The screenshot every local-service owner has sent me at some point: 50 leads this month, cost per lead down, the campaign "working." Then the follow-up a week later. Half the numbers go straight to voicemail. A third are two states away asking about a service you do not offer there. A couple are clearly bots filling the form with garbage. The dashboard says you won. The phone says you did not.

Short answer: Google Ads sends bad leads because your settings and your conversion goal tell it to. Loose match types and the presence-or-interest location option invite off-intent, out-of-area searches, and optimizing to raw form fills teaches Google to fetch the cheapest fills. Tighten targeting and optimize to a qualified event.

The takeaways

Your match types and location setting manufacture the junk upstream, before your offer or landing page ever get a say.
Google's "Presence or interest" location option (its recommended default) shows your ad to people who merely searched for your area, not people in it.
Optimizing to raw form fills trains the algorithm to find you more cheap fills; only the cost per qualified lead tells you what a customer actually costs.

Why does Google send bad leads when the lead volume looks fine?

Bad leads at healthy volume are almost always a targeting-and-optimization problem, not an offer problem. Google's bidding does exactly what you reward it for. If every form fill counts as a conversion of equal value, the algorithm learns that cheap, easy fills are the goal, so it goes and finds more of them. Volume looks great precisely because the leads are low-intent. A tighter, more expensive qualified buyer is harder to find, so the system quietly drifts toward the opposite. The lead count is not lying to you; it is answering the wrong question.

Is the "Presence or interest" location setting sending you out-of-area leads?

Often, yes, and this is the single setting most local advertisers never check. Under a Search campaign's location options, Google defaults to targeting people by "Presence or interest": people in, regularly in, or who have shown interest in your targeted locations (Google Ads Help). "Shown interest" means someone in another state who searched "plumber in Tampa" can see your Tampa ad and fill your form, and there is no way you can book them. Switch the setting to "Presence" (people actually in your area) and a whole class of impossible-to-serve leads disappears. For most businesses with a real service boundary, presence is the correct choice, and it is one radio button.

How do loose match types manufacture junk searches?

Broad match hands Google permission to show your ad on searches that only loosely relate to your keyword, and that is where a lot of off-intent leads are born. Someone searching for "free," "DIY," "jobs," or "how to do it myself" is not your buyer, but a broad keyword can still trigger for them. The defence is boring and it works: open the Search Terms report every week, read the actual queries that spent money, and add the irrelevant ones as negative keywords at the ad group and campaign level. Pair that with exact or phrase match on the terms that convert. The point is to stop paying for searches that could never become a customer, while keeping the ones that do.

Why does optimizing to form fills get you more junk?

Because a form fill is the easiest thing on the page to produce, and Smart Bidding optimizes for whatever you told it to count. When the account only sees the first form submission, a bot, a price shopper, a wrong-fit inquiry, and a qualified buyer all look identical: one conversion each. The algorithm has no way to tell them apart, so it chases the cheapest of them. This is why accounts that add form validation and lead scoring often watch their raw lead count fall while the phone starts ringing with real prospects. You did not lose leads. You stopped counting the ones that were never real.

How do you judge Google Ads lead quality honestly?

By cost per qualified lead, not cost per lead, which means you have to define "qualified" and feed it back to Google. Tag each lead in your CRM or call tracker (booked, out of area, spam, wrong service), then send that judgment back as an offline conversion so bidding optimizes toward the leads that actually progress. Now the honest number appears: what it costs to win someone who can and will buy. A $45 cost per lead can hide a $1,500 cost per real customer once you strip out the fills that never answered. The raw count was cheerful; the qualified count is true.

Read the number that survives contact with reality

Every one of these fixes points the same way: stop trusting a self-reported count and start judging on the outcome you can verify downstream. That is the whole discipline of reading ad performance honestly. It is also the principle behind how Adscalr scores creatives, where the composite weights real ROAS and revenue per install rather than a raw conversion tally that any lucky form fill can inflate. If you want the longer version of that argument, our ad intelligence work is built around not fooling yourself with the flattering number.

The lead volume was never the problem. What you asked Google to optimize for was.

Why Google Ads Sends You Bad Quality Leads