Why Are My Google Ads Not Working? A Contractor's Diagnostic Guide

Your ads are running, the money is going out, and the phone is quiet. How to tell which of three very different problems you actually have.

By Gavin Sevastian · Updated August 4, 2026 · 14 min read

The money leaves your account every day and the phone stays quiet. If you have been asking why your Google Ads are not working, the maddening part is that the dashboard looks about the same whether your ads never appeared at all, appeared to people who were never going to hire you, or worked exactly as intended while nobody counted the call.

Most articles treat that as one problem. It is three, and they have almost nothing in common. Fixing the wrong one is worse than doing nothing, because you end up rebuilding a working account on the strength of a number that was never measuring what you thought it was.

What follows is a diagnostic, run in the order a specialist would run it. Cheapest checks first, rarest causes last. Stop as soon as you get a clear no. The first three checks take about ten minutes. A full pass through everything below is an afternoon.

If you are new to running ads at all, the beginner's guide to Google Ads for a small business covers the ground underneath this page.

1. Start Here: The Ten Minute Check That Tells You Which Problem You Have

Three checks, no platform expertise required. They will not fix anything. They tell you which of the three failures you are looking at, which is the part everyone skips.

1. 1

   Put your own phone records next to the Google Ads conversion count

   Same month, same dates. You are looking for one thing: did the phone ring on days the dashboard says nothing happened.

2. 2

   Run the Ad preview and diagnosis tool

   This is Google's own tool for checking whether your ad is appearing. Do not Google yourself instead.

3. 3

   Check whether call reporting is switched on

   It lives in your account settings. If it is off, calls are not being tracked at all.

1.1 Compare your own phone log against what Google says it counted

This single comparison splits the measurement problem away from the other two, and it costs nothing.

Pull last month's calls from your phone system or your notes. An actual count by date, not a rough sense of it. Put that next to the conversion count in Google Ads for the same dates.

If the phone rang on days the dashboard shows zero, you do not have an advertising problem. You have a counting problem, and every keyword change you were about to make would have been made against blind data.

It is easy to skip this because it feels too simple to be worth doing. It needs no login and no expertise, and it decides everything that comes after.

1.2 Look up your ads with Google's tool, not by searching your own name

Google built a tool for exactly this question. It "helps you understand why your ad or ad asset might not be appearing," and it shows you what your ad looks like on a results page.

It also protects your own data. Using the tool "avoids accumulating ad impressions and impacting your performance statistics." If you have been searching your own trade and city to check the ad is up, you have been logging impressions nobody clicked and dragging down your own click-through rate.

Use the Ad preview and diagnosis tool instead. It answers the question properly and leaves no mark.

1.3 Check whether call reporting is switched on at all

Open your account settings and find the call reporting section. Turning it on "uses a Google forwarding number to track calls as conversions. Turning it off shows your direct business number but prevents Google Ads from tracking calls."

That second half is the trap. Owners turn this off because they do not like an unfamiliar number appearing on their own ads, then spend months wondering why the account shows no calls. Both things are true at once. The phone is ringing, and Google is not tracking it. Details are on Google's call reporting page.

Where to go next, based on what you found:

• Barely any impressions, spend well under budget, or the preview tool shows nothing: go to Failure One.
• Plenty of clicks and spend, almost no calls: go to Failure Two.
• Your phone log disagrees with the dashboard, or call reporting was off: go to Failure Three.

2. Why Are My Google Ads Not Working? Three Failures, One Identical Dashboard

The headline numbers in Google Ads (spend, clicks, conversions) sit downstream of all three failures. That is why they cannot tell the failures apart, and why staring at them harder does not help.

2.1 The ads are not showing

A delivery failure. You cannot have a targeting problem if you were never in the auction, so nothing downstream matters until this is cleared.

2.2 The ads are showing to the wrong people

A matching failure, and the one that eats the most time. Deciding which searches are worth buying is judgment, not a setting.

2.3 The ads are working and the calls are not counted or not answered

A measurement or operations failure. The dashboard and the owner are both telling the truth, about different things. Most articles fold this branch into the other two.

