Google Ads has a reputation problem with small business owners, and it's mostly earned. Almost everyone knows somebody who "tried Google Ads," spent a few thousand dollars, got a handful of junk calls, and swore off the whole thing.
Here's the uncomfortable truth behind most of those stories: the platform did exactly what it was configured to do. The configuration was just wrong. Google's defaults are designed to spend your budget broadly, the interface nudges you toward automation you can't see inside of, and a new account with no guardrails will cheerfully buy clicks from people who were never going to hire anyone.
The fix isn't a secret strategy. It's a short list of boring disciplines, plus the honesty to recognize when ads are the wrong move entirely. This guide covers both.
Start with search campaigns only. Ignore everything else.
When you create a campaign at ads.google.com, you'll be offered a menu: Performance Max, Display, Video, Demand Gen, Search. For a small service business starting out, the answer is Search, and only Search.
The reason is intent. A search ad shows up when someone types "water heater replacement wilmington nc." That person has a problem, today, and is actively looking for someone to pay. Every other campaign type interrupts people who weren't looking: banner ads on recipe sites, videos before YouTube clips, feeds and placements across Google's network. Interruption advertising can work for big brands with big budgets and patience. For a small business that needs the phone to ring this month, intent is the only game worth paying for.
Two specific traps to sidestep:
- Performance Max. Google pushes it hard, and it bundles search, display, video, and more into one automated black box. It can perform once it has lots of conversion data to learn from. A brand-new account has none, which means PMax learns by spending your money in places you can't fully see or control. Skip it until search alone is profitable and well-tracked.
- Default network checkboxes. Even inside a Search campaign, Google opts you into the Display Network and Search Partners by default. Uncheck both at setup. You want your budget on Google search results, period, until you have data that says otherwise.
While you're in setup: target only the geography you actually serve (and set location options to people in your targeted area, not people "interested in" it), and run ads only during hours someone will actually answer the phone. A missed call from an ad is the most expensive kind of missed call, because you paid for it twice.
Negative keywords: the discipline that separates winners from donors
Your ads don't just show for the keywords you pick. They show for whatever searches Google decides are related, and Google's definition of related is generous, because Google gets paid either way.
Negative keywords are your veto list: terms that block your ad from showing. Google's own documentation on negative keywords covers the mechanics; the strategy is simple. Before launch, load the obvious ones for your trade:
- "free" in most combinations, because free-seekers don't become customers.
- "jobs," "careers," "salary," "hiring," because job hunters search your service terms constantly.
- "DIY," "how to," "yourself," because researchers aren't buyers today.
- "cheap" if your positioning isn't cheap, and competitor brand names unless you're deliberately bidding on them.
- Services you don't offer and areas you don't serve, because adjacent searches will leak in.
Then, and this is the part most people skip, check the search terms report weekly, every week, forever. It shows the literal searches that triggered your ads. The first month's report is always humbling: real money spent on searches you'd never have approved. Every week you add the junk as negatives, and every week the account gets a little more efficient. This single twenty-minute habit is most of the difference between accounts that work and accounts that burn.
Landing page match: the click is only half the purchase
You don't pay for customers on Google Ads. You pay for clicks. Whether clicks become customers is decided entirely after the click, on your page, and this is where otherwise well-built campaigns die.
The rule is message match: the page someone lands on must obviously continue the search they just made. Someone who searched "emergency roof tarp service" and clicked an ad about emergency tarping should land on a page about exactly that, not your homepage, not a generic "our services" page that makes them hunt. Every second of hunting bleeds people who hit the back button, and you paid for every one of them. (For trades like roofing, where emergency intent and project intent are completely different mindsets, separate pages per intent is the whole ballgame.)
A landing page that deserves paid traffic has:
- A headline echoing the search. They should think "right place" in under two seconds.
- One obvious action. Tap-to-call number and a short form, visible without scrolling. Three fields beat eleven.
- Proof on the page. Reviews, photos of real local work, years in business, licenses. Strangers from ads have zero context on you; the page has to build trust from nothing.
- Speed on a phone. Most of these clicks are mobile. A page that takes ages to load on a phone in a driveway is a paid bounce.
There's a money reason beyond conversion, too: Google scores ad and landing page relevance, and that quality assessment affects what you pay per click. Better match literally lowers your prices. If your current site can't support pages like this, fix that before funding ads; it's the foundation our website and SEO work exists to build, and ads without it are pouring water into a cracked bucket.
Budget floors: the honest math nobody puts in the sales pitch
Here's the part many agencies soft-pedal: Google Ads has a practical minimum viable budget, and it's higher than most owners hope.
The logic is simple. Use Google's Keyword Planner inside your account to look up what clicks in your trade and city actually cost; for competitive local services, individual clicks are often the price of a decent dinner, sometimes much more. Then assume only a fraction of clicks become leads and only a fraction of leads become jobs, because that's reality even for good campaigns. Multiply it out and a small monthly budget buys you a handful of clicks, maybe a lead or two, and, statistically, nothing you can learn from. You can't tell a bad campaign from bad luck on ten clicks.
So be honest with yourself on two numbers:
- Monthly spend must buy enough clicks to generate real data, which for most local trades means a budget that would sting if it returned nothing for a month, because early on it might.
- You need runway for two to three months, because the first month is tuition: gathering search terms, cutting waste, fixing pages. Accounts usually get meaningfully better with this iteration, but only if you survive long enough to iterate.
If funding that runway means skipping payroll or rent, the answer isn't a smaller ads budget. It's no ads budget yet. The SBA's guidance on marketing budgets is a sane reference point for keeping total marketing spend proportionate to revenue. Ads should be fed from margin, not from hope.
When NOT to run Google Ads
This is the section that saves the most money. Don't run ads if:
- Your website can't convert. No mobile-friendly, fast, proof-loaded landing page means every click is rented traffic poured onto sand. Foundation first.
- You have little to no review base. Ad clicks are strangers. Strangers check reviews. A competitor with hundreds of reviews will harvest the trust your click paid to create.
- Nobody reliably answers the phone. If calls go to voicemail during business hours, fix operations before buying more calls. (This is also where automated follow-up earns its keep; speed-to-answer is most of paid-lead ROI.)
- The math doesn't math. If your average job's profit can't comfortably cover a realistic cost per acquired customer, ads are a machine for converting your savings into Google revenue. Thin-margin, low-ticket, one-time services struggle here; high-ticket and repeat-customer businesses have room.
- You're already booked out. A full schedule plus ads equals paying for calls you'll answer late or never. Raise prices instead; it's the free version of the same lever.
- You can't track results. Without call tracking and conversion tracking wired up, you'll be making budget decisions on vibes. Set up tracking before the first dollar, not after the first doubt.
None of these are permanent disqualifiers. They're sequencing. Most businesses that fail at ads didn't fail at ads; they ran ads third when ads should have been step six.
The boring playbook, in order
Build the converting page, build the review base, wire up tracking, then launch one tightly-geographed search campaign with aggressive negatives, check the search terms report weekly, and judge nothing for ninety days. It isn't exciting. It's just what working looks like.
Want the foundation handled before you spend a dollar on clicks?
That's our lane. Omnyra is a veteran-owned shop in Wilmington, NC, with 1,500+ small business sites built in the last 90 days. We build done-with-you websites live on a call with you: first draft in 24 hours, live in 7 days, guaranteed. Tiers start at $500, and the Max tier automates your follow-up cadence (5 texts and 3 emails per lead), so the clicks you eventually pay for actually turn into booked work. Pay-in-4 and Klarna available.