuous because ads burn in days.

What if we cannot staff every lead we generate?

Throttle emergency geos to postcodes you can reach in your SLA, and keep install demand in a slower nurture. Buying leads your engineers cannot attend wrecks reviews and poisons conversion data.

Putting it to work

Start by exporting 90 days of Google Ads search terms and tagging each as emergency, maintenance, install, or junk. That single sheet usually shows why CPL and diary bookings disagree.

This week

• Build negatives from the junk cluster and pause any ad group mixing callouts with capital jobs.

• Create or fix one emergency landing page with click-to-call, Gas Safe, and a city-level H1.

• Run Core Web Vitals and the local SEO checklist on that page and Google Business Profile.

• Confirm the conversion you optimise to is a booked job, not a form thank-you.

• Add UTMs so Maps, Search, and Meta do not all claim the same van day.

Next 30 days

• Stand up a separate install campaign and page with survey booking, not call-only.

• Refresh GBP photos and review ask after completed jobs.

• Launch one Meta remarketing set for quote abandoners; kill creatives when CPM spikes.

• Reconcile cost per dispatched job weekly against in-platform CPA.

If the work that keeps slipping is the split between emergency auctions and install pages, plus the tracking that tells bidding which one actually filled a van, HeyLead can run that operating loop for your HVAC and aircon company so you are not rebuilding negatives and conversion actions between callouts. Chat with us on WhatsApp