Google Ads for small business works when the budget matches the cost of a click in your sector, the conversions counted are real enquiries, and the account starts from the Search network with controlled keywords, not from Performance Max. The cost of an enquiry is worked out before spending (CPC divided by conversion rate): with the UK CPCs measured in September 2026 it runs from about £40 to over £700 depending on the sector, and the sensible minimum budget follows from it.
Sounds like accountant's maths, right?
It is. And it's exactly the maths almost nobody does before opening an account.
The mistakes we find in the accounts we inherit all look alike:
- conversions that count technical events, with a cost per conversion under one euro that makes everything look perfect;
- one single campaign for every service, with the budget ending up on the most expensive keyword;
- Performance Max from day one, fed with the wrong data.
That's where this article comes from.
So why listen to us?
Because among the projects we run there are two extremes that contain almost everything a small business needs to know.
Contesini Consulenze, two energy advice branches in the province of Mantua, Italy, got 299 leads in 7 months at €7.20 each, with 68% becoming opportunities. Spend: about €300 a month.
Valore Solare, solar panel installations, paid €212 per enquiry for 9 enquiries in its first month. In 11 months it reached 1,506 enquiries at an average of €70.78.
Here you'll find the decisions in order, with the maths done on DataForSEO data (Google UK, September 2026) and on our own accounts:
- what a click costs in small business sectors (and so what an enquiry costs);
- why start with Search and not Performance Max;
- how to structure the account, choose keywords and negatives, and where to send the ads;
- how to count real conversions;
- what the minimum budget is, sector by sector.
This is the companion piece to the guide to SEO for small businesses and the round-up of marketing tools for small businesses; the service is described on the Google Ads page.
Let's start with the number that decides everything else.
What a click costs in small business sectors (from £1.91 to £35.90)
CPC (cost per click) decides everything else, and it varies more than tenfold from one sector to another. The table shows Keyword Planner values via DataForSEO for common commercial searches, with the cost of an enquiry calculated at a 5% landing page conversion rate: an average in our accounts between B2B (lower) and local services with a phone number (higher).
| Search | Monthly volume | Average CPC | Top-of-page bid | Cost per enquiry at 5% |
|---|---|---|---|---|
| house renovation london | 480 | £1.91 | £1.84-6.59 | £38 |
| builders manchester | 880 | £2.29 | £1.74-5.78 | £46 |
| pneumatic components | 170 | £2.84 | £0.59-2.57 | £57 |
| led wall | 1,000 | £3.03 | £1.34-4.20 | £61 |
| packaging machines | 720 | £4.58 | £1.48-5.31 | £92 |
| builders london | 720 | £5.47 | £2.19-9.08 | £109 |
| dentist manchester | 2,900 | £6.37 | £2.30-6.94 | £127 |
| plumber london | 5,400 | £7.31 | £4.28-33.15 | £146 |
| accountant london | 3,600 | £9.48 | £3.13-12.60 | £190 |
| cnc machining services | 880 | £23.30 | £3.08-28.15 | £466 |
| emergency plumber london | 1,600 | £30.09 | £15.27-55.82 | £602 |
| divorce solicitor london | 480 | £35.90 | £8.30-39.48 | £718 |
The table says three things people often miss.
One: in industrial B2B the problem is often volume, not cost.
"Subcontract machining" gets 50 searches a month in the UK. Even buying every impression, with a 10% CTR (the share of people who click) you get 5 clicks.
At 5%, that's one enquiry every four months.
You need dozens of keywords to build volume.
Two: when a generic B2B term is expensive, look at who you're bidding against.
On "cnc machining services" (£23.30) the first page is held by online quoting platforms: Xometry, Protolabs, Geomiq, Fractory, PCBWay.
A machine shop will do better on narrower terms like "pneumatic components" at £2.84.
Three: in local services a high CPC is offset by a higher conversion rate, because the conversion is a phone call. For hotels ("hotels in london", £1.73) and restaurants the maths is in the sector guides.
And where does that 5% come from?
It's a working assumption, and we'll say so.
On Macropix, between July 2025 and June 2026, the average CPC was €0.81 and the cost per conversion €44.93: a rate of 1.8%. Normal in B2B with quotes worth tens of thousands of euros.