The branch matters more than any individual fix. Raising your budget is the correct lever in the first branch, a waste of money in the second, and actively harmful in the third, because you are scaling spend against data you already know is wrong.

3. Failure One: Your Ads Are Not Showing

Work these in order of how cheap they are to rule out. The last two are invisible from any campaign report, which is why they belong on a list and not in your memory.

3.1 The budget is gone before lunch

Google publishes two ceilings. For most campaigns, the daily spending limit is two times your average daily budget on any given day, and the monthly limit is 30.4 times it. The mechanics are on Google's page about average daily budgets.

On a small budget the effect is blunt. A busy Monday can spend double, and then your ads stop for the rest of the day. Nothing on the overview screen says you were absent from eleven in the morning onward, which is precisely when a homeowner with a leak is searching.

Google says the same thing from the other direction on its diagnosis page: a preview of your ad "may no longer appear in the tool when your campaign reaches its average daily budget."

If you are not sure whether your daily number is realistic for your trade and area, start with what Google Ads actually cost for a contractor.

3.2 Eligible is not one status, it is four

In Google's developer documentation, an ad's policy approval state is one of four values:

• Approved. "Serves without restrictions."
• Approved (Limited). "Serves with restrictions."
• Area of interest only. "Will not serve in targeted countries, but may serve for users who are searching for information about the targeted countries."
• Disapproved. "Will not serve."

The middle two are the ones that catch people. A limited ad serves, but not the way you assumed it does. An ad in the area of interest state does not serve in the countries you targeted, and may instead serve to people elsewhere who are searching about them.

Both spend your budget. Both look live at a glance. Google lists all four in its policy approval status reference.

3.3 The boring administrative causes, which are the most common

Run down this list before you touch anything strategic. Each is binary and takes under a minute:

• The card on file was declined.
• The campaign or ad group is paused, or the ad itself is.
• An ad was disapproved on policy grounds and nobody noticed.
• The ad schedule is set to hours when nobody is searching.
• Keywords are sitting in a low search volume state, frozen out of auctions at zero cost with no error anyone would spot.

Then two that no campaign report will surface. Ads "won't serve if your account is undergoing billing verification or has been temporarily suspended due to security concerns, like an account takeover." Both are on the same diagnosis page.

If every setting in your campaigns checks out and delivery is still dead, look at the account and billing area before you rebuild anything.

4. Why Bidding More Does Not Guarantee You Show Up

If you believe a higher bid means you appear, you will never look at anything else on this page. It is the most common wrong mental model in this branch, and correcting it is worth three minutes.

4.1 Ad Rank decides whether you show, then where

Google calculates Ad Rank twice. It "is calculated first to determine whether your ad is eligible to show, and a second time to determine where your ad is ranked relative to other eligible ads."

An ad can lose the auction completely, not merely place lower down the page.

Bid is one input. Ad and landing page quality, the rank thresholds your ad has to clear, how competitive the auction is, and the context of the search all feed in too. A competitor bidding less than you can appear when you do not. Google explains the inputs on its Ad Rank page.

4.2 Impression share turns "am I showing enough?" into a number

You do not have to guess at this. Impression share is the impressions you received divided by the total impressions you were eligible to receive, and Google reports it in the account.

The value for you is that under-delivery stops being a feeling. A low figure sends you back to the previous section, to budget and approval status and targeting, instead of into your keyword list looking for something to change. The definition is on Google's impression share page.

4.3 Why raising the bid is usually the third thing to try, not the first

Quality inputs and delivery blockers are cheaper to fix and are far more often the actual constraint. A disapproved ad, a card that failed, a schedule that runs at the wrong hours, a location setting pointed at the wrong people: none of those get better when you pay more per click, and all of them cost nothing to correct.

Bidding is a real lever and there are accounts where it is the right one. It is just rarely the right one first. For the fuller picture of how bidding decisions work, read the guide on choosing a Google Ads bid strategy.

5. Failure Two: Your Ads Are Showing to People Who Will Never Call

This is the branch where everything looks healthy. The budget gets spent, the click count is respectable, the cost per click is not outrageous. And none of that traffic was ever going to book a job.

