Google Ads for Small Business: What It Costs and What It Takes

What Google Ads really costs a small business, how long it takes to work, and how to tell before you spend a dollar whether it suits your business at all.

By Gavin Sevastian · Updated August 4, 2026 · 15 min read

Most owners want two things settled before they spend anything. Will paid search work for a business my size, and what does it actually cost?

The second question has published numbers behind it, and they are all in here. The first has an answer most pages will not give you: sometimes no. Some businesses should not run this channel at all, and you can work out which group you are in before you spend a dollar.

Google Ads for a small business is not a different product from Google Ads for a national chain. Same auction, same interface, same click prices. What differs is that a small budget has no room to absorb a mistake, which is why several of the leaks further down cost you far more than they cost a company spending forty times as much.

Start here: the 15 minute test of whether Google Ads fits your business

Google Ads buys demand that already exists. It does not create it. So the first three checks are whether that demand exists in your area, whether you can catch it when it arrives, and whether the numbers work on the jobs you actually sell.

Check that people actually search for what you sell

Keyword Planner is free with a Google Ads account and does not require a live campaign or any spend, so you can check demand before committing money.

When you read the results, separate two kinds of terms. A term that describes the job is someone with a problem: furnace repair, garage door won't open, emergency plumber. A term that describes the product is someone reading: furnace prices, best shingles, how much does a deck cost. Both show volume. Only one of them books work this week.

There is no universal number to clear here, and anyone who hands you one is guessing at your economics. The volume only has to be large enough that, at your trade's click price and a realistic conversion rate, it funds enough leads to matter against your average job. Which means this check and the arithmetic check two sections down are really one test run in two halves.

1. 1

   Create a Google Ads account without launching a campaign

   Skip the guided campaign setup when it prompts you. Keyword Planner does not require an active campaign or any spend, so nothing starts serving while you look.

2. 2

   Set the location and currency to your own market

   Set the geographic target to your city or service area rather than the whole country, and confirm the currency matches what you would actually be billed in. National volume tells you nothing about whether your van can be busy.

3. 3

   Pull job terms and product terms separately

   Enter ten to fifteen phrases a customer would type when they need the work done, then a second batch describing the product or the price. Compare the two lists before you form an opinion about demand.

Check that you can answer the phone during the hours the ads run

Paid leads decay fast. The MIT and InsideSales lead response study, which tracked more than 15,000 leads and over 100,000 call attempts across six companies, found the odds of reaching a lead are roughly 100 times lower at thirty minutes than at five, and the odds of qualifying one about 21 times lower. You can read the study yourself.

That is a genuine disqualifier, not a nag about being organized. A contractor on a roof at 2pm with nobody covering the phone is paying full retail for voicemail, and the caller is already talking to the next company on the page.

The expensive version of the same problem is running ads on days you cannot service. You pay the same click price on a Saturday you never work as you do on a Tuesday you do, and the caller who gets nothing forms a permanent opinion of your business.

Check the arithmetic on your average job

Lead cost does not scale down to match your ambition. The market sets it, and you either clear it or you do not.

The chain is short. Cost per lead, divided by the share of leads you close, is what it costs in advertising to book one job. Then you compare that number against what the job earns you after materials and labour.

You know your own close rate. If you do not, the only public anchor worth using is SearchLight Digital's contractor dataset, covering 888 contractors, which reported a 41.7% book rate on Google Ads leads. Run that against the published US home services median cost per lead of $90.92 and you land near $218 in advertising behind every booked job. Those are two separate datasets, so treat it as a shape rather than a forecast.

Those are also US dollars from US campaigns. Your own numbers will be in CAD, and the method does not change, only the inputs. What you are testing is simple: can the job you actually sell carry that number and still be worth the truck roll? A roof replacement absorbs it without noticing. A small service call does not, unless it reliably becomes a second job. Are Google Ads worth it for contractors runs the longer version of this calculation.

What Google Ads for small business actually costs

What a click costs in home services

Two published reports cover this, and both come from LocaliQ (WordStream is its brand, so treat them as one source rather than two agreeing with each other).

