Google Ads for Local Service Businesses: Winning the Jobs Worth Having

The short answer

Most local service businesses treat their Google Ads account as the thing that decides whether advertising works. It rarely is. Two other things decide it, and both sit outside the account entirely: how fast an inquiry gets answered, and which jobs the business agreed to bid on in the first place.

Consider two businesses in the same trade and the same city. One has a carefully structured account, tight match types, a clean negative list, well-written ads. The other has a mediocre account with sloppier targeting — but a person picks up the phone on the second ring, every time, including at seven in the evening. The second business wins, and it is not close. Someone standing in a room with water spreading across the floor is not evaluating ad copy. They are calling the first listing that answers, and if it does not answer within the length of their patience, they are back in the results calling the next one.

The second lever is quieter and costs more. A local service business can usually perform a dozen distinct job types, and the default assumption is that the account should bid on all of them, because all of them are revenue. That assumption is what fills a calendar with the cheapest work in the category while the jobs that actually carry the business go unbid. Advertising capability is not the same decision as advertising profitably, and conflating them is the most expensive structural error in this vertical.

Everything below is built around those two levers. The targeting, scheduling, landing page and measurement sections are downstream of them — they are how the two levers get expressed inside an account, not separate topics.

Emergency versus planned intent — the defining split

This is the structural distinction that everything else in a local service account depends on, and running both types of demand through the same campaign is why so many of these accounts read as mediocre in aggregate while containing one excellent business and one broken one.

Emergency intent and planned intent are not two ends of a spectrum. They are two different businesses with different economics, different sales cycles, different failure modes and different definitions of success. They share a trade, a van and a phone number, and almost nothing else.

What emergency intent actually is

An emergency search comes from someone with an active, worsening problem. A burst pipe. A dead heating system in a cold week. A locked door with keys inside. A power failure across half a building. The defining property is not the topic — it is that the problem is getting worse while the person searches, so the search is short, the reading is minimal, and the decision window is measured in the time it takes to make one or two phone calls.

Three things follow from that, and all three are unusual in advertising. First, these searches convert on contact rather than on consideration — there is no nurture sequence, no comparison, no second visit. Second, they tolerate premium pricing better than any other demand in the category, because immediacy is the thing being purchased and the alternative is ongoing damage. Third, and this is the part that gets ignored, an emergency click is worth close to nothing if nobody answers. Not less — close to nothing. The person does not leave a voicemail and wait. They go back to the results page.

That asymmetry is the whole argument for treating emergency work as its own campaign. It is the demand most worth having and the demand most completely destroyed by an operational gap. Everything about how you buy it should reflect that: separate campaign, separate budget that cannot be eaten by planned-work clicks, serving hours matched to real answering coverage, ad copy that states availability plainly, and a landing page whose only meaningful action is a tappable phone number.

What planned intent actually is

A planned search comes from someone considering work that nothing is forcing them to do this week. A bathroom renovation. A system replacement before it fails. A rewire during a refurbishment. Landscaping in the spring. The person has time, and they will use it — reading, comparing, collecting quotes, talking to a partner, waiting for a budget cycle.

The economics invert almost completely. Job values are usually higher, but the conversion happens over days or weeks rather than minutes. Speed still matters, but for a different reason: the first business to respond with a real quote sets the frame that every later quote gets compared against. What kills planned work is not slowness on the first response — it is the absence of a second, third and fourth touch. A quote sent and never followed up is the standard way this demand is lost, and it is lost silently, because nobody reports a follow-up that did not happen.

Planned intent also justifies work that is pure waste on emergency terms. Longer, more informative landing pages. Photo galleries of completed work. Financing information. A form that captures enough detail to quote without a site visit. Remarketing. A follow-up sequence with a defined cadence and an owner. None of that helps a person whose ceiling is dripping, and all of it matters for someone planning a kitchen.