There is exactly one report that tells you this, and plenty of contractors have never opened it.

1. 1

   Open the search terms report in your Google Ads account

   Set the date range to the last 30 or 90 days, so you are looking at a period long enough to show a pattern.

2. 2

   Sort by cost, highest first

   Not by clicks, not by impressions. You want to see what your money actually bought, in order of how much of it went there.

3. 3

   Write down every term that could never become a booked job

   Do not change anything yet. Just build the list. Reading it honestly is worth more than any single fix you could make this afternoon.

5.1 The search terms report is the difference between what you bought and what you got

Your keywords are a hypothesis. They are your guess about what a customer is going to type. Search terms are what people actually typed before your ad appeared, and the two lists are rarely as close as owners expect.

Google calls it "a list of search terms that a significant number of people have used, and that resulted in your ad being shown." The full description is on Google's search terms report page.

Sitting with this report will teach you more about your own account than a month of reading articles about Google Ads. It is also uncomfortable the first time, because a lot of accounts are paying for searches the owner would never have chosen if anyone had asked.

5.2 The report is incomplete by design, and that is permanent

Google withholds some of it. In Google's own words, "some search terms that don't have enough query activity are omitted from the search terms report in order to keep with our standards on data privacy." What you get back are the terms used by a significant number of people.

For a national retailer that omission barely registers. For a Mississauga HVAC company it is a different story. Local searches are long, conversational and often unique. Somebody types a full sentence describing the noise their furnace is making, once, and nobody else in the country types that exact string. Those are the queries that sit below the threshold and never surface.

Absence from the report is not evidence that a search never happened. It is not a reporting delay either. It does not clear as the account ages or as you spend more, and Google publishes no number for where the cutoff sits.

5.3 No match type is literal, including exact

No match type means quite what its name suggests. All of them expand to close variants, and that behaviour cannot be switched off. Exact match is not exact.

Negative keywords are your main defence, and they work in the opposite direction to positive ones. Positive keywords expand: buy one and Google matches things it judges related. Negatives do not. Block a term and you have blocked that term, not the family it belongs to.

That asymmetry is why exclusion work is laborious and not clever, and why the list is never actually finished.

6. Location Targeting Is the Setting Most Often Quietly Wrong

If your business can only serve people you can physically drive to, this one is worth checking today.

6.1 The default includes people who are merely interested in your area

The setting is documented in plain terms: "The default advanced location option in Google Ads will use both physical location (where a user is located or is regularly located) and / or location of interest to determine where ads can appear."

The default is not "people in Toronto." It is people in Toronto plus people interested in Toronto. Google presents presence or interest as the recommended option and documents the whole thing openly on its location targeting page.

There is no conspiracy here, and it is worth saying so because a lot of articles reach for one. That default is tuned for every advertiser on the platform, and many of them genuinely do want people who are merely interested in a place.

A trade that has to load a van does not. Changing it is covered on Google's page about targeting ads to geographic locations.

6.2 How Google infers interest, and why it reaches further than you think

Interest is inferred from behaviour, including what Google describes as past searches that indicated a location of interest. It is not about where somebody is. It is about what they have shown curiosity in.

Then the line that surprises most owners. Location of interest "isn't restricted to a person's country, or the Google search domain the person is searching on."

Concretely: a Toronto plumber on default settings can pay for a click from somebody in Calgary who was searching about Toronto last week, or from somebody overseas researching a move. The click is real, the cost is real, and the person cannot hire him. For more on what this market looks like specifically, the Toronto page covers it.

6.3 Where this compounds with an approval status

Put the location default next to the approval statuses from section three. Interest targeting widens who can see the ad. The area of interest only status, as Google defines it, narrows serving away from the countries you targeted and toward people searching about them.

They are different mechanisms at different scales, and they push in the same unhelpful direction: money going to people whose relationship with your service area is curiosity rather than proximity. Neither shows up as a problem in the headline numbers, which is the whole reason this section exists.

7. Who Is Actually Clicking Your Contractor Ads

The wrong-people categories are predictable enough that you can draft most of an exclusion list before opening a single report.