• All industries, from the 2026 Google Ads benchmarks covering 13,474 US search campaigns: $5.42 USD per click, an 8.18% conversion rate, $66.69 per lead.
• Home and Home Improvement, from the same report: $8.33 per click, an 8.05% conversion rate, $90.92 per lead. It is the second most expensive category by click price, behind attorneys and legal services.
• Home services, from the 2025 home services report covering 3,211 US campaigns: $7.85 per click, a 7.33% conversion rate, $90.92 per lead.

The two home figures differ because the categories are drawn differently and the samples cover different periods. They agree on the thing that matters: a lead in the trades costs noticeably more than the all-industry median.

Underneath that, trades differ enormously.

Read across the roofing and painting rows. Roofing has the lower click price of the two and costs well over a hundred dollars more per lead, because its conversion rate is roughly a third of painting's. Click price on its own tells you very little about what a lead will cost you.

Every figure above is a US median, in US dollars, from US campaigns. They tell you the shape of the market and the spread between trades. They are not a Canadian budget. If you are pricing this in CAD, pull your own terms in Keyword Planner in your own currency and run the same arithmetic on your own numbers.

How a click price becomes a monthly budget

Cost per click, divided by conversion rate, gives cost per lead. Multiply by the leads you need in a month and you have a budget.

Run it on the home services medians of $7.85 per click and a 7.33% conversion rate. That is about 13.6 clicks per lead, or roughly $107.

Now compare that to the cost per lead the same report publishes: $90.92. The two do not match and they never will, because a median of ratios is not the ratio of medians. Use the arithmetic to understand the shape of your own numbers, and use published figures when you want something to plan against.

Against the published $90.92, ten leads a month is around $909 in ad spend and thirty is around $2,728. Thirty matters because it is the volume Google's Target CPA guidance suggests you have before judging performance, which is a reasonable bar for any account.

Trade level changes the number a lot. Ten roofing leads at the published median is closer to $2,282. Ten painting leads is around $1,384. Ten handyman leads is about $541. The trade-by-trade breakdown goes further into why the spread is that wide.

How Google actually charges you

Your daily budget is an average, not a cap. Google's documentation on daily and monthly spending limits says that for most campaign types you will never pay more than twice your average daily budget on a given day, or 30.4 times it in a month. The 30.4 is just 365 divided by 12. So a $30 daily budget can spend $60 on a busy Monday and is capped near $912 for the month.

Then the part almost nobody publishes. Ad scheduling does not reduce your monthly spend. Google states that the monthly limit stays at 30.4 times your average daily budget regardless of how many days the campaign is scheduled to run. A Monday to Friday campaign compresses the full monthly target into roughly 21 serving days, so it spends about 1.4 times the daily budget on the days it is actually on.

Work it through on that $30 daily budget. Running seven days a week it targets around $912 a month and paces near $30 a day. Restrict it to weekdays and the monthly target does not move, so it now paces near $42 on every day it runs. Owners who switch off weekends to save money and then see higher weekday spend usually assume something broke.

The practical instruction is the reverse of how most people set it up. Start with the monthly number you can genuinely live with, then use Google's own advice on the average daily budget and divide by 30.4 to get the daily figure you enter.

How the auction works, and why a local business can beat a bigger bidder

Your bid is one of six inputs, not the whole thing

Google's Ad Rank documentation names six factors:

• Your bid
• The quality of your ads and landing page
• The Ad Rank thresholds
• How competitive the auction is
• The context of the person's search
• The expected impact from your ad assets and other ad formats

Google's glossary entry on Ad Rank puts the consequence plainly: even if your competition has higher bids than yours, you can still win a higher position at a lower price by using highly relevant keywords and ads.

That is the structural reason a two-truck operation can outrank a national franchise in one town. Relevance is something a local business is naturally better at, because it genuinely does one thing in one place.

Quality Score is a thermometer, not a thermostat

A lot of guides tell you to raise your Quality Score to win the auction. Google's own page on Quality Score says the opposite in its own words: "Quality Score is not an input in the ad auction. It's a diagnostic tool." The same page adds that it "is not a key performance indicator and should not be optimized or aggregated with the rest of your data."

Hold the distinction, because it blurs easily. Ad and landing page quality genuinely is an Ad Rank factor. The 1 to 10 number in your reporting column is a comparison against other advertisers whose ads showed for the exact same search over the previous 90 days. One is a factor. The other is a readout.