How the split shows up in the account

• Separate campaigns, always. Not separate ad groups. Shared budgets are the mechanism by which cheap planned-work clicks consume the money that emergency demand needed, and campaign-level budgets are the only place you can prevent that.
• Separate schedules. Emergency campaigns should serve when someone answers. Planned campaigns can serve during evenings and weekends, because a form submission at nine at night is perfectly serviceable the next morning if the follow-up process is real.
• Separate conversion definitions. For emergency work, the conversion is a connected call. For planned work, it is a completed inquiry with enough information to quote. Counting both as "a lead" hides the fact that one of them is a job and the other is the start of a process.
• Separate copy. Emergency ads state availability, response time and coverage area. Planned ads state credentials, work quality and the quoting process. Copy written for one reads as noise to the other.
• Separate negatives. Planned-work vocabulary is emergency-campaign waste and the reverse is also true — "cost of", "ideas", "designs" and "how long does it take" belong nowhere near an emergency campaign.

The negative side of this deserves more attention than it usually gets, because emergency campaigns are unusually vulnerable to research traffic that shares their exact vocabulary. The full method for building those exclusions from your own data is in the negative keyword list for service businesses, and the categories that matter most here are DIY instruction and price research — both of which look enormously relevant on an emergency term and convert at effectively nothing.

Job selection is a bidding decision, not a capability question

Here is the question almost every local service account gets wrong at setup: which services should we advertise? The default answer is "the ones we offer", and it is wrong in a specific and expensive way.

A plumbing business can unblock a drain and it can install a heating system. An electrician can replace a socket and can rewire a property. A cleaning company can do a one-off clean and can hold a recurring commercial contract. In each pair, the two jobs share a trade and share nothing else — not the value, not the margin, not the time on site, not the competition in the auction, not the likelihood of repeat work.

The cheap job in each pair is almost always the one with the most bidders. That is not coincidence. Low-value jobs have low barriers to performing them, so more businesses can do them, so more businesses bid on them, so the cost of buying one rises toward the point where the job stops making money. Meanwhile the high-value job — the one requiring specialist certification, a bigger crew, a longer engagement — has fewer bidders precisely because fewer businesses can deliver it.

How to actually make the decision

The decision needs numbers, but they are your own numbers and nobody else can supply them. Before touching the account, put your service lines in a list and fill in four things for each one, from your own job records rather than from impression.

1. Typical job value, and the spread around it. A service line whose jobs range enormously in value is a different bidding problem from one that is consistently priced, because the cheap end of a wide range will dominate the volume you actually buy.
2. Gross margin after materials, labor and travel — not revenue. A high-ticket job with thin margin and a long time on site can be worth less than a smaller job you finish in an hour.
3. What follows the job. Some service lines produce a customer; others produce a transaction. A first job that reliably leads to a maintenance relationship can justify an acquisition cost that the same job in isolation cannot.
4. How much of your capacity it consumes. A job type that occupies your best technician for two days has an opportunity cost that never appears in any advertising report, and it is real.

Then rank the list by margin contribution per unit of capacity consumed and start advertising from the top. Not from the top of the volume list. Not from the list of what you are best known for. From the top of that ranking, and stop when the remaining service lines cannot plausibly absorb the cost of buying a customer.

The predictable objection is that the cheap job is a foot in the door — the small repair that becomes the big installation later. Sometimes true, and it is testable rather than assumable. Look at your own records: of the customers who first came for the cheap job, what share came back for anything else, and how long did it take? If you cannot answer that from records, you do not have the evidence, and the safe assumption is that the cheap job is a cheap job.

What to do with the jobs you stop bidding on

Not advertising a service line is not the same as hiding it. It stays on the site, stays on your Google Business Profile services list, stays in your organic content, stays available to anyone who calls. The distinction is narrow and important: you continue to serve that demand, you stop paying auction prices to acquire it. That is usually the correct arrangement for the cheap end of a trade — it arrives through the free surfaces, through referrals, and through customers you already have, and it does not need to be bought.

Geo-targeting by drive time, not by radius

Radius targeting is the default because it is one field and a number, and it is wrong for almost every business that sends a vehicle to a customer. A circle drawn on a map treats every direction as equivalent. Roads do not.

The pattern repeats in every city with any geography to it. A location on the far side of a river with one bridge sits inside the circle and takes forty minutes to reach. A location past a rail crossing, a mountain road, a single arterial that seizes at peak hours, or a lake you have to drive around — all of these read as near on the map and behave as far in the van. Meanwhile a place well outside the circle on a clear highway is twenty easy minutes away and is being excluded.