7.1 People who want to do the job themselves, or buy the part

DIY searchers, how-to searchers, and people shopping for a product instead of for a person. All of them are topically about your trade. None of them are calling a contractor.

Those two groups look similar in a keyword list and behave completely differently. Somebody pricing a unit is comparing products and may well fit it himself. Somebody with a job that needs doing is looking for someone to do it. Same trade, same vocabulary, different person at the other end.

7.2 People who want a job, not a quote

Salary, apprenticeship, hiring, training, certification, licensing. Somebody searching what a roofer earns in Ontario is not about to book a roof.

These are cheap to block, they arrive constantly, and they never stop arriving, which makes them the first exclusions worth building and the ones that keep paying.

7.3 People in the wrong place or the wrong trade

Out-of-area searchers, which is the section above and is a settings fix instead of a keyword fix.

Then adjacent trades, where the ad is topically close and the job is not one you do. A roofing ad picking up siding and eavestrough work. A plumbing ad catching basement waterproofing. An HVAC ad pulling appliance repair. Every one of those is a defensible match on paper and a wasted click in the bank account.

The honest part: this is the branch that consumes the week. Deciding which searches are worth buying, in what form, at what price, from a report that is incomplete by design, is a judgment call and not a task, and it has to be made again as the market moves. The categories above get you started. They do not get you finished.

8. What Normal Looks Like, So You Know If You Are Actually Broken

Before you conclude the channel does not work, it is worth knowing roughly what published figures look like for this kind of business. Not so you can chase a number, but so you can tell the difference between an account that is underperforming and an account performing normally in an expensive category.

8.1 Published paid search benchmarks for home improvement

LocaliQ publishes search advertising benchmarks by category, and the row that fits most of the trades reading this is Home and Home Improvement. Their stated method is that the figures come from over 16,000 of their own customers' campaigns running April 2024 through March 2025, across Google Ads and Microsoft Ads. The numbers are in the Search Advertising Benchmarks report.

Two caveats belong beside every one of those figures. Home and Home Improvement is a single bucket, so it cannot tell you anything about your trade specifically. And the figures are US dollars, which is why they are labelled that way in the table above and not converted.

Use them as a sanity check on whether your account is in a normal range for an expensive category. Do not use them as a target.

8.2 Why a bad week is often not a signal

At the click volumes a single local contractor generates, a run of zero calls is frequently ordinary variance and not evidence of anything. Small numbers are noisy. A quiet fortnight followed by a busy one is what small numbers look like, and reacting to the quiet fortnight is how working accounts get taken apart.

The opposite mistake is just as expensive, and I am not going to pretend otherwise by telling you to be patient. Sometimes a dry spell is exactly what it looks like, and waiting another month costs you a month.

Telling those apart means looking at your actual click counts, your conversion history and your season. No article can do that for you.

What you can do is check the economics underneath it. If your budget only ever buys a handful of clicks a week, then a week with no calls was always likely, and the conversation you need is about what Google Ads actually cost for a contractor instead of about the campaign.

9. Failure Three: The Phone Rang and Google Did Not Count It

This is the branch that gets folded into the other two almost everywhere you read about it, and it is the one where a wrong conclusion does the most damage. The fix is not in the campaign at all. It is in the plumbing underneath, and until you check it you cannot trust a single number you are making decisions from.

1. 1

   Confirm call reporting is on

   It sits in your account settings. If this is off, nothing below this line is able to work.

2. 2

   Confirm the Calls from ads conversion action is active

   Google says to check that it is active and set as a primary goal, not a secondary one.

3. 3

   Check what the minimum call length is set to

   Open that conversion action and look at the call length setting. Any call shorter than this number produces no conversion.

4. 4

   Open the Call details report in Report Editor

   Google names this report specifically. It shows actual call durations instead of only the calls that qualified as conversions.

9.1 There are two separate call tracking paths and having one does not give you the other

Google tracks calls through more than one mechanism, and they have different setup requirements. Calls from ads depends on call reporting and a Google forwarding number being assigned to your ad. Calls to a phone number on your website is a separate path that also needs the Google tag plus a specific event snippet for website call conversions.