What the column is actually good for is its three parts: expected click-through rate, ad relevance, and landing page experience. Each points at a different place to go and look. The headline number points nowhere.

A separate auction runs for every search, every time

There is no leaderboard you climb once and hold. Google runs different auctions for each ad location, and calculates Ad Rank each time someone searches.

That explains something owners often panic about. Your position moves day to day while nobody touches anything, because the competitors, the wording of the searches and the context around them all changed. A four week trend is worth reacting to. A Tuesday is not.

Month one versus month three: what actually changes

Month one is calibration, not performance

Google documents a learning period, and its guidance on the duration of the learning period is specific: it can take up to three weeks or one to two conversion cycles for a bid strategy to calibrate, though it can be faster when there is plenty of conversion data. A conversion cycle just means the typical lag between the click and the lead in your trade, which for an emergency call is hours and for a roof replacement can be weeks.

Google names three things that start that clock: a new strategy, a setting change, and a change to the composition of the campaign. Change the bid strategy on Tuesday and restructure the ad groups on Sunday and you have restarted it twice without ever seeing a calibrated week. Google also notes its algorithms keep learning after the Learning label disappears, so the label going away is not a finish line. Which strategies trigger it, and what each one is for, is covered in Google Ads bid strategy.

Month two is where the waste gets cut

By the second month there is a real search terms report to work from, and it usually contains things nobody expected to pay for: people looking for a job rather than a contractor, homeowners trying to do it themselves, someone hunting for a part, and searches from towns you do not serve.

The keywords you chose and the searches you paid for are two different lists. Match types expand, so what you typed into the account is a starting instruction rather than a boundary. The report is where you find out what Google decided that instruction meant.

Cutting that traffic is the only part of the work where you are removing spend rather than moving it around, and none of it exists on day one. A week-one account and a week-eight account are barely the same account.

Month three is the first fair read

By month three there is usually enough volume for a verdict to mean something. The same Target CPA guidance cited earlier recommends measuring over the last 30 days and including at least 30 conversions before judging it.

Now the part that applies to a lot of small accounts. If your budget produces six or eight conversions a month, month three still is not 30 conversions. You are measuring with an instrument too coarse for the job, and you will be tempted to act on differences that are only noise.

That does not mean the account is failing. It means the honest read takes longer, and anyone promising a clean verdict at 30 days on a small budget is telling you what you want to hear. Small counts are noisy, so two weeks with zero leads on a budget that expects six or eight a month carries almost no information. Zero clicks in week two is a different thing entirely: that means the ads are not running, and it is checkable today.

Nine ways small accounts lose money, and why each one hurts more at your size

Every one of these can happen in a large account too. The difference is absorption. A large budget can carry two or three of them and still look fine in the totals. A small one cannot, and one of them can be the difference between the channel working and the channel not.

Leaks in what you are measuring

1. Running without working conversion tracking

With no conversion data flowing back, nothing else on this page can be measured. Cost per lead is unknown, the search terms report cannot be sorted by what actually worked, and automated bidding optimizes toward a goal it has no feedback on, so it happily buys the cheapest clicks it can find. It is not that this failure is exotic. It is that every other number becomes unreadable when it is missing.

How you spot it: clicks and impressions are climbing and the conversion column is empty, or shows a figure that never moves.

The fix: test the whole path before the first click, with a real form submission and a real phone call from a real phone. A green status label in the interface is not proof. If the phone is already quiet, why your Google Ads are not getting calls walks the diagnosis.

2. Counting leads instead of booked jobs

Cost per lead is a proxy for the number that pays your bills, which is cost per booked job weighed against what that job earns you. The ad platform cannot see your invoices, so it cannot tell a price shopper from a signed contract. Only you can, and if you never feed that back, the account optimizes toward whatever is cheapest rather than whatever is worth having.

How you spot it: you can say what a lead cost last month but not what share of leads became work, or which campaign the good jobs came from.

The fix: record the outcome of every paid lead somewhere, even a spreadsheet with four columns. Source, date, quoted, booked.

3. Assuming the search terms report shows everything

It is the best window you have into what you actually paid for, but it is not a complete ledger. Google states that some search terms are omitted from the report when they do not have enough query activity, to keep with its standards on data privacy.

How you spot it: the clicks listed in your search terms report add up to noticeably fewer than the clicks in your campaign totals.