How to build the area honestly

• Start from drive time, not distance. Decide the longest travel you will accept for each job type — and it should differ by job type, because a large installation absorbs travel that a small repair cannot — then map which places are actually reachable in that time.
• Test it at the hours you serve. A drive time measured at ten in the morning is fiction for a business that mostly travels at rush hour. Check the same routes at the times your technicians actually move.
• Exclude the pockets deliberately. Most platforms let you subtract locations from a targeted area. Use that to carve out the places that look close and are not, rather than shrinking the whole radius and losing good territory in every other direction.
• Set the location option to people in your area rather than people interested in it, unless you have a specific reason otherwise. A local service business selling a physical visit does not want someone in another country researching your city.
• Treat the edge of the area as a different economic proposition, because it is. The same job at the boundary costs more in unbilled travel, fits worse into a day of scheduled work, and displaces jobs closer in.

That last point deserves its own paragraph, because it is the part that gets waved away. Travel is unbilled and it consumes the scarcest resource the business has: technician hours. Two jobs of identical price are not identically valuable if one is fifteen minutes away and the other is fifty. If you are bidding the same for both, you have decided that travel is free, and it is not — it comes out of margin and out of the number of jobs a day can hold.

The practical expression of this is location bid adjustments that reflect job economics rather than click performance. Bid down at the edges. Bid up in the core where a technician can complete more work in a day. And where an area is genuinely unprofitable to serve, exclude it rather than bidding it down and hoping — a reduced bid still buys clicks, just fewer of them.

Ad scheduling against your actual answering capacity

Here is the honest version of ad scheduling, and it is a much shorter conversation than the usual dayparting discussion: do not buy clicks at hours when nobody answers the phone.

Most local accounts do exactly this, and they do it without ever making the decision. The campaign runs on the default all-day schedule because that is what a campaign does. The office is staffed at defined hours. Nobody sat down and compared the two. So the account spends money every evening buying inquiries from people who reach a ringing phone, and those clicks are charged at exactly the same price as the ones that connect.

The reason it survives is that nothing in the reporting flags it. There is no metric called money spent on unanswered inquiries. The clicks appear, the cost appears, the conversions simply do not — and the absence gets attributed to keywords, or competition, or the landing page, or anything except the fact that the business was closed.

The comparison to run this week

1. Pull your account cost by hour of day and by day of week. Every platform reports this and few people look at it.
2. Write your real answering coverage next to it. Not your published opening hours — your real ones, including who is genuinely reachable on a Saturday and what happens to a call at eight in the evening.
3. Highlight every hour where the account spends money and no human answers. That is your first number, and it is usually larger than anyone expects.
4. Decide, hour by hour, between two options: stop serving, or start answering. There is no third option that involves continuing as is.
5. If the answer is start answering, cost the coverage against what those hours are currently spending. An after-hours answering service is a real business decision with a real price, and comparing it against the spend already going into unanswered hours is the honest way to make it.

One more scheduling trap specific to this vertical: the hours that produce the most searches are not necessarily the hours that produce the best jobs. Late-evening emergency demand is often genuine, urgent and willing to pay a premium, and it is also the demand most local businesses are least equipped to serve. That gap is an opportunity for the business that decides to close it deliberately — and a straightforward waste for the one that buys the clicks without closing it.

Call handling is the real conversion step

Every ad platform reports a conversion at the moment a call starts or a form is submitted. That moment is not the conversion. It is the start of a process that the platform cannot see and that determines the entire value of what you just bought.

Think about what actually happens on the other side of a missed call from an emergency search. The person is standing in front of a problem that is getting worse. They called you because you were near the top of the results. The phone rings out. They do not leave a voicemail, because a voicemail does not stop water. They hit back, they call the next listing, and if that one answers, you are finished — not just for this job, but for every future job that customer would have brought, and for whoever they recommend you to. You paid full price for the click and delivered the customer to a competitor.

