Marketing for home services in Australia

Every home-service business carries a picture of its customer: the homeowner with water coming through the ceiling, ringing whoever answers first. SoudCoh's research sample of Australian trade and home-service advertisers tells a quieter and more useful story. Most demand is ordinary work. Most of it arrives on a phone. A real share of it arrives when the office is shut. And the trade — more than the device or the time of day — decides whether it comes in as a call or a form.

Throughout, an enquiry means a recorded phone call or a submitted form, never a tap on a phone number.

Mobile is where the work is

For the typical advertiser in the sample, 76% of enquiries come from mobile clicks, from a near-identical share of the spend. Mobile enquiries cost a little more — a premium of +9% per enquiry against desktop for the same advertiser — but a blanket mobile bid-down trades a small saving per enquiry for a large loss of volume. Fix the mobile page first, and adjust bids by device only where your own data shows a large gap.

Mobile searchers also behave differently once they land. The share of enquiries that are phone calls reads 58% vs 41%, mobile against desktop, for the median advertiser: desktop searchers are more likely to write, mobile searchers to ring. But the device split hides a bigger one.

The trade decides the door

Plumbing and electrical enquiries arrive mostly as calls on any device; cleaning enquiries arrive mostly as forms, even from phones. The reason is the job. A leak or an electrical fault is urgent and hard to describe, so people ring. A bond clean or a carpet clean is planned and easy to describe, so people write.

So design the contact route around the job, not the screen. Trades that live on calls need a visible number, call tracking and someone answering. Trades that live on forms need a short form, an instant acknowledgement and a fast reply. Every site still needs both doors, because every trade gets some of each.

Demand does not keep office hours

About one in four calls from ads — 26% for the typical advertiser — arrive before 7am, after 6pm or on a weekend. Homeowners search when they are home. Weekends pay their way, too: for advertisers running ads all week, a weekend enquiry costs +4% against a weekday one.

No single day dominates the week's demand either: the busiest day carries 16% of the week's enquiries, and Saturday and Sunday carry a steady share.

Switching ads off at night and at weekends is a common 'saving' that the data does not support as a default. Whether those hours pay depends on one thing no bid can fix: whether someone answers the phone. Either answer them — a call-answering service costs less than the ads that produced the call — or schedule the ads to the hours you can actually pick up. See matching ad hours to the hours you answer.

'Near me' is a decision, not a location

'Near me' searches produced 17% of enquiries for the median trade advertiser, and they converted clicks into enquiries at 1.3× the rate of other searches. A customer who types 'near me' has usually decided to hire and is choosing who. The search itself says nothing about where they are, so it only works when location targeting reaches people physically in the service area — not people merely 'interested in' it — and when the landing page shows the service area and a phone number in the first screen. See why near-me searches convert better for trades.

What an enquiry costs

Cost per enquiry varies widely by trade. The table sets out the three trades where SoudCoh publishes a benchmark.

The spread matters more than any single median. Cleaning enquiries cost far less than plumbing or electrical ones and far more of them arrive as forms; plumbing and electrical enquiries cost more and mostly arrive as calls. A rise in cost per enquiry that stays inside your trade's range is a different conversation from one that leaves it. The Google Ads benchmarks for Australian trades set out the full ranges and definitions, and our trade pages go deeper for plumbing businesses, electricians, cleaning businesses, HVAC businesses, pest control, roofers, painters, removalists and builders.

The keyword list is where most businesses think their advertising lives: the phrases the business, or its agency, chose to bid on. In SoudCoh's research sample it is closer to a starting hypothesis. For the median trade advertiser, 59% of enquiries came from searches that were not on the keyword list — even allowing for plurals and word order — and those searches took 64% of search-term spend. The symptom ('toilet keeps running'), the suburb, the odd phrasing: a large part of the demand a trade campaign captures arrives in words the business never chose.

Discovered demand is not free: an enquiry from a search that was not on the keyword list cost 1.19× as much as one worded like an existing keyword. That modest premium is the price of learning. Harvest discovery rather than switching it off — promote repeat winners to their own keywords and keep the net open. The full analysis is in trade enquiries come from searches you never chose.