The fix: compare those two numbers once, and treat the gap as traffic you cannot see rather than traffic that does not exist.

Leaks in who sees the ad

4. Google's default location setting includes people who are not there

Straight from Google's page on preventing clicks outside your targeted locations: the default setting is "People in, regularly in, or who've shown interest in your included locations (recommended)." Read that last clause again. Somebody three provinces away researching contractors in your city qualifies as having shown interest, and you pay full price for the click.

How you spot it: calls asking whether you cover a city three hours away, or form fills with postal codes you do not recognize.

The fix: change the location option away from that default to the one covering only people in your service area, ideally before the first click rather than after the first month.

5. Negative keywords do not work the way positive ones do

Positive keywords expand. Negative ones do not, and Google says so plainly on its negative keywords page: they will not match to close variants or other expansions, so you have to add synonyms and singular or plural versions yourself. There is a second catch on the same page. If your negative word falls after the sixteenth word of somebody's search, your ad can still show.

How you spot it: the same unwanted theme keeps reappearing in the report in slightly different wording after you thought you had excluded it.

The fix: treat exclusions as a list you maintain. Adding negatives at launch is a start, not protection.

6. Broad match with no conversion history to guide it

Broad match reaches the widest set of searches, and what makes it usable on a mature account is a substantial pile of conversion data steering it. A new account has none of that. The setting is the same, the outcome is not: the widest possible net with no feedback telling it what to keep, and a small budget reaches the bottom of that net fast.

How you spot it: a search terms report full of phrases you would never have thought to bid on, most with one click and no conversion.

The fix: earn it. Start narrow, get tracking right, and widen match types once there is conversion data worth steering with.

Leaks in structure and destination

7. Copying a big advertiser's campaign structure

Structure looks free and is not. Every campaign you add divides the same money into smaller piles. Split $60 a day across six tidy campaigns and each one gets $10, which at the roofing click price in the table above does not reliably buy a single click per day. Six campaigns then spend a month collecting numbers too small for anyone, human or algorithm, to draw a conclusion from.

How you spot it: campaigns showing a handful of clicks for the whole month, and a report where no single line has enough data to act on.

The fix: run fewer campaigns than feels tidy. Consolidation is a small-budget advantage.

8. Sending paid clicks to the homepage

A homepage is built for someone who already knows who you are and is having a look around. A paid click is a stranger with one specific problem and very little patience. Message match matters more the more the click costs: at $10 a click you cannot afford to spend the visitor's first ten seconds on a navigation menu.

How you spot it: every ad in the account points at the same URL, and the clicks arrive while the phone stays quiet.

The fix: one page per service you advertise, using the same words the ad used.

9. Applying whatever Google recommends

Recommendations arrive with a score attached and a one-click Apply button, and some of them are genuinely good. But the party making the suggestion is paid when you spend more, not when you earn more, and that shapes which suggestions get made. The auto-applied ones are the sharper version of the problem, because they change your account without you reading them first.

How you spot it: settings that changed on their own, or a recommendations score you feel vaguely guilty about.

The fix: run each one through a single question. Will this book more profitable jobs? Some pass. Switching auto-apply off costs you nothing.

The businesses Google Ads does not suit

Three situations where the honest answer is that this channel is neither your problem nor your solution.

When nobody is searching

Google Ads buys demand that already exists, which means it needs people typing something. If your service is new, unfamiliar, or solves a problem homeowners do not know they have, there is no inventory to buy. You will find this in fifteen minutes in Keyword Planner: a few dozen searches a month across an entire province, most of them from people reading rather than buying.

That is the wrong channel for the job rather than a failure of the ads. Businesses in this position need something that interrupts people who were not looking, which is what social platforms are built for, and that is a different budget with a different set of expectations attached.

When the job value cannot carry the lead cost

Take the roughly $218 USD in ad cost per booked job from the top of this page and work backwards through your own business rather than forwards from the ad account.

Say your average ticket is $500 and you keep 40% of it after materials and labour. That job earns you $200 and cost $218 to buy. You lost money before you started the truck. To break even at that margin you need a ticket around $545, and to make it genuinely worth doing, meaningfully more than that.