The mechanism behind response-time decay is worth stating carefully, because it is usually asserted with a statistic and it does not need one. A person who has just searched is in a state that ends. While they are still in it, they are actively looking for someone to solve the problem, they have the search results open, and they are willing to be sold to. An hour later they have called someone else, or the crisis has passed, or they have moved on to something else and the problem has been mentally shelved. Nothing about them is the same. You are no longer responding to an inquiry — you are cold-calling someone who once had a question. The value did not decline gradually. The context that gave the lead its value disappeared.

What to actually put in place

• Count missed calls. Not answered calls — missed ones, by hour, as a standing number somebody sees weekly. This is the single most useful measurement most local service businesses do not have.
• Make the voicemail useful and then call back anyway. A voicemail that states a callback window is better than a generic greeting, and neither substitutes for actually returning the call within minutes rather than the next day.
• Return every missed call from an advertised number, without exception, as fast as physically possible. The odds are worse than answering live and they are not zero, and the cost of trying is one phone call against a click you already paid for.
• Give form submissions a named owner and a target response time, and check it against reality rather than against intention. Forms decay the same way calls do, and they fail more quietly because nobody hears a form ring out.
• Route by availability, not by hierarchy. A lead assigned to the person who is supposed to handle it, who is on a job site, is a lead that waits. A lead assigned to whoever can respond now is a lead that gets answered.

The routing side of this is a systems problem rather than an advertising one, and it is where the largest recoverable losses in most local service businesses sit. The full design — ownership, fallbacks, escalation, and what happens when the assigned person does not respond — is covered in the post on building a lead routing system that does not drop leads, and it is the natural companion to this section: advertising decides how many inquiries arrive, and routing decides how many of them survive the first hour.

A measurement gap Nepali advertisers should know about

Google publishes a list of countries where call reporting and Google forwarding numbers are available, and Nepal is not on it. The practical consequence is specific and worth stating plainly: a phone number can still be added to an ad as a call asset, but the forwarding number, call reporting and call-duration conversion tracking that advertisers in eligible countries rely on are not available. Calls generated by Google Ads in Nepal cannot be measured natively by the platform.

For a market where the phone is a dominant conversion path, that is a structural handicap rather than an inconvenience, and it has to be worked around rather than ignored. The realistic approaches are a third-party call tracking provider that supports numbers in your market, or a disciplined manual process — a dedicated number used only in ads, a log of every inbound call against its source, and a weekly reconciliation against the account. Neither is as good as native reporting. Both are far better than treating call volume as unknowable and optimizing on form fills that represent a fraction of real demand.

Local Services Ads versus Search — what to check before assuming

Local Services Ads are a distinct Google product, not a campaign type inside a standard Search account. They have their own signup and verification path, their own screening requirements for the business and in some categories its staff, their own placement, and a pricing model based on leads rather than clicks. That last difference is the one people notice first, and it is genuinely different: you are buying a contact rather than a visit.

The conceptual case for the product is straightforward. Verification and screening act as a filter, the placement is prominent for exactly the local service queries this post is about, and paying per lead moves some measurement burden off the advertiser. The conceptual cautions are equally straightforward: verification takes time and paperwork, the lead-dispute process matters more than most advertisers expect, and reviews and responsiveness feed directly into how much volume the product sends you — which loops right back to the answering-speed argument above.

Where the product is available to you, the sensible framing is complement rather than replacement. Local Services Ads and Search campaigns occupy different placement and buy demand slightly differently, and running both lets you see which produces jobs rather than which produces cheaper contacts. Where it is not available, nothing in this post changes — the emergency and planned split, the job selection ranking, the drive-time area and the answering discipline are all properties of the business, not of a product.

Google Business Profile is the prerequisite nobody wants to hear

This is the unglamorous section, and it is the one that changes outcomes most reliably for a business that has not done it.

A local service search does not return a list of ads. It returns a page with a map result set on it, and for a large share of local intent that map result set is what people actually engage with — it shows proximity, hours, reviews and a call button in one glance, which is precisely the information an emergency searcher needs and precisely the reassurance a planned-work searcher is looking for. It is also free, and it is available in Nepal.