Half the enquiries live in the long tail

Trade demand is episodic: a burst pipe, a failed inspection, a lease ending. Much of it is phrased once and never again. For the median advertiser, 54% of enquiries came from searches that received no more than two clicks in total, and 33% came from searches that appeared in only one calendar month. Even from the third month of running onwards, 25% of each month's enquiries still came from searches that had not appeared in any earlier month.

No keyword tool would suggest most of these searches; only keywords with room to match can reach them. The campaign has to recognise intent, not memorise phrases.

'Exact' match is not exact

Many owners respond by asking for exact match only, to keep the account on a short lead. It no longer does that. Google now treats exact match as 'same meaning', and 47% of the enquiries that exact-match keywords produced came from searches worded differently from the keyword. That is mostly helpful — close variants catch misspellings, reorderings and plurals — but it means an exact-only account still needs a search-term review and a negative list. Exact match alone is not control. See exact match is not exact.

The report you rely on shows about half the picture

That last point matters more than most owners realise. Google's search-terms report accounts for only about 52% of the money a typical trade Search campaign spends; the rest goes on searches it never shows you. Enquiries are no more visible than spend: the report names the search behind roughly 51% of a typical advertiser's enquiries.

A negative-keyword list can only ever police the visible part. That is why control has to come from structure — tight service themes, sensible match types, location settings — and not from the search-terms report alone. Review the report weekly, but judge a campaign on enquiries and cost per enquiry, not on how tidy the visible terms look. More in the search-terms report hides half your spend.

Short and specific beats long and wordy

Longer searches are not always better searches. Three-word searches converted clicks into enquiries at 1.24× the rate of searches of five or more words. The searches that convert most often are three or four words long, often a service plus a place or a problem; one- and two-word searches convert slightly below average. Very long searches tend to carry questions and research — 'how much does it cost to…', 'why does my…' — and convert below average. Long-tail room is still worth having because it adds volume, but those searches need the tightest cleaning.

Segments worth reading separately

Some kinds of search behave differently enough to deserve their own line in every review, and several overturn defaults that rules-based cleaners still apply.

If most demand arrives in words nobody chose, structure matters more than the length of the keyword list. Its job is to make sure that whatever a customer types, the ad that answers is about their job, the page it opens says the same thing, and the account can tell you which services and areas are paying.

Size is not the point. What separates a good account from a poor one is whether the searches, ads and landing pages line up, and whether someone reads the search-terms report every week. Be wary of any pitch that sells 'more keywords' as the fix.

Service campaign first; suburbs as a satellite

Every account SoudCoh builds starts with a service campaign: one ad group per service, each with its own tight keyword set (a median of 18 keywords per ad group), three responsive search ads and a matching service page. Where suburb-named searches matter, a per-suburb location campaign runs beside it on a minority of the budget, with its own budget so the two can be compared. The Search campaigns page describes the setup.

That comparison is the point. Clicks on per-suburb campaigns carried a premium of +49% over service-campaign clicks from the same advertiser: a search naming a suburb signals intent to buy, so those auctions are more contested. Whether the extra intent pays for the dearer click is close to a coin toss: for 57% of advertisers running both, the suburb campaign produced fewer enquiries per dollar than the service campaign. Run both legs with separate budgets and let cost per enquiry decide once both have two months of data.

A small core does most of the work

A comprehensive build switches on hundreds or thousands of keywords, but a small core carries the account. The top tenth of keywords by spend carry 53% of all keyword spend for the typical trade advertiser, and only 29% of switched-on keywords have triggered even one impression. That is harmless — Google matches each search to the closest keyword, so near-duplicates sit silent — but it makes keyword count a vanity metric. Watch the core daily: bids, search terms, ad copy and landing page. Manage the long tail in bulk and prune it quarterly.

Match types: phrase is the workhorse

Phrase match carries 69% of keyword spend across SoudCoh's research sample, with exact match holding the proven core and broad match a small remainder. Phrase match keeps the service words fixed and lets the customer add the rest — the suburb, the urgency, the symptom — which is the balance most trade accounts need. SoudCoh's builds launch every keyword as an exact-and-phrase pair, with effectively no broad match at launch.