On Contesini, with the phone call as the conversion, the rate was much higher.
Search before Performance Max (because Sunscape counted 1,487 "conversions" at €0.61)
Performance Max optimises on the conversions you give it, and in most of the accounts we inherit the conversions are fake. That's why in our accounts it's the wrong way to start, even though Google suggests it to every new account and many consultants recommend it to small businesses because it "does everything by itself".
Two examples, both real.
Sunscape in 2024 counted 1,487 "conversions" at €0.61 each. They were technical events.
Valore Solare had 15,267 in a month against 5,467 clicks.
More conversions than clicks?
Exactly. And a Performance Max fed with that data would have spent the whole budget generating more technical events.
The order we follow:
- a Search campaign with hand-picked keywords and verified tracking, for the first two or three months;
- a Performance Max alongside (not instead of) Search, only once the account counts at least 30 real enquiries a month, with brand exclusion and the campaign-level negatives available since 2025.
On Contesini the local Performance Max came later, on top of local campaigns already at €7.20 per lead.
On Display remarketing, which a small business can switch on for a few tens of pounds a month, we wrote how to do remarketing.
Further reading: when Google Ads makes sense and what to expect from the channel.
How to structure a small business account (three services, four campaigns, not twenty)
One campaign per product or service line, one ad group per intent within each, the brand separate. With three services and a £1,000 monthly budget that's three campaigns plus one, not twenty.
Geography depends on the sector.
In industrial B2B, city keywords are marginal: in the Italian data we analysed they make up 1-2% of the organic traffic of Macropix, Elesa and Camozzi. The campaign covers the whole country (or Europe) and no city goes into the keywords.
In local services it's the opposite. "Accountant london", "builders london", "plumber london" have the city built in.
There the campaign is set on a radius or a list of towns, with the "presence" option and NOT "presence or interest", which by default spends on people in Scotland searching for London.
The city also changes the price: "builders manchester" costs £2.29, "builders london" £5.47.
Contesini has two branches and two campaigns, one per branch, each on its own radius. Because for a local business every click from outside the area is wasted: a €7 lead from outside the area is worth nothing.
And the competition in B2B?
Lower than you'd think.
In our desktop scans of "packaging machines" and "cnc machining services" no ads appeared.
A single scan is a snapshot, so it doesn't mean nobody runs ads on those queries. It does mean advertising competition is lower than in local services, where the top-of-page bid for "emergency plumber london" reaches £55.82.
If you also sell into a second country, set it up separately: a separate account or at least separate campaigns. For our clients in Ticino we run a separate account in Swiss francs targeted on the canton. The keywords are the same as in Italy, the CPCs and competition are not, and a mixed account blends two markets that convert differently.
Further reading: Google Ads for B2B companies and how to define your ideal customer before choosing your targeting.
Keywords, match types and negatives (you'll find them on the results page before spending a penny)
Start with phrase and exact match. Broad match, which Google suggests together with smart bidding (automated bids), on a new account with no conversion history buys searches that have little to do with your product. And with £20-30 a day there's no room to find that out slowly.
You can pull negatives from the results page before you even start. Three examples.
"Packaging machines". The first page includes Machineseeker, a marketplace for used machinery, and one of the "People also ask" questions is "Where can I find packaging machinery for sale?".
For anyone selling new machines: "used", "second hand", "auction" are negatives.
"Cnc machining services". Three of the four "People also ask" questions are about the hourly rate.
People comparing prices. Worth knowing before you write the ad.
Industrial trades in general. The usual negatives go in on day one: "jobs", "salary", "apprenticeship", "course".
Then the search terms report, every week for the first two months.
Doesn't the algorithm do that?
No. It's the work no automation does for you, and in the maths above it's often the difference between a cost per enquiry that pays and one that doesn't.
Which page the ads should land on (not the home page)
Ads land on one page per campaign, never on the home page: the same service as the ad in the headline, the clickable phone number at the top if the conversion is a call, and a form that asks what your sales team would ask on the first call anyway.
For Valore Solare: area and type of roof. For an engineering firm: quantity, material and lead time.
Doesn't a longer form lose contacts?
Yes, some. It lowers the conversion rate and raises quality.