All of it rests on the same foundation: "Tracking call conversions from ads and from a website and importing call conversions require Google forwarding numbers." The details are on Google's page about phone call conversion tracking.

The practical consequence is that having one path working tells you nothing about the other. It is entirely normal to find calls from ads counted properly while every call from the website is invisible, or the reverse. Nothing in the interface announces the gap. You just see a conversion number quietly missing half its input.

To check the website side, Google points to Tag Assistant. An "Inactive Tag" status usually means the tag or the snippet was never implemented correctly, whatever your web developer told you at the time.

9.2 A call that rings short produces no conversion at all

Google's definition, in full: "You set a minimum call length, and every call that lasts at least that long is counted as a conversion."

Consider what that means for a short call. It is not recorded as a weak conversion or a low quality one. There is no conversion.

A homeowner rings, your apprentice picks up, tells them you will call back after the current job, and it is over in forty seconds. That is a real lead from a real ad, and depending on your setting it produces no conversion anywhere in your account.

Google names the report that resolves this: Call details, in Report Editor. It shows actual durations instead of only the calls that cleared the bar.

9.3 One of the tracking methods is an estimate, not a count

There is a version of call tracking that runs without a forwarding number. Clicks on call ads and assets can be tracked directly, in which case "a click on call ad or asset counts as a conversion based on Google's estimation of whether a meaningful phone call occurred after a person clicked your ad."

One of the numbers in your account may therefore be a model output and not a tally of things that happened. That explains dashboard gaps in both directions: counted calls you have no record of, and real calls that were never counted.

Then the objection almost every Ontario contractor raises when a tracking number comes up, which is that they will not use one because they need to see who is calling.

Google forwarding numbers are available in Canada, and Google states that "Caller ID still works for calls routed through Google forwarding numbers (except for some calls in India and Japan) so you can know who's calling you." That is on the call reporting page.

The number your customer dials changes. What appears on your phone when it rings does not.

10. The Calls You Never Answered Look Exactly Like Calls That Never Came

Lead handling is not a separate topic sitting outside your advertising. It is inside the loop, and it presents to you as an ad problem.

10.1 Voicemail, missed calls, and after hours

To your reporting, to your dashboard, and to the bidding system, a missed call and a call that never happened are the same event. There is no column anywhere that says you were paid for this click and then nobody picked up.

You will find plenty of articles quoting a percentage of calls that small businesses miss. Most of those figures come from companies selling call tracking software, which makes them marketing and not measurement, so I am not going to repeat one at you.

Measure your own instead. Call reporting records whether each call was connected, so your actual answer rate is a number you can have this afternoon rather than one you have to borrow.

10.2 Your ad schedule and your phone coverage have to agree

If your ads run at seven in the evening and nobody answers the phone after five, you are buying clicks at their full price and converting none of them. This looks exactly like ads that do not work.

Either extend the coverage or narrow the schedule. Both are legitimate. Doing neither, while paying for both, is the expensive option.

10.3 How fast you call back is part of the ad's performance

There is a well known piece of research on this in the Harvard Business Review, The Short Life of Online Sales Leads. Its finding, in the authors' own summary, is that most companies are not responding nearly fast enough to the online enquiries they receive.

A homeowner with water coming through the ceiling is unlikely to be sitting and waiting for one callback. Whatever your response time currently is, it is a variable in how well your advertising performs, and it is one of the few on this page you control completely.

11. Why Bad Data Compounds

The instinct is to treat tracking as a technical detail you will sort out once the campaign is producing. Invert that. Tracking is not the thing you check to see if the account is working. It is one of the things making the account work or not work.

11.1 Automated bidding optimises toward what it can see

Automated bidding learns from the conversions it is shown. If calls are invisible to it and form fills are not, it will buy more of the traffic that produces form fills, because as far as it can tell that is the only thing your advertising ever achieves.

Over time that reshapes the whole account toward the wrong customer, and no individual setting will reveal it. Everything looks configured correctly. It is optimising perfectly toward a picture of your business missing half the evidence. There is more on how these systems behave in the guide to choosing a Google Ads bid strategy.