Exact match is not a cheaper way to buy a lead. For the same advertiser, phrase-match clicks cost 17% less than exact-match clicks and convert less often, and the two effects cancel out: a phrase-match conversion costs about the same as an exact-match one. Judge match types on cost per enquiry, never on cost per click. Broad match has a place too. Where it ran — a minority of the sample — its enquiries cost 1.09× as much as exact-match enquiries from the same advertisers. It is a tool for well-tracked, well-cleaned campaigns, not a default for an account nobody reviews.

Most lost impressions are lost to rank, not budget

The typical trade Search campaign in the sample appears in 21% of the searches it is eligible for, and for 92% of advertisers more impressions are lost to ad rank than to budget. Raising the budget does nothing for impressions lost to rank. Under automated bidding, 'lost to rank' also includes auctions the algorithm chose not to chase because they looked too expensive, so the levers are relevance and a realistic cost-per-enquiry target — not a bigger daily budget.

Relevance can be measured. The most common Quality Score drag is expected click-through rate: for the typical advertiser, 58% of scored keyword spend is rated 'below average' on it, with landing-page experience close behind. Expected click-through is judged against every other advertiser on the same search, so in trades the ad has to beat directory listings and big brands on the first line — price, speed and proof beat slogans. And 54% of scored keyword spend goes to keywords with a Quality Score of 4 or less. Some high-intent, high-competition searches never score well, but each one should be a deliberate choice, reviewed monthly.

Keep Search campaigns on search

One structural setting deserves its own check. For 21% of trade advertisers in the sample, campaigns labelled 'Search' were also serving ads on the Display Network. The 'Include Google Display Network' box is ticked by default when a Search campaign is created, and Google can shift delivery towards Display without any change appearing in the change history. Where Display ran inside Search campaigns it took 23% of their impressions for a sliver of the spend, and a Display click produced a real enquiry at about 4% of the rate of a Google Search click in the same campaigns.

Search partners fared better, but not well: cheaper clicks that converted at 16% of the Search rate for the median trade advertiser, so each enquiry cost more. 100% of the Search campaigns created through SoudCoh's campaign builder launch with Display and Search partners switched off, targeting people physically in the service area. More in the Display Network on Search campaigns.

The first three months of a new account follow a pattern that surprises most owners. In SoudCoh's research sample, the launch arc has three phases: early proof that demand exists, a honeymoon month, and a second month that costs more.

Before day one

Nothing SoudCoh builds goes live on the strength of a generator alone: 97% of the campaigns SoudCoh's builder creates go into the account paused, and are switched on only after a person has read the ad groups, keywords, copy and landing pages against the business — the cheapest point to catch a service it does not offer or a promise it cannot keep.

Days one to seven: the phone should ring

Search captures demand that already exists, so enquiries should arrive almost immediately. 84% of new launches in the sample recorded an enquiry from their first three days of ads, and nearly all did within the first week. A silent first week is far more often a broken call-tracking number, a form that does not fire or a phone nobody answers than a bad keyword. Check those before touching bids.

The honeymoon, then the rise

A new account's second month is usually dearer than its first. The typical cost per enquiry in month two ran +31% against month one, and 68% of new launches paid more per enquiry in their second month. A fresh account tends to serve the most urgent, ready-to-book searches first, before the bidding widens its reach into harder searches. Month two is the real baseline.

Cleaning pays most at the start

The first months are also when search-term cleaning matters most. Control compounds: each negative removes a pattern for good, so the junk that reaches the account shrinks as the list matures and the review gets faster. Two months into daily cleaning, the share of reviewed spend that needed negating had fallen for the median advertiser (−43%). After that, cleaning becomes maintenance.

If discovery brings in most enquiries, control stops it bringing in junk. The two are usually presented as opposites — tight accounts against loose ones — but in SoudCoh's analysis they work as a loop: launch with control, let the account discover, fence what it finds, promote what works, and go round again. The companion paper, Launch, discover, control, sets out the method in full.

In a well-run account, waste is a leak, not a flood

In SoudCoh's daily search-term reviews, only about 4.4% of reviewed spend was clear-cut waste to negate for the median trade advertiser — 1.5% in cleaning and 4.5% across other trades and home services. The leak differs by trade, because the words overlap differently with jobs, DIY and retail searches, but it is a leak, not the flood that 'wasted spend' pitches describe. The big gains in trade search come from structure, bids and landing pages; cleaning protects those gains by stopping the leak from growing. An agency that promises large savings from cutting wasted spend alone is describing an account nobody has looked after.