The exception matters as much as the rule. If that first job reliably turns into a maintenance customer for the next several years, the arithmetic changes completely, because you are not buying one job, you are buying a relationship. Recurring services and genuinely high lifetime value change this. One-off small tickets do not. How to get plumbing leads works through a version of this with one trade's numbers.

When the problem is not lead volume

Google Ads amplifies the business you point it at. It does not fix a weak offer, a thin review profile, or a close rate that is quietly the real bottleneck. It makes each of those more expensive per unit, because now you are paying retail for the leads you were already losing.

A business that converts one inquiry in ten does not have an advertising problem yet. Doubling the inquiries doubles the cost of that leak. Fixing the close rate first makes every future ad dollar work twice as hard, and it is free.

There is a related trap in how owners judge all this. Paid leads arrive with an invoice attached and referrals do not, so ads look expensive by comparison. Referral acquisition is not free, it is just invisible: it costs you reputation building, follow-up, and hours you never billed. The comparison that means something is cost per booked job against what the job earns, and you can run that on both channels.

If you cannot say what percentage of your current inquiries turn into booked jobs, that number is the project, not the ads.

Google Ads is not the only paid option, and sometimes not the best one

Local Services Ads: different product, different economics

Local Services Ads are a separate product from Search, often called the Google Guaranteed pack. You clear a verification process before you can run them, and the economics differ enough that the comparison is worth doing rather than assuming.

SearchLight Digital published a dataset covering 888 contractors, 1,774 campaigns and 126,650 leads. Its reported figures:

• Cost per lead: $53 on Local Services Ads against $104 on Google Ads
• Book rate: 43.9% against 41.7%
• Cost per customer: $233 against $472
• Average ticket: $1,826 against $2,465

That last row matters, and most write-ups leave it out. The cheaper leads were also smaller jobs, so neither channel wins on cost per lead alone. The comparison is worth running on your own trade, and LSA vs Google Ads goes deeper into where each one lands.

SEO: same searches, different timeline and different risk

They chase the same clicks on the same page. The difference is when they arrive and what happens when the money stops.

Most established contractors end up running both, using ads to fill gaps and to control which services get promoted, and organic to lower the average cost of everything over time. Google Ads vs SEO for contractors has the longer comparison.

How to decide which to start with

If your trade has a Local Services category and you can clear verification, that is a reasonable place to start, because the published cost per lead is lower. Verification is the friction and it takes longer than people expect, so begin it early. How to get Google Guaranteed covers what is involved.

If your average job is large enough that job size matters more than lead price, Search earns its premium, because you control which searches you show for and exactly what page they land on.

And if you have no website and no conversion tracking, neither one is ready for your money yet. That is the cheapest finding on this page.

What running a Google Ads account actually involves, week to week

Whether you hire someone or do it yourself, it helps to know what the work is. The setup is a weekend. What happens in week seven decides whether the account earns money.

The recurring loop

The job is a weekly loop. Read what people actually searched for. Exclude what does not belong. Check the ads still match the pages they point at. Confirm the tracking is still recording. Then decide where next week's money goes based on what the first four told you.

None of those is difficult on its own. What makes them work is that they happen again the following week, and the week after that. Each pass is small and unremarkable, which is exactly why the loop gets skipped, and why the waste it prevents never shows up as a dramatic before-and-after. It was never on a chart in the first place.

Why the ground keeps moving

Google-wide, about 15% of all searches have never been searched before. Your trade's vocabulary moves slower than the global query stream, but it does move, which is why an exclusion list that was complete in March has holes in it by June.

The platform moves too. Defaults change, bidding options get renamed, match behaviour gets adjusted, and new ad formats get switched on automatically. Every one of those changes is a small nudge toward the platform's preferences rather than yours.

The consequence is the part owners underestimate. An unmanaged account does not plateau, it decays, and the decay is quiet. Nothing errors. The ads keep running, the invoices keep clearing, the cost per lead drifts up slowly enough that no single month looks wrong, and nobody can point at the day it broke.

The part that takes the longest to learn

Most of the skill is not knowing what the settings do. It is knowing which of Google's suggestions to refuse, and where to look first when something is wrong.

Order matters more than any individual tactic. Start with tracking, because on an account whose conversion tracking never fired, nothing you change afterwards can be measured. Rewrite the ad copy on that account and you will not even be able to tell whether the new copy helped, so you have spent three weeks and learned nothing.