On Contesini 68% of leads became an opportunity. And that number matters more than cost per lead.
If the website is what you have and there's no time for a new landing page, the existing service page will do, as long as it has a form above the fold and no menu leading elsewhere.
How to build a landing page for a long sales cycle is in B2B landing pages; if traffic arrives and enquiries don't, the diagnosis is in why your website does not convert.
How to count real conversions (if they exceed clicks, tracking is broken)
If conversions exceed clicks, tracking is broken. Valore Solare had 15,267 conversions and 5,467 clicks in the same month, and even 1,487 conversions at €0.61 as on Sunscape are a warning sign: no B2B sector produces enquiries for less than a euro.
The conversions we count in a small business account are three:
- form submission, measured on the thank-you page or on a confirmed submit event;
- a call from the ad or from the number on the page, with a minimum duration of 30-60 seconds;
- a click on WhatsApp or email, where these are real channels.
Everything else (scroll, time on page, contact page views, menu clicks) stays as a secondary event, outside optimisation.
Won't the numbers collapse?
They will. That's good news.
On Sunscape, once tracking was rebuilt, 2026 counts 144 enquiries. Fewer than the 1,487 "conversions" of 2024, but these are people who asked for a quote.
You check it by hand: ten test submissions, ten rows expected in the CRM and in Google Ads, with the same timestamp.
The next step is importing offline conversions (the lead that becomes an opportunity). On Contesini it's what showed that 68% of leads were worth something. The basics of measurement are in how to use Google Analytics.
Do your conversions outnumber your clicks?
Valore Solare counted 15,267 against 5,467 clicks in the same month, and every automated bid was working on that data. In 30 minutes we'll check together what your account is really counting and what a real enquiry costs you.
Book a strategy consultationWhat is a sensible minimum budget for a small business (at least 10-15 enquiries a month)
The minimum budget is the one that produces enough enquiries to tell chance from signal: in our accounts, at least 10-15 enquiries a month. Below that threshold a good month and a bad month blur together, and automated bidding has no data to learn from.
The maths, sector by sector:
- Dental practice in Manchester: CPC £6.37, 5% rate, 10 enquiries = 200 clicks = about £1,270 a month.
- Builder: in Manchester, CPC £2.29, 10 enquiries at 5% = about £460 a month; the same builder in London, at £5.47, needs about £1,090.
- Engineering subcontractor, UK-wide: at a realistic B2B rate of 2%, 10 enquiries = 500 clicks. On "pneumatic components"-type terms at £2.84 that's about £1,420 a month, and you need 30-50 keywords to find 500 clicks in niches with 50-170 searches per keyword. On "cnc machining services" at £23.30 it would be over £11,000, which is why that term is left to the platforms.
- Manufacturer with a category like "led wall": CPC £3.03, 2% rate, 10 enquiries = about £1,500 a month, but with 1,000 searches a month no single keyword yields 500 clicks: you need "led video wall" (720) and the related terms too. Macropix, on the Italian equivalent, has managed €175,295 over eight years, about €1,800 a month, for 2,906 enquiries.
- Local branch with the phone call as the conversion: Contesini got 43 leads a month on about €300 of spend. With a 10% rate and a CPC around £2, £300-500 a month is enough.
So when do you raise it?
When cost per enquiry is stable and the enquiries are being followed up. Not before.
Valore Solare went from €2,000 to €14,000 a month in 11 months, with cost per enquiry down 40% compared with the first quarter.
Raising the budget in the first month, when an enquiry costs €212, burns money to learn the same lesson faster.
Management comes on top of the media budget (the spend on Google).
- Below £500 a month of spend a consultant costs more than the campaign: better to learn to run it yourself with a simple Search campaign.
- Above £1,500 the management cost pays for itself through negatives and bids, as the maths above shows.
For market-wide figures we have collected the Google Ads statistics 2026.