Which means a business with a neglected profile is paying for clicks to compensate for a free surface it is losing on. That is a bad trade in any market. Worse, the two surfaces are not independent in the reader's mind: someone who sees your ad frequently checks the map listing before calling, and a listing with wrong hours, three reviews and no photos undoes whatever the ad achieved.

The order of work

1. Verify the profile and make sure the business name, address and phone number match your site and your other listings exactly. Inconsistency here is the most common quiet problem and the easiest to fix.
2. Get the hours right, including holiday hours, and make them match your real answering coverage rather than your aspirational ones. A listing that says open when nobody answers produces the same failure as an ad running at unanswered hours, with a review attached.
3. Fill in the services list properly, using the language customers use rather than internal job names. This is also where the job-selection decision shows up: the services you have stopped advertising still belong here, because this surface costs nothing.
4. Define the service area, and define it by the same drive-time logic the campaign uses. Two different definitions of where you work is a fixable inconsistency.
5. Build a review habit rather than a review campaign. Ask at the moment the job is finished and the customer is visibly satisfied, which is the only moment it works, and respond to every review including the unhappy ones.
6. Add real photos of real work. Not stock. For trades in particular, the before-and-after of completed jobs does more persuasive work than any sentence you can write.

None of that is advertising, and all of it changes advertising outcomes. The wider argument for treating local discovery as an architecture problem rather than a listing chore is made in the post on answer engine optimization for local businesses, which covers how the same profile data feeds the assistants and answer surfaces people increasingly use before they ever reach a results page. For advertisers working in the Nepali market, the groundwork underneath it — categories, review behavior, and how the map pack actually behaves here — is in the local SEO guide for Nepali businesses.

Landing pages for local services

Local service landing pages fail for reasons that are almost boringly consistent, and none of them are design problems. They fail because the page does not answer the four questions a local searcher has, in the order they ask them.

• Do you cover where I am? State the service area in plain text — named towns, districts or neighborhoods, not a vague regional claim. Someone who cannot confirm you cover them will not call to find out.
• Can I reach you right now? The phone number goes at the top, is tappable on mobile, and is not an image. If a person has to scroll or squint to find a number on an emergency page, the page has failed regardless of how it looks.
• Are you legitimate? Licensing, registration, insurance and certifications where they apply to your trade, stated with the actual numbers and bodies rather than a badge graphic. For work involving gas, electricity, structural change or entry to a home, this is not decoration.
• When will you actually come? A specific, realistic response promise beats a superlative every time. "Someone answers this number until nine at night" is worth more than "fastest response in the city", and it has the additional advantage of being checkable.

That last one is where most pages overreach and it is worth being blunt about the cost. A promise you cannot keep converts the click and then loses the customer, generates the review that damages the map listing you just fixed, and does it at the exact moment the person is most annoyed. Promise what your actual dispatch capacity supports on a normal day, not on your best day.

The emergency and planned split applies to pages as much as to campaigns. An emergency page should be short, single-purpose and dominated by the phone number. A planned-work page can afford length, galleries, financing details and a form that collects enough to quote. Sending both intents to the same page means one of them is being served badly, and it is usually the emergency one, because the page was designed to look thorough. The broader mechanics of matching a page to the query that reached it are covered in the post on landing page optimization for Google Ads.

Seasonality and demand spikes

Most local service trades have a demand shape rather than a demand level, and the accounts that handle it well do two things the accounts that handle it badly do not: they move before the season rather than during it, and they treat capacity as the binding constraint rather than budget.

Moving before the peak

Planned work has a lead time, which means the searches that produce spring installations happen before spring. An account that increases budget when the season is visibly underway is bidding into the most competitive fortnight of its year, against every competitor doing the same thing, for demand that was decided weeks earlier. The planned campaign should ramp ahead of the season. The emergency campaign should hold through it, because emergency demand appears with the weather rather than ahead of it.

Weather-driven spikes and the capacity ceiling

A cold snap, a storm or a heat wave produces a sudden surge of genuine emergency demand across a whole city at once. Two things are true simultaneously: this is the most valuable demand of the year, and your capacity to serve it is fixed at whatever your crew can do in a day.