That is not a clever insight. It is just expensive to learn, because you learn it by getting it wrong on live accounts with real money behind them.

Doing it yourself, hiring a freelancer, or hiring a manager

When doing it yourself is the right call

This is not a consolation prize. Plenty of contractors run their own accounts well, and the conditions where it makes sense are specific enough to check yourself against.

Your budget is small enough that a management fee would be a large share of it. You sell one or two services rather than nine. Your service area is a tight radius rather than half a province. And you have a standing weekly slot you will actually keep, including in the weeks you are busy, because that is exactly when the account drifts.

Two things you have to commit to, or it does not work. Install conversion tracking properly and verify it with a real submission and a real call. Then work the weekly loop above without skipping it.

What you are actually buying when you pay someone

You are buying time and pattern recognition. The interface is the same one you can open yourself. What differs is how many accounts the person has watched fail, and in what ways.

You will meet two fee structures. A percentage of spend scales with your budget, which creates an obvious incentive to grow the budget. A flat monthly fee does not, though it does mean a small account and a large one cost the provider very different amounts of effort for the same money.

At the small end a percentage produces a fee that is hard to run a business on, which is part of why flat fees are common below a certain account size. If someone quotes you a percentage on a small budget, ask what happens to their attention when a larger client calls in the same week. What Google Ads agencies charge covers the structures in more detail.

The questions to ask before you hire anyone

Use these on any provider, including us.

1. Who owns the ad account, and what happens to it if we part ways?
2. How do you verify conversion tracking rather than assume it works?
3. What will you refuse to do that Google recommends, and why?
4. Does the monthly report show leads, or booked jobs?
5. How long before you would expect an honest read on whether this is working?

The answers matter less than whether they are specific. Vague answers to any of the five are the signal. If you want to ask us the same five, that is a fair use of a first call.

Writing ads that will not get you in trouble in Canada

Almost nobody covers this, and it is useful whether you run the ads yourself or not. The Competition Bureau regulates advertising claims in Canada, and a Google headline is an advertising claim.

The general impression test

The Bureau explains that when courts assess marketing information, they consider both the literal meaning and the general impression it creates. That second part is what catches people. A headline can be technically true and still leave a false impression, and the false impression is the problem.

This matters more in paid search than in most places, because a search headline is a single short line. The qualifier that makes a claim accurate rarely fits, so it gets dropped, and what is left reads as a promise you did not intend to make.

Advertised prices and the fees that follow

The Bureau defines drip pricing as offering a product at a price that is unattainable because the customer must also pay additional charges or fees, and treats it as false or misleading unless those extra fixed charges are imposed by the government, such as sales tax.

That is exactly the shape of a "$99 furnace tune-up" headline attached to a mandatory trip charge, diagnostic fee, or fuel surcharge. If nobody can actually get the service for $99, the number in the ad is not a price, and government-imposed charges such as sales tax are the only add-ons the definition carves out.

Countdowns that never actually end

The Bureau names fake urgency directly, including limited-time offers or countdown clocks that are renewed or frequently repeated after the advertised deadline, along with untrue claims about limited inventory or high demand.

The contractor version is a "spring offer ends Friday" that has been in the ad rotation since last spring, or a countdown on a landing page that resets every Monday. The Bureau also states that liability rests solely with the originator of the misleading information, though its own examples deal with supply chains of manufacturers, distributors and retailers, so how that plays out between a business and whoever runs its ads is not something the page addresses.

Ontario contractors deal with a crowded, expensive auction on top of all this. If you work the GTA, the local picture is worth its own read in Google Ads in Toronto.

Common questions about Google Ads for small business

Where to start

If you take one thing from this page, take the order. Demand first, then your ability to answer the phone, then the arithmetic, and only then the account. Most failed campaigns failed at step one or step three, and nobody checked.

1. Pull your own trade and city in Keyword Planner, and separate job terms from product terms.
2. Decide honestly whether the phone gets answered during the hours the ads would run.
3. Run the job-value arithmetic. Average job, margin, close rate, against a realistic cost per lead in your own currency.
4. Install conversion tracking and verify it with a real form fill and a real phone call, before the first click.
5. Change the location setting away from Google's default so it covers only your service area.
6. Work the weekly loop, and treat exclusions as maintenance rather than a task you finish.