Google Ads for plumbing companies in the UK

A Leeds owner opens Google Ads on a wet Tuesday and sees spend of £84 before lunch. The search terms look busy. The diary does not. Two of the three “leads” are homeowners asking how to bleed a radiator themselves. The third is a call that rang 11 seconds, went to voicemail, and never came back. That is the real UK plumbing Ads problem: you are not short of clicks. You are short of jobs that cover a van, a Gas Safe engineer, and a same-day slot.

Google Ads still average about 4.2x ROAS across industries, and Search CPC has climbed (the 2025 WordStream UK Trades benchmark puts Search CPC at around £2.40, up roughly 10% year on year). Trades do not live in that average. Emergency drain, boiler, and leak queries in London, Manchester, and Birmingham routinely bid against national aggregators, comparison sites, and “near me” bait. Optmyzr’s 2024 PPC Waste Report found ~22% of spend goes to low-intent queries in service-sector accounts - irrelevant queries, neglected negatives, or a homepage that does not match the ad. Mobile takes 63%+ of paid search clicks, then conversion lags desktop by 30-40% because the page is slow, the click-to-call is buried, and smart bidding never saw the booked job.

This piece is for marketing owners inside UK plumbing firms, not for DIY householders. It covers how Search, call tracking, and landing pages have to work as one system if you want truck rolls instead of quote tourism.

Why UK plumbing Search looks profitable until the van calendar tells the truth

Last-click in Google Ads will happily count a “conversion” when someone taps Call now, hangs up, and Googles a cheaper out-of-hours number. Automated bidding then treats that tap as success and buys more of it. Plenty of firms still run one campaign called Plumbing, one ad group, and a keyword list that mixes “emergency plumber Leeds” with “how to unblock a toilet with washing up liquid”. Smart Bidding cannot tell a £180 callout from a £0 tutorial click if both fire the same conversion tag.

You also lose 20%+ of conversion data to ad blockers and privacy changes. If you only count a thank-you page and never pass the booked job (or at least a qualified call over a duration threshold) back into Enhanced Conversions, the algorithm optimises on noise. High CPC pressure then tempts you to switch a new account to manual CPC “to stay in control”. That usually freezes learning just as you needed more signal, not less.

Performance Max will drive bulk volume if you let it. For plumbing, that bulk often includes bathroom-fit shopping, parts, and national lead-gen partners. Google’s own case studies cite 4x+ ROAS for service businesses with strong first-party data - but those are best-in-class accounts, not a baseline to plan against. In a typical independent firm with a slow homepage and no call duration import, it just spends. If someone tells you they will “just send traffic to the homepage,” treat that as a warning, not a plan.

A Birmingham multi-van shop we modelled last winter had a 3.1% form conversion rate on branded terms and a 0.4% rate on “blocked drain near me”. The blocked-drain ads were cheaper per click. They were also cheaper because they were attracting people who wanted a YouTube fix, not a van. Cost per booked job on the emergency cluster was more than 4x the branded cluster once you ignored form fills under 30 seconds of talk time. The fix was not a new platform. It was cleaner signals and a landing page that said the same thing as the ad: same-day, Gas Safe, call now, no DIY tips in the hero.

If paid search is leaking after the click, a short look at on-site behaviour with HeyLead Insights usually shows the miss: thumb-zone call buttons that sit below a 1,200-word blog, forms asking for a postcode plus a photo plus a preferred time before anyone will ring, or a 4-second INP on mobile so the tap never registers as a conversion.

How to structure Google Ads for plumbers around job value and call quality

Do not structure by “brand vs generic” as if you were a SaaS company. Structure by economics. A burst pipe at 9 p.m. in Zone 2 is not the same product as a quote for a new bathroom in Guildford. Separate campaigns (or at least tightly themed ad groups with their own budgets and conversion actions) for emergency callouts, planned installs (boilers, cylinders, unvented), and commercial contracts if you actually take them. Match RSA headlines to the job: “Blocked drain, engineer today”, not “Family plumbers since 1998”.

Location targeting should follow where you will actually send a van in under an hour, not the whole of Greater London because the map looks impressive. Use presence, not “interest in”. Exclude postcodes you routinely decline. If you cover two cities, two campaigns beat one blended CPA that hides Manchester winning and Liverpool losing.

Run Google Local Services Ads alongside Search. LSAs sit above standard Search ads for local trades in the UK, charge per lead rather than per click, and are verified against Gas Safe status for plumbing and heating. Their presence in the SERP changes bid strategy for branded and emergency terms: you are competing for click-share with a Google Guaranteed unit as well as aggregators, so branded and high-intent emergency campaigns need tighter match, cleaner landing pages, and conversion actions that still make Search worth the remaining auctions.

Conversion setup is the campaign. Count a call only after a duration you know is a real conversation (many UK trades use 45-90 seconds). Import offline conversions when a job is booked or invoiced, even if that is a weekly CSV at first. Turn on Enhanced Conversions with the unified toggle so first-party data offsets the 20%+ you otherwise lose. Do not let every form, every click-to-call, and every “get a quote” button share one conversion action called Lead. Smart Bidding will then buy the cheapest lead, which is rarely the best job.

Search terms still matter even though the industry talk is “signals over keywords”. Google still needs seed keywords and a ruthless negative list. Weekly, dump search terms. Negate DIY (“how to”, “fix myself”, “washer”, “O-ring”, “YouTube”), jobs you do not do (heat pumps if you do not install them, oil boilers, power flushing if it is not on the van), and competitor brand terms you cannot win profitably. Keep that hygiene boring and consistent. AI Max can deliver 12-18% lower CPCs versus standard Search in some accounts. It also expands match. Without negatives, cheaper clicks are just cheaper waste.

