Why Your Google Ads ROI Is Broken (And It's Not Your Ads)
Owner-operators spend years optimising creative, keywords, and bidding strategies. Across the accounts we've reviewed, most of them are fixing the wrong stage of the funnel. Here's where the money actually leaks, with the math.
By Yasmin & Arsen | P&C Marketing Agency | Reading time: 8 minutes
Last quarter, an anonymised Saskatchewan roofing company asked us a simple question.
Why aren't we seeing more revenue from Google Ads?
The owner believed the channel was underperforming. He was spending $14,000 a month. Leads were coming in consistently. But revenue had been flat for almost a year.
On paper, the account looked healthy. Cost per lead was $87. Close rate was 18%. The agency report said the campaigns were performing within benchmark.
Then we asked one question.
How many of those leads are actually buyers?
The owner pulled his CRM. We pulled it up together. Of the previous 300 leads, 184 were tire-kickers, comparison shoppers, or people outside their service area. Real qualified leads: 116. Real cost per qualified lead: $225. Real close rate against qualified leads: 47%.
The agency report said cost per lead was $87 and close rate was 18%. Both numbers were technically correct. Both numbers were also useless.
His Google Ads account didn't have a creative problem. It didn't have a bidding problem. It had a Stage 1 problem — and nobody was looking at Stage 1.
The 5-Stage Pipeline, briefly
Before we go further, a quick frame. P&C's 5-Stage Pipeline Audit is the methodology we run every engagement through. The five stages are:
| Stage | What it covers | The question it answers |
|---|---|---|
| Stage 1: Traffic | Who you're paying to attract | Are the right people seeing us? |
| Stage 2: Capture | Landing pages and offers | Do they tell us who they are? |
| Stage 3: Qualification | Lead scoring and intake | Are they actually buyers? |
| Stage 4: Conversion | Sales process and follow-up | Do they close? |
| Stage 5: Tracking | Revenue attribution and reporting | Can we trace deposits to source? |
This post is about Stage 1. In a previous post we covered Stage 5. Together, they bracket the funnel. Get those two right and the middle stages get measurably easier.
Why Stage 1 is where the silent money goes
Most Google Ads accounts don't have a traffic problem. They have a filtering problem.
Stage 1 is where, in our experience reviewing SMB accounts, 30-60% of most Google Ads budgets quietly disappear. Not in a dramatic way. Not in a way that shows up on the dashboard. The clicks happen. The form-fills happen. The "conversions" register. The account looks fine.
But the leads that come through are wrong — wrong intent, wrong budget, wrong geography, wrong stage of decision. By the time anyone notices, three months of spend has gone to acquiring leads the sales team can't close.
Most agencies never catch this because most agencies don't look downstream of the click. They look at the platform metrics — clicks, conversions, cost per click — and report on those. The platform doesn't know which clicks became real customers. The agency doesn't ask.
Quick test: If you can't answer the question "How many of my last 20 leads were genuinely qualified buyers?" within five minutes, you probably have a Stage 1 visibility problem.
The four ways Google Ads leaks at Stage 1
Repeatedly across client accounts, the same four traffic-stage leaks appear. Each one tends to drain 10-25% of budget before the campaign has a chance to work, in the accounts we've reviewed. Fix all four and the economics of the account often change dramatically before a single new advert is written.
Leak 1: Broad match keywords without negative discipline
Broad match is Google's default setting and Google's most profitable setting — for Google. It lets the platform spend your budget on adjacent search terms it thinks are related to your keywords. Sometimes it's right. Often it's not.
The Saskatchewan roofing client had "metal roofing installation" set to broad match. Over three months, Google had matched that keyword to searches including:
| What Google matched (bad) | What we wanted matched (good) |
|---|---|
| metal roofing DIY | metal roof installation Saskatoon |
| metal roof colours | metal roof replacement quote |
| metal roof history | commercial metal roofing contractor |
| metal roof for chicken coop | emergency roof repair near me |
The left column was burning $4-9 per click. None of them were buyers. The right column was the actual demand. Same keyword, two completely different traffic profiles.
The fix takes 90 minutes. Pull the last 90 days of search terms report. Mark every term that isn't a buyer with high intent. Add those to the negative keyword list. Repeat weekly for a month. After that, monthly maintenance.
Across the SMB accounts we've worked on, this single fix typically reduces wasted spend by 15-30% within the first 60 days. Not from cutting budget — from removing the leaks in how the budget gets spent.
Leak 2: Geography that includes where you don't actually serve
Google's default location targeting is "people in or interested in your target locations." That second half — "interested in" — is where money disappears.
A Saskatoon roofing company doesn't need clicks from Regina, Calgary, Edmonton, Vancouver, or Toronto. Yet we've seen all five appear in location reports. Someone in Toronto searching about Saskatchewan businesses gets counted as in-target. They click. They're never going to buy from a Saskatoon roofer.
Here's what bad versus good location targeting looks like in practice:
| Setting | Default (leaking) | Fixed (tight) |
|---|---|---|
| Target type | Presence OR interest | Presence only |
| Radius | Province-wide | 50km from service base |
| Exclusions | None | Cities outside service area |
This one fix saved the roofing client $1,840 a month. We found it in their settings on day one. It had been quietly burning for two years.
Leak 3: No bid adjustments for device, time, or audience
If your call-driven business is bidding the same amount at 11pm on a Sunday as it is at 10am on a Tuesday, you're paying for clicks that can't convert. If your appointment-based business bids the same on mobile as desktop without checking which actually closes, you're guessing.
Every account we review has at least one of these settings on default. The data to fix them is in Google Ads, free, sitting there. Nobody pulls it because pulling it and acting on it isn't billable in the same way creative refreshes are.
The exercise: pull conversion data by hour-of-day, day-of-week, device, and audience segment. Anywhere conversion rate drops more than 40% below average, apply a negative bid adjustment. Anywhere it rises more than 40% above, apply a positive one.
Across the accounts we've worked on, this typically produces a 10-20% lift in cost per acquisition. Same spend. Better-allocated spend.
Leak 4: Bidding for clicks when you should be bidding for revenue
This is the leak that connects Stage 1 to Stage 5. Most SMB accounts bid on "Maximise Clicks" or "Maximise Conversions" — where conversions means form-fills, including the spam and the tire-kickers.
Google then optimises for what you told it to optimise for: more form-fills. Including the bad ones. Including the ones that waste your sales team's time.
The fix requires Stage 5 to already be working. You need offline conversion tracking installed, deposited revenue flowing back to Google Ads, and a "Maximise Conversion Value" bidding strategy with revenue as the conversion value. Then Google starts optimising for the leads that actually become customers, not the ones that fill out forms.
This is also why the four stages can't be fixed out of order. Stage 1 leaks are amplified by Stage 5 gaps. In our experience, Stage 1 fixes are roughly three times more effective when Google has real revenue data to optimise against.
What this looked like in real numbers
Back to the anonymised Saskatchewan roofing client. The initial diagnosis came from reviewing the previous 300 leads — that's where the 184 unqualified / 116 qualified breakdown comes from. The table below shows the first 60 days after the four Stage 1 fixes were implemented:
| Metric | 60 days before fixes | 60 days after fixes |
|---|---|---|
| Monthly spend | $14,000 | $14,000 |
| Total leads | 161 | 94 |
| Qualified leads | 62 | 81 |
| Cost per qualified lead | $225 | $173 |
| Closed deals | 29 | 38 |
| Revenue produced | $184,000 | $241,000 |
| ROAS | 13.1x | 17.2x |
Fewer total leads. More qualified leads. Same spend. 31% more revenue.
The agency they'd had before us would have looked at the "Total leads" row dropping from 161 to 94 and called it a problem. They'd have pushed for more budget to "rebuild volume." That's the entire industry playing the wrong game.
The right game is the bottom row. The wrong leads cost money to acquire and then cost money to disqualify. The right leads compound.
More leads is often the wrong answer. Better leads is usually the right one.
The 30-day Stage 1 fix
Here's the order. Don't skip steps.
Week 1: See what Google is actually doing with your money
Pull the search terms report for the last 90 days. Print it. Highlight every term that isn't a buyer-intent search. The percentage you highlight is your Stage 1 waste rate. Across the accounts we've worked on, anything above 25% is a major leak.
Then pull location data. Compare where your clicks come from to where your closed deals come from. If they don't match, your geo targeting is wrong.
Week 2: Negative keywords and geography
Add every highlighted search term to your negative keyword list. Switch location targeting to "presence only." Add explicit city exclusions for places you don't serve.
Repeatedly across client accounts, this single week of work reduces wasted spend by 15-25%. Same budget, redirected to actual buyers.
Week 3: Bid adjustments
Pull conversion data by device, day, hour, and audience. Apply negative bid adjustments where conversion rate is more than 40% below average. Apply positive ones where it's more than 40% above.
Most accounts we review need 8-12 bid adjustments. Most accounts have zero.
Week 4: Connect Stage 1 to Stage 5
This only works if Stage 5 is already fixed — if you have offline conversion tracking installed and deposited revenue feeding back to Google. If you don't, stop here and fix that first.
If you do, switch your bidding strategy to "Maximise Conversion Value" with deposited revenue as the conversion value. Then leave it alone for 30 days. The algorithm needs that long to learn what real revenue looks like in your account.
The boring truth, again
Google Ads ROI is mostly not a Google Ads problem.
It's a pipeline problem at Stage 1 (who you're paying to attract) compounded by a pipeline problem at Stage 5 (what you're telling Google to optimise for). The middle stages — landing pages, lead capture, qualification, sales conversion — matter, but they matter less than the brackets.
An owner-operator running a $500K-$5M business doesn't need a better Google Ads agency. They need someone willing to look at the account as one stage of a pipeline, not as the pipeline itself.
The Saskatchewan roofer didn't need more traffic.
He didn't need more leads.
He didn't need a bigger budget.
He needed to stop paying for people who were never going to buy.
Want to know where your Google Ads budget is leaking?
We'll record a free 10-minute Loom showing the biggest traffic and tracking leaks we can identify in your account, and where we'd start fixing them.
No call required. No pitch unless you ask.
→ Book your free Google Ads Leak Audit
After you book, we'll request read-only access to your Google Ads account and your last 90 days of spend data. The audit lands in your inbox within 5 business days. You keep it whether you hire us or not.
P&C is a marketing accountability firm. We make marketing measurable in revenue, not impressions. We work with owner-operated businesses across Canada, the UK, and the US.
Related reading:
- If You Can't Trace Your Last 10 Deals, You Don't Have Marketing. You Have a Marketing Expense. — The Stage 5 companion to this post
- Coming soon: The 5-Minute Pipeline Self-Audit (free PDF, no email gate)