The expensive junk is not where most template lists look. The biggest single source of waste — 44% of all the spend SoudCoh's reviews flagged — was searches for a different service, or for something unrelated that happens to share a word: a neighbouring job the business does not do, which looks relevant enough to click. Job-seeker and DIY searches, the usual targets of template negative lists, were only 12% of flagged spend. Template lists (jobs, courses, DIY tutorials, free downloads; never free on its own, which would block free-quote searches) are cheap insurance worth adding on day one, but the expensive leaks are specific to each business and only show up in its own search terms.

Old negatives block new customers

Negative lists only grow unless someone prunes them. 52% of trade advertisers in the sample now block, through their own negative keywords, at least one search that produced an enquiry over the preceding year. The typical cost is small — blocked searches had produced 0.4% of the year's search-term enquiries for the median advertiser — but it is concentrated: in the worst case in the sample, searches that a business's own negative list now blocks had produced 6% of its paid-search enquiries over the year. It is the cheapest fix in paid search, because the enquiries were already proven. See negative keywords that block customers.

Rules alone make this worse. 1.5% of the searches a keyword-rule cleaner flagged for negation had converted in the very window it was judging, and 2% of searches that simple waste rules flagged had produced a genuine enquiry at some point in the year. That is why every automated negative in SoudCoh's process goes through a person and a simulation against a long window of converting searches.

'Relevant but not converting' is a question, not a verdict

About 75% of the spend a daily search-term review flags is on searches that look relevant but have not converted yet — a judgement call, not junk. Negating them on a cost threshold would shrink the account. 7% of the relevant-but-unconverted searches held back from negation went on to produce an enquiry in a later month, and 53% never triggered an ad again, so adding them as negatives would only have lengthened the list. The better moves are a bid change, a sharper ad, a better landing page, and more time. Negate only when a pattern repeats with meaningful spend and no enquiries over a long window, or when the owner confirms the job is not wanted.

Competitor names: segment, do not ban

Searches naming another trade business took only 0.6% of search-term spend for the median advertiser, so blocking them all saves very little. And for 72% of the advertisers that paid for competitor-name clicks, those searches produced at least one real enquiry: people search a known name, find it busy or out of area, and call the next credible option. Negate a rival's name only after a long window of spend with no enquiries, using the full brand phrase — or give competitor traffic its own low-bid campaign and judge it on its own numbers.

Promote the winners

The other half of control is promotion. Only 11% of the searches that produced an enquiry through phrase or broad match were ever added as keywords in their own right, and even among searches that produced three or more enquiries, only 33% had been promoted. A search that is merely caught cannot be bid up, written for or protected from a stray negative. A monthly promotion pass — repeat converters moved to exact match in the right ad group, with an ad that repeats the customer's words — is cheap and compounds.

Once an account is cleaned regularly, the remaining junk is a rounding error: after every safety check, the negatives that could safely be added covered just 0.6% of monthly spend for the median advertiser. The typical trade advertiser carries about one negative for every five keywords (a ratio of 0.2); a list growing much faster is a sign the structure is too loose. From there, the money is in cost per click, conversion rate and budget allocation.

Every decision in this paper — which searches to keep, which services to fund, whether month two is working — rests on the Conversions column. For many trade businesses, that column is not counting what the owner thinks it is. The problems are ordinary, several are Google's defaults, and each can be checked in minutes.

Taps are not enquiries

For a third of the sample (32%), taps on a phone number, email or WhatsApp link were being counted as conversions in Google Ads. A tap shows someone meant to call, not that they did. Where taps were counted, they made up 45% of everything Google Ads reported as a conversion for the median advertiser, and the Conversions column overstated real enquiries — connected calls and submitted forms — by +92%. A campaign reporting nearly twice the enquiries it produced shows a cost per enquiry at roughly half the truth, and Smart Bidding chases the people most likely to tap.

The fix costs nothing: move taps and profile actions to secondary, and let bidding optimise to calls and forms only. We explain how in phone-tap conversions inflate trade leads.