That combination changes what good looks like. During a spike the correct objective is not more leads — it is filling every available slot with the highest-value work available, and being honest with everyone else. Concretely: hold the emergency campaign, consider pausing the low-value job types entirely for the duration, and put a realistic wait time on the page rather than taking bookings you cannot honor. A business that overbooks a storm week converts a windfall into a month of complaints and a damaged review profile, which then costs it the next spike too.

The off-season question

The instinct in a slow season is to cut advertising, and it is sometimes right and sometimes exactly backwards. It is right when the demand genuinely is not there — nobody is searching, and buying a larger share of nothing accomplishes nothing. It is backwards when demand is merely lower while competitors have withdrawn, which is when a modest budget buys visibility it could not buy in season. The way to tell them apart is search volume for your own terms, not your booking calendar, because your calendar reflects both demand and your own advertising decisions and cannot separate them.

Where local service accounts actually fail

• Money is spent at hours nobody answers. The most common failure in the category, and the one with the least resistance to fixing — it is a schedule setting and an honest conversation about coverage.
• One campaign carries both emergency and planned demand. Shared budget means the cheaper clicks win the money, and the more valuable intent gets starved by whichever type happens to be cheaper that week.
• Every service line is advertised because every service line is offered. The cheapest, most competitive job type absorbs the budget, the calendar fills with low-margin work, and the account is judged on a lead count that says nothing about any of this.
• The service area is a circle. Jobs at the edge cost more in unbilled travel than the pricing assumed, and territory across a bridge or a rail line is treated as though it were next door.
• Missed calls are not counted. If nobody knows the number, nobody can act on it, and it is invisible in every advertising report ever produced.
• Quotes go out and are never followed up. This is the planned-work equivalent of a missed call, it is worth more per instance, and it produces no signal at all — a quote that dies of neglect looks identical in the account to a quote that was rejected.
• The Google Business Profile is stale while the ad budget grows. Paying to compensate for a free surface that is losing, and the two are visible to the same person in the same session.
• Conversion tracking counts inquiries rather than jobs. Every optimization decision downstream is then made against a number that does not distinguish a five-figure installation from a call about a price list.

That last one propagates further than people expect, because automated bidding optimizes toward whatever you told it to value. If every form fill and every call counts the same, the system will efficiently find you more of the cheapest ones. The audit sequence for getting the measurement right before anything else is in the conversion tracking audit, and for a local service business the specific thing to get right is that different job types must not be worth the same number.

The metric that lies

Here is the pattern, and it is so common in local service accounts that it is nearly diagnostic. Something changes in the account — a budget increase, a broadened match type, a new campaign, a bid adjustment. The next month, call volume is clearly up. Cost per lead may even be down. Every number on the report that a business owner looks at has improved.

And booked jobs are flat. Revenue is flat. The technicians are not busier.

The mechanism is job mix. The change did not bring more of the same demand — it brought more of a different, cheaper demand. Broadening a campaign in a local service business almost always broadens toward the low-value job type, because that is where the query volume is, and it is where the cost per click is lowest, which is exactly why an efficiency-seeking system drifts there. So the call counter rises, filled with inquiries about the cheap job, while the demand that pays for the business sits unchanged.

Call volume cannot detect this, and neither can cost per lead. Both treat every inquiry as one unit. The second number that catches it is booked jobs and revenue segmented by job type, compared against the same segmentation from before the change. If the total went up but the mix shifted toward your cheapest line, the change made the business busier and not better — and busier at low margin is a worse position than it was before, because capacity is now committed.

The version of this that hurts most is the one where the low-value work crowds out the high-value work. Your technicians have a fixed number of hours. Filling them with the cheap job means the profitable job that calls on Thursday gets pushed to next week or declined. The advertising report shows growth. The business is going backwards, and nothing in the account will tell you.

Preventing it is a governance question more than a reporting one: changes get recorded, and the effect of each change is read against segmented job outcomes on a fixed window rather than against a lead count. The discipline for that — what gets changed, how often, and what evidence is required before and after — is set out in the bid governance framework, and it applies with particular force here because local accounts are small enough that one person changes things reactively and nobody writes it down.

What this cannot fix — the honest scope

Everything above assumes a business that can do the work well and wants more of it. Where that assumption does not hold, advertising is the wrong instrument and will make the underlying problem more expensive rather than less.