The mistakes we find in small business accounts (and what fixing them is worth: from €94 to €39.44)
The mistakes we see most often are eight, and nearly all of them are about tracking and structure, not budget. In order of frequency, in the accounts we have inherited in recent years:
- conversions that count technical events, with a cost per conversion under one euro that makes everything look fine (Sunscape, Valore Solare);
- one single campaign for all services, with the budget ending up on the most expensive keyword rather than the one that pays;
- broad match plus "maximise conversions" from day one, with no history;
- "presence or interest" targeting on a local campaign;
- automated recommendations applied in bulk: Google suggests them every week and some broaden keywords or switch on the Display network without anyone deciding to;
- ads landing on the home page, or on a page with no form;
- no negatives after the first month, with the search terms report never opened;
- campaign switched off after three weeks because "it didn't work", in other words before there were 30 enquiries to learn anything from.
What's fixing them worth?
On Macropix the cost per enquiry went from €94 to €39.44 between 2024 and 2025.
It came from three of these points (negatives, structure by product, clean conversions), not from a new Google feature.
On the change under way in the results pages, with ads inside and around the AI Overview, we wrote Google Ads and AI Overviews.
Paying more than €94 for an enquiry?
On Macropix, negatives, structure by product and clean conversions brought the cost per enquiry down to €39.44 in a year. Let's look together at which of these eight mistakes are in your account.
Book a strategy consultationGoogle Ads or SEO for a small business? (both, in this order)
Google Ads brings enquiries from the first week and stops the day the budget stops; SEO takes from six months to a few years and then stays. Sunscape went from 1 to 38 keywords on page one in six months; Macropix on "ledwall" in Italy took a few years.
On Macropix the two channels have run side by side for years. In 2025 Ads delivered enquiries at €39.44 while organic held 79 keywords out of 135 on page one.
For a small business starting from scratch, the sequence that works is this:
- Ads straight away on commercial keywords;
- SEO in parallel on product and service pages;
- after 12 months, move the Ads budget to the keywords organic can't reach.
The organic side is in the companion guide SEO for small businesses, with guides for each sector; how the two share the work is in SEO and Google Ads together.
For sectors with their own rules there are dedicated guides: construction companies, freelancers and professional services, payroll consultants, dentists, ecommerce, and for charities Google Ad Grants.
For estate agents, the reasoning on who to target (sellers, not buyers) and on minimum budget is worked through with real CPCs in Google Ads for real estate.
Frequently asked questions
It depends: on the sector's CPC and on how many enquiries you need to make decisions. With September 2026 UK CPCs, a local trade outside London sits at £300-500 a month, a dental practice or a London builder around £1,000-1,300, an accountant or professional firm in London about £1,900, an industrial business on £3-5 keywords between £1,500 and £2,500. Contesini got 43 leads a month on about €300.
Yes, it's worth it for small companies too, if they have a product people search for by name and a margin that can carry the cost per enquiry. Macropix, a small business in Milan, paid €39.44 for a quote request for an LED wall in 2025. It isn't worth it for someone selling a product nobody searches for yet, or who can't reply to enquiries within a day.
It depends: later yes, at the start no. It needs verified tracking and at least 30 real conversions a month, otherwise it optimises on the wrong data. In the first months a Search campaign with phrase and exact match gives more control and shows which searches bring enquiries.
It depends on what you measure: the first enquiries arrive in the first week, a stable cost per enquiry after 2-3 months. Valore Solare went from €212 per enquiry in the first month to an average of €70.78 over 11 months; Macropix's cost per enquiry fell from €94 to €39.44 in a year of work on the account.
It depends on the spend. Below £500 a month, yourself, with a simple Search campaign and the search terms report opened every week. Above £1,500, management pays for itself through negatives, structure and tracking: the three changes that more than halved the cost per enquiry on Macropix. In between, it depends on how much time you have.
Where to start (with five searches and a calculator)
Not the lowest CPC, but the cost per enquiry: from £38 to £718 across the 12 searches in the table.
Not 15,267 conversions, but 144 real enquiries.
Not Performance Max on day one, but Search for two or three months.
Not a big budget straight away, but once the enquiry cost settles: from €212 to €70.78.
Now, two concrete things.
One: take the five searches a customer would use to find you, read the CPC in the Keyword Planner and divide it by 0.05 (or by 0.02 if you sell B2B). That's the cost of an enquiry. Multiply it by ten and you have your first month's budget.
Two: if that number is lower than the margin on a sale, open the account starting with tracking. Ten test submissions, ten rows in the CRM.
If you'd like us to do the maths on your data, book a strategy consultation.
See you next time!