Two call routes, counted once per caller

Calls arrive two ways: straight from the ad, through Google's call reporting, and after someone has read the website, through a tracked number swapped onto the site for paid visitors. Both are enquiries, but they are measured by different machinery, and most businesses in the sample did not capture both — only 32% did. An account that captures only ad calls undercounts what its strongest pages produce, and steers budget away from them.

Then count people, not rings. About one call in five to a tracked number — 19% — comes from someone who has already called: chasing a quote, confirming a time, calling back after a missed call. Those calls matter to the business but they are not new enquiries. Google's 'one' counting option stops a single ad click being counted twice, but a caller who rings back later can still count again, so apply a unique-caller rule wherever your call tracking allows it, or cost per enquiry looks better than it is.

The 60-second line hides short, decisive calls

About half of answered calls from ads — 49% — last less than a minute, below the 60-second line Google uses by default to decide whether a call counts. Short calls are a mix: wrong numbers and hang-ups, but also the quick 'can you come Thursday?' that books a job. A fixed 60-second rule hides the second group from bidding and keeps the first out of reports; neither is ideal. Listen to a sample of short calls before deciding where your threshold should sit, and revisit it by service — a blocked-drain call is often short and decisive. See short ad calls and the 60-second rule.

The one-dollar default makes ROAS meaningless

77% of trade advertisers in the sample leave Google's default value of one dollar on every enquiry, so the 'conversion value' and ROAS columns simply count enquiries in dollars. That is harmless until someone reads the ROAS column, or switches to a value-based bid strategy — at which point Google optimises for the number of enquiries while everyone believes it is optimising for revenue. Only 9% told Google Ads what a job was actually worth, and none did it through website or call tags: the only real values were won jobs uploaded after the fact.

A website cannot know what a job is worth when the form is sent; the price is set on site, days later. Real value has to come back from the job book — until then, treat every ROAS figure as a placeholder. More in the default conversion value makes ROAS meaningless.

One lead record, with its source attached

SoudCoh's method is to write every form, call and booking to one lead record, with its source captured on arrival — click ids, landing page, device and, once matched, campaign, ad group, keyword and search term — alongside the existing Google tags rather than instead of them. Won-job values then go back to Google Ads as an offline conversion, secondary at first, so they inform reporting before they change bidding. The tracking page describes the setup, and how to tell whether your leads are tracked properly walks through the checks.

A trade business does not get paid for enquiries. It gets paid for jobs, and jobs vary enormously in value — which is why cost per lead, the most common way of judging trade advertising, is a poor guide on its own.

A handful of jobs carry the month

The biggest tenth of won jobs carried 41% of recorded revenue for the median business. A single large job — a switchboard upgrade, a full re-pipe, a leak that became a bathroom — can swing a month's return on ad spend more than any campaign change. Judge campaigns on rolling quarters, look at the median job as well as the total, and keep a separate eye on the services that produce the big tickets, because they are rare enough to be cut by mistake.

Judge each lead by the job it wins

Cost per lead treats a blocked-drain call and a hot-water replacement as equal. They are not. Give planned services — installations, replacements, hot water, gas and compliance — their own campaigns and budgets so a steady stream of small call-outs cannot starve them. Reply to forms as fast as you answer the phone, and judge calls and forms by the jobs they win, not by how urgent they sound. For how that plays out across channels, see lead generation for service businesses.

Most trade enquiries start on a phone. The website's job is narrow and demanding: prove in the first screen that you do this exact job, in this area, and make it easy to ring or to write.

A page for every service

SoudCoh sends ads to service pages rather than the home page. Service pages make ads more relevant, lift Quality Score, and give Google and AI assistants something specific to cite. That matters because landing-page experience is the second most common Quality Score drag, rated below average on 52% of scored keyword spend for the typical advertiser. What converts is what is on the page: the job, the area, proof and an easy way to get in touch.

Suburb pages follow the same rule: build them only for the areas you want, and make them real — local jobs, local photos, travel times, not a template with the suburb name swapped.

What the first screen needs

Across the trade websites SoudCoh builds, three components recur: an explanation of what happens after someone enquires, clear suburb coverage, and a page for each service. They answer the questions a homeowner has before calling — do you do this exact job, do you come to my suburb, and what happens after I press send? None needs a big budget; they need specific, accurate content. The pages that convert put the phone number and a short form in the first screen, name the suburb and the job, and show proof: licences, reviews, photos of real work.