Which is also why adding budget makes this worse. If your measurement is wrong, more money buys more of whatever the wrong measurement is pointing at. The leak scales with the spend.

11.2 Changing things every few days guarantees you measure noise

Google publishes a calibration window: after a bidding change "it can take up to 3 weeks or 1-2 conversion cycles for the bid strategy to calibrate to the new objective." That is on Google's page about bid strategy changes.

An owner who adjusts something every Tuesday because the week was quiet never gets past that window. The system is permanently recalibrating, the data is permanently uncalibrated, and every conclusion drawn from it is about the adjustment and not about the market.

12. The Rare Causes, and Why They Belong Last

These are the explanations owners reach for first. They are all real. They also sit near the bottom of the list, and starting here means chasing the rarest cause while an ordinary location setting sits misconfigured.

12.1 Click fraud

Popular, and occasionally correct. Usually not the reason your phone is quiet. Google filters invalid clicks automatically.

12.2 Restricted categories and account level holds

Some trades and some countries face advertising restrictions that block ad or keyword approval outright. Worth naming, because when it applies it explains everything at once and no amount of campaign work will move it.

The billing verification and security suspension states from section three belong in the same bucket: rare, real, and invisible unless you go and look.

12.3 Nobody is searching for my service

Usually false, and worth naming because it is the conclusion contractors arrive at when the actual cause was a delivery failure or a setup too narrow to reach anybody in the first place.

Genuine absence of demand does exist. Very specialised work in a small market is a real diagnosis. It just belongs at the bottom of the list, after you have confirmed you were actually in the auction.

13. Can You Fix This Yourself, or Is This a Full Time Job?

Honest answer: a meaningful part of this you should do yourself, and I would rather you did. The seven-item list at the end of this page is that part, and it costs nothing. The rest is not hard in the sense of being complicated. It is hard in the sense of never being finished.

13.1 Why this resists a checklist

The individual checks are easy. The problem is where they live. The full diagnosis spans nine separate surfaces: a preview tool, a budget report, a policy column, a settings page, a query report, an account setting, a conversion action, a tag debugger, and a billing page. There is no single view that shows you all of them.

Worse, three of those nine give you a clean-looking answer while something is genuinely wrong, and you have already met all three above. In each case the interface tells you something reassuring and accurate, and the reassuring accurate thing is not what you asked.

Then add the calibration lag, which means you wait weeks to learn whether a change helped, and low volume, which means the answer you get after waiting is often ambiguous anyway.

And it does not stay fixed. Search behaviour shifts, competitors enter and leave the auction, seasons move the mix of what people need, and an exclusion list that was right in March is leaking by August. This is recurring work and not a repair, which is the honest reason it becomes somebody's job.

13.2 And a fair question: should you be advertising right now at all?

Search is not the only paid option Google offers a home service business, and the comparison is worth making instead of assuming. There is a fuller breakdown in the guide to Local Services Ads for contractors.

Sometimes the right answer is neither, for now. If you cannot answer the phone reliably, or you are already booked six weeks out, buying more demand is solving a problem you do not have. Advertising adds demand. It does not fix what happens after the phone rings.

If you are weighing the cost of doing this with help, what Google Ads agencies charge sets out the shapes those arrangements take.

14. What To Do This Week

Work these in order and stop as soon as one of them gives you a clear answer. The earlier items are the cheapest to rule out and the most often responsible.

1. Compare last month's phone log against the conversion count for the same dates.
2. Run the Ad preview and diagnosis tool, and stop Googling yourself.
3. Check whether call reporting is switched on in your account settings.
4. Check budget pacing and the approval status on every active ad.
5. Set location targeting to presence if you can only serve people you can drive to.
6. Open the search terms report, sort by cost, and read the top fifty.
7. Open the Call details report and find out what is actually happening on the phone.

If you would rather not do this yourself

That is a reasonable position, and it is what we do at AdClaw. If you want somebody to run the diagnosis properly and tell you which of the three problems you actually have, start here. No obligation, and if the answer turns out to be one of the seven things above, we will just tell you that.