RSA copy should name the constraint the customer actually has: Gas Safe, no call-out fee if you truly waive it, 60-minute arrival windows you can keep, evening slots, parts on the van. Do not promise “from £49” if the real ticket is £180 plus parts. Misleading RSAs inflate CTR and destroy close rate. Extensions: call, callouts, location, structured snippets for services, and sitelinks to dedicated pages (blocked drains, boilers, leaks), never to the blog.

Emergency plumber landing pages: what converts at 2 a.m.

UK emergency traffic is thumb-first, often on 4G, often standing in a flooded kitchen. Your page has one job: get a qualified call or a short form with a phone number. Hero: service + city + time promise you can keep. Primary CTA: tap to call. Secondary: “text photos” only if someone actually replies after hours. Reviews with job type and area (“burst pipe, Croydon, 40 minutes”) beat a generic 4.9 stars widget.

Proof that moves a plumber’s buyer: Gas Safe number visible, insurance, “parts on the van”, photos of real jobs (not stock bathrooms), and a plain English price band if you use one. Do not hide VAT. Do not make them complete a 9-field form to request a call back on a leak. Form friction is how paid search “converts” at 0.4% while the phone rings for competitors.

Match the query. “Boiler not firing, Worcester” should not land on a general plumbing homepage with a 6MB gallery. Dedicated URLs let you write RSA that matches, track CPA by job, and keep Quality Score from collapsing. Use UTM discipline so GA4, the call tracker, and the job sheet can be reconciled. Last-click will still double-count if you also run Meta; treat in-platform CPA as directional and reconcile to booked jobs weekly, not to a dashboard screenshot.

If Google Ads management is already live and CPA is drifting, resist rebuilding the whole account in one afternoon. Constant structure changes reset learning. Change conversion quality, landing alignment, and negatives first. Then give bidding two to three weeks of clean data. Realistic results on a messy UK trades account take 3-6 months, not a promised CPL on day one. Nobody can guarantee a specific ROAS before they have seen your account, pages, and job mix.

A Manchester two-van firm that stopped buying radiator tutorials

The marketing lead at a two-van firm in Greater Manchester inherited an account that spent about £2,400 a month. CPA in Google looked acceptable at £38 per “lead” - a £41 CPA that looked fine until they filtered calls under a minute. The office manager said half the calls were “can you just tell me which valve”. Search terms were full of how-to queries. The landing page was the homepage, with a 3.8s LCP on Android and a call button below three paragraphs of company history.

They did not switch to manual bidding. They split emergency drain and boiler into their own campaigns, imported calls over 60 seconds, and built two short pages with tap-to-call and a Gas Safe badge in the first viewport. Negatives killed the tutorial cluster in one sitting. After 11 days, recorded conversions dropped. Booked jobs did not. Cost per dispatched job fell because bidding finally had a conversion worth chasing. They left PMax off until Search terms were boring.

That is the pattern. Volume is a dead metric if the van is idle. Quality leads, not more leads, is the language owners already use when they are tired of dashboards. Tighter positioning, a clearer same-day offer, and decent attribution beat adding Bing, Meta, and a third call-tracking vendor in the same week.

When the funnel is this operational, a specialist partner is often cheaper than a junior in-house two days a week who resets the account every time learning looks “stuck”. If you want a second pair of eyes on structure before you raise spend, a free audit is enough to see whether the leak is queries, the page, or the conversion action.

FAQs

Should a UK plumbing company use Performance Max or Search first?

Search first, with job-type campaigns, call-quality conversions, and negatives. Add PMax once you can see which queries become invoices. Google’s own case studies cite 4x+ ROAS for service businesses with strong first-party data - but those are best-in-class accounts, not a baseline to plan against. It will not fix a homepage and a single “Lead” tag.

Is automated bidding just inflating spend?

It inflates spend when the conversion is junk. It behaves when the conversion is a duration-qualified call or a booked job. Manual CPC on a new account often feels safer and usually starves the model of the signal it needs. Fix the event, then let bidding run with fewer, broader campaigns rather than daily bid tweaks that reset learning.

What CPC should we expect in London versus the North?

There is no honest universal number. Emergency terms in inner London fight national aggregators and will sit well above any published cross-industry average. Judge cost per booked job after parts and labour, not headline CPC.

Do we need a dedicated landing page for every service?

You need one for every cluster that has different economics and different RSA promises: drains, boilers, leaks, and commercial if you sell it. Four tight pages beat 40 thin ones. Align ad copy to the H1. Check titles, meta, and snippets so the SERP does not promise a price the page then hides.

How long before Google Ads for plumbing companies in the UK looks stable?

Plan on 3-6 months of clean tracking, negatives, and page work, not a guaranteed CPL in week one. Anyone who promises a number before seeing the account, the diary, and the pages is selling comfort, not media.

Putting it to work

This week

• Pull 30 days of search terms and mark DIY versus dispatchable jobs.

• Stop bidding on conversion actions that fire on sub-minute calls.

• Point emergency RSAs at one mobile-first page and test tap-to-call in the first screen.

• Run Core Web Vitals and UTMs on that URL so the next 14 days of data is usable.

• Write five negatives you should have added months ago (how to, DIY, YouTube, parts-only, out-of-area towns).

Next 30 days

• Split emergency versus install budgets and reconcile Google conversions to the job sheet weekly.

• Turn on Enhanced Conversions and, if you can, offline import for booked jobs.

• Only then test a contained PMax asset group, watching branded overlap.

• Review after-hours bid schedules against who actually answers the phone.

Start by matching last month’s Google “leads” to vans that actually rolled, then fix the conversion action that taught bidding the wrong lesson. HeyLead can take that UK plumbing Search loop (query hygiene, call-quality bidding, and the emergency page handoff) off your plate so you are not rebuilding the account between callouts. Chat with us on WhatsApp