Build the contact route around the job. For call-led trades — plumbing, electrical, most urgent work — the number must be tappable, tracked and answered, including outside office hours. For form-led trades such as cleaning, the form must fit on one phone screen, acknowledge instantly and promise a reply time you keep. Every page still needs both doors. See the website guides for electricians, plumbers and cleaning businesses, and a trade website that converts calls and forms.

Some of the questions buyers once typed into Google now go to ChatGPT-style assistants. SoudCoh tests how those assistants choose their sources, using buyer questions about marketing for trades.

Assistants cite the page built for the question

When ChatGPT-style assistants answered 'who does marketing or websites for [trade]', they mostly cited agency pages built for that exact trade — 60% of the pages cited in SoudCoh's tests — not agency home pages or directories. Asked about electricians, an assistant looks for a page that is about electricians, and a general 'we do digital marketing' home page rarely qualifies. The pages that get cited state plainly who they are for, what the work involves and what it costs. If you want to be recommended for a trade, publish a page for that trade.

An AI answer is a draw from a short list

Ask an AI assistant the same buyer question three times and only a minority of the websites it cites — 25% in SoudCoh's tests — appear in all three answers. There is no single 'AI ranking' to win. Being on the list — through a page that directly answers the question, a directory profile and third-party mentions — matters more than being first in any single answer. Measure AI visibility by asking many times, on more than one assistant, never once.

Analytics undercounts AI discovery

Do not expect analytics to show you this. On SoudCoh's own website, visits that analytics could attribute to AI assistants were 0.2% of sessions; most AI-influenced visits arrive with no referrer and are filed as 'Direct'. When someone copies a name from an AI answer into a browser, or taps a link inside an app that drops the referrer, nothing travels with the visit to say where it came from; some assistants tag their links, which a dedicated channel can catch. Ask every new enquiry how they found you, store the answer on the lead, and give AI assistants their own channel in analytics. See how ChatGPT chooses which business to recommend.

Google still gets the short search

Two in five non-brand Google searches that showed SoudCoh's site named a place (41%), while phrased questions barely registered. Questions are moving to AI assistants; Google still gets the short commercial search. So build both — service-plus-place pages for the short search and plain-English answers for the questions — with local SEO underneath.

The findings in this paper apply at every size, but a business meets them in a different order as it grows: an owner-operator is short of time, a growing company of measurement, a multi-site brand of consistency.

On budget

We recommend a minimum of A$5,000 a month in ad spend for most businesses; the largest companies we work with invest up to A$30,000 a day. The reason for a floor is learning: an account needs enough enquiries each month for search terms, match types and services to be compared, and at low volumes a single large job — or a single quiet week — can make a month look like a trend. Whatever the level, spread the budget across all seven days, and check whether impressions are being lost to rank before adding money; for most advertisers, rank is the tighter limit.

Owner-operators

For a business where the owner is also the one on the tools, the most valuable changes cost nothing but attention. Make sure the Conversions column counts real enquiries. Keep the ads to the hours someone can answer, or put a call-answering service behind the number. Build one tight service campaign with a page for each service, and read the search terms once a week. Resist the pitch to add hundreds of keywords.

Growing trade companies

As crews are added, the bottleneck moves from finding work to knowing which work pays. This is the stage to capture both call routes, count unique callers, answer forms as fast as calls, and start the monthly spreadsheet that connects won jobs to ad groups. It is also the stage to test a per-suburb satellite campaign against the service campaign with separate budgets, and to give planned services — installations, replacements, compliance — their own campaigns so reactive call-outs do not starve them.

Multi-site service brands

At scale, small leaks compound across sites. A shared negative list can silently remove a service everywhere at once, so it needs a quarterly audit against converting searches and a record of why each negative exists. Won-job values should flow back into Google Ads as offline conversions — secondary at first — so budget can move between services and regions on revenue per ad dollar rather than cost per lead. And because a handful of large jobs carry a large share of revenue, judge regions and services on rolling quarters and on the median job, not on a single month's return — with one definition of an enquiry across every site.

Each point below comes from a finding in this paper. The first four take an afternoon and cost nothing; the rest are standing habits. None requires an agency — though if you would rather someone else ran them, talk to SoudCoh.