• It cannot fix a capacity problem. If you are already booked out, more inquiries produce longer wait times, more declined jobs and more people who had a poor experience trying to hire you. The correct response to being fully booked is raising prices or hiring, not advertising.
• It cannot fix an uncompetitive offer. If your pricing, availability or coverage is genuinely worse than the alternatives in your market, buying more visibility means more people encounter that and choose otherwise. This is the answer nobody wants and it is sometimes the true one.
• It cannot fix a reputation problem. Ads deliver people to a business they will then look up. A weak review profile intercepts them at exactly that moment, and no bid changes it.
• It cannot substitute for referral and repeat business in a trade where those dominate. For many established local service businesses, advertising is a supplement to an existing customer base and a marginal source of new work, not a replacement for either.
• It cannot make a low-value job type profitable. If a job cannot carry the cost of buying a customer, no account structure will change that arithmetic. Refusing to bid on it is the fix.
• It cannot compensate for an operational gap. This is the whole argument of the post. An unanswered phone, an unfollowed quote and an unassigned lead each destroy the value of the click after you have already paid for it, and no amount of account work reaches them.

There is also a scale threshold worth naming. A one-van operation with a full calendar and a steady referral flow may simply not need paid search, and the honest recommendation in that case is a well-maintained Google Business Profile, a site that answers the four questions above, and a phone that gets answered. That combination is free, and for a genuinely local single-operator business it is frequently sufficient. Paid search becomes the right tool when you have capacity you want to fill, a job type whose economics can carry an acquisition cost, and someone whose job it is to answer.

A starting sequence

If you are setting this up or rebuilding it, this is the order. It deliberately front-loads the parts that are not advertising, because those are what determine whether the advertising is worth doing.

1. Count your missed calls for two weeks, by hour. Do nothing else first. This number is usually the largest single recoverable loss in the business and almost nobody has it.
2. Write down your real answering coverage — the hours a human genuinely picks up, not your published hours — and decide what happens to a call outside them.
3. Rank your service lines by margin contribution per unit of capacity consumed, using your own job records. Decide which ones you are willing to buy customers for and which ones you are not.
4. Fix the Google Business Profile completely: verified, correct hours matching real coverage, full services list, drive-time service area, real photos, a review habit with an owner. This is free and it comes before spending.
5. Define the service area by drive time at the hours you actually travel, and carve out the pockets that look near and are not.
6. Split the account into emergency and planned campaigns with separate budgets, separate schedules, separate copy and separate conversion definitions. Do not use ad groups for this.
7. Restrict the emergency campaign to hours somebody answers. If that leaves valuable hours unserved, cost out real coverage for them against what those hours would otherwise spend.
8. Build the landing pages to match: a short, phone-first page for emergency intent, a longer, proof-heavy page with a quoting form for planned work. State the service area in plain text on both.
9. Set up call measurement before launch. Where native call reporting is unavailable in your country, decide now whether that is a third-party provider or a disciplined manual log, because retrofitting it later means a blind first quarter.
10. Give leads an owner and a target response time, with a fallback for when the owner does not respond. Advertising without this is buying inquiries you have not decided who will answer.
11. Launch narrow. One emergency campaign on your best job type and one planned campaign, both in your core area. Expand only from evidence, and only after the first read on booked jobs by type.
12. Review booked jobs and revenue by job type monthly, not lead counts. That is the number that tells you whether any of this worked.

Where to start if you only do one thing

Count the missed calls. It costs nothing, takes two weeks of paying attention, and it will tell you whether your problem is upstream in the account or downstream in the operation. In most local service businesses I have looked at, the answer is downstream — and no amount of restructuring, bid tuning or copy testing touches a phone that rings out.

Then make the job selection decision, honestly and in writing, before the next budget conversation. Being able to name the job type you have deliberately stopped bidding on, and defend the reason out loud, is the clearest signal that the account is being run on business economics rather than on lead counts.

I run the paid side of Arcetis on this exact order of operations, and the pattern holds across markets: the businesses that win local service advertising are not the ones with the best-structured accounts. They are the ones that answer, and that know which jobs they are buying.

Related reading