Why am I getting junk leads from Google Ads?

Is it really junk, or just cheap?

Junk leads fall into three groups, and each one has a different fix.

• Fake. Bots, made-up contact details and form fills with no intent behind them. This is a traffic and form security problem.
• Real, but the wrong fit. People with no plan to buy: price shoppers, DIY researchers, people outside your service area. This is a targeting and qualification problem.
• Fine, but measured badly. Decent leads counted in the same bucket as newsletter sign-ups and accidental form touches, so the numbers look worse than the pipeline.

Pull the leads from the last 30 days and tag each one as junk, wrong fit or good. If most of the junk shares a pattern, such as the same source, the same time of day, the same landing page or obvious nonsense data, the cause is traffic or bots.

Why does Google Ads send junk leads by default?

Google optimises toward whatever you tell it counts as a conversion. Without tracking that separates good leads from bad ones, Smart Bidding treats every form submission and every call as equal. It then finds more of whatever is cheapest to produce, and the cheapest leads are almost always the worst.

Every junk lead you count as a conversion trains the algorithm to find more people like it.

Where do junk leads come from?

Five sources cause most of