Three Checks to See Whether Google Ads Is Wasting Your Money

Your Google ads are running, the money is being spent, and the report shows leads. Yet the phone rings less often than the numbers promise. Usually the reason isn’t that Google “doesn’t work” — it’s one of three things: what the account counts as a lead, who the ads are shown to, and whether there is enough budget for the algorithm to learn.

You can check everything below yourself, without a marketer. Or at least know exactly what to ask your agency.

Check 1. Does your account count leads correctly?

Google Ads learns from whatever you’ve defined as a “conversion” — that is, from what the account treats as a lead. If the wrong thing is recorded there, the algorithm will diligently look for more of the wrong thing.

First — are conversions set up at all? Without them, Google doesn’t know which clicks brought customers and optimizes blindly, for clicks or views. This happens more often than you’d think: tracking was set up once, then the website changed, and it quietly broke.

Next — what exactly is being counted. The most dangerous conversions are the ones that are just a click: on a phone number, or on a Viber, Telegram or WhatsApp button. A click is intent, not a lead:

• a click on a phone number from a laptop calls no one, yet the “lead” has already been counted;
• one person may click several times — and if conversion counting is set to “Every”, that’s several leads instead of one, or none;
• a click on a messenger button isn’t a message: the person opened the chat and closed it;
• the algorithm sees these “leads”, decides the ads are performing well and looks for even more people who click instead of getting in touch.

In some countries there’s an extra catch. Where Google doesn’t offer call forwarding numbers — Ukraine is one example — it can’t count real calls from ads at all. In that case, “calls” in the report are almost always clicks on the phone number.

What to count instead of clicks:

• forms — the simplest and most reliable option: only a submitted form counts as a lead (a “Thank you” page or a successful-submission event), not a click on the button;
• call tracking — a separate phone number for ads that counts real calls and shows which ad they came from. Even there, it matters what you count: in practice, out of four calls an hour only one is a real prospect — the rest are missed, repeat or spam calls. So it makes sense to count an answered call lasting 30–60 seconds as a lead;
• offline conversions — sales from your CRM uploaded back into Google Ads. Then the ads learn not from leads, but from the people who actually bought.

How to check it yourself:

1. In Google Ads, open Goals → Conversions → Summary. If the list is empty, there are no conversions.
2. Look at the status of each action. “No recent conversions” or “Inactive” means tracking is broken.
3. See which actions are marked as primary — those are what the ads learn from. If the primary one is a phone or messenger click, the account is learning from intent.
4. Compare a week’s numbers with your own log of calls and enquiries. A gap of several times is a reason to dig in.

What to ask your agency: “Are these calls and enquiries, or clicks? How do you see real calls? Are we uploading sales from the CRM?”

Check 2. Is the account protected with negative keywords?

Ads are shown not for the words entered in the account, but for what people actually search. You can see the difference in the Search terms report.

Broad match isn’t dangerous in itself. It’s the setting where Google shows your ads for “similar” searches too — that’s how you find customers you hadn’t thought of. The danger is broad match without negative keywords: then “similar” searches bring people looking for something free, a job, instructions, another city or a different product altogether. These impressions eat 20–30% of the budget. A well-managed campaign keeps relevant searches at 97–98%.

Protection takes two tools:

• negative keywords — words your ads must not be shown for, in a specific campaign or ad group;
• negative keyword lists — shared sets applied to all campaigns at once: “free”, “jobs”, “DIY”, “used”, names of other cities. Without them, every new campaign has to be protected from scratch, and something always gets missed.

Why search terms don’t get cleaned up:

• there’s no routine. A new campaign should be reviewed daily for the first few days, then every 2 days for broad match and weekly for exact phrases;
• Google’s auto-applied recommendations can remove manual negative keywords on their own or switch keywords to broad match — and the junk traffic comes back;
• the whole search term gets excluded instead of the root word, so the next day the same search arrives in different words.

How to check it yourself:

1. Open the Search terms report for the last 30 days and sort by cost. The top 20 rows show where the money really went. Are they all about your product or service?
2. Look at Tools → Shared library → Negative keyword lists: do they exist, and are they applied to your campaigns?
3. Open Tools → Change history and filter by negative keywords to see when they were last added.
4. Go to Recommendations → Auto-apply and check what’s switched on. Anything that changes keywords or negatives is better turned off and decided manually.

What to ask your agency: “Which negative keyword lists are applied? When did you last clean up search terms?”

Check 3. Is there enough budget for the algorithm to learn?

Google’s Smart Bidding strategies learn from conversions. To find customers reliably at a reasonable cost, a strategy needs a critical number of leads — roughly 30 in 30 days. With fewer, the algorithm is guessing rather than learning.

That’s the main trap of a small budget. If the budget allows 5–10 leads a month, the strategy lacks data: it stays in “learning” for a long time, the cost per lead jumps around, and it looks as if the ads don’t work. It’s even worse when a small budget is split across several campaigns: leads are scattered, and no campaign reaches critical mass. A practical benchmark is around $15–20 a day per campaign.

What to do about it:

• don’t split the budget — one or two campaigns where leads accumulate beat ten campaigns with one lead each;
• don’t set a target cost per lead until you have ~30 leads a month — until then, use Maximize conversions without a target;
• don’t keep tweaking: every budget change of more than 20% or a change of strategy triggers a new 7–14-day learning period. Raise the budget in steps of 10–20% every 5–7 days.

Low bids are the second reason. Google runs an auction for every impression. If competitors pay more per click and have better ads, yours is shown rarely or at the bottom of the page.

How to check it yourself:

1. Count how many conversions each campaign got in the last 30 days. Fewer than 30 means the strategy lacks data.
2. Check the campaign status: “Limited by budget” is the first hint.
3. Add the Search lost IS (budget) and Search lost IS (rank) columns to the campaigns table. The first shows how many impressions you lost because of budget, the second — because of bids and ad quality. Tens of percent means you’re handing customers to competitors.

What to ask your agency: “How many conversions a month does each campaign get? Is that enough for the strategy? How many impressions do we lose to budget and to bids?”

Summary

Three checks, 15 minutes:

1. Conversions: they’re set up, and the ads learn from forms, calls and sales — not clicks.
2. Negative keywords: broad match is protected by negatives and shared lists, and search terms are cleaned up regularly.
3. Budget: each campaign gets enough leads to learn and doesn’t lose the auction.

If even one check fails, part of your budget is being wasted. See how we set up lead tracking on our web analytics page, and how we run search campaigns with negative keywords in search ads.

Don’t want to dig into it yourself? Request a free Google Ads audit — we’ll review your account and tell you what to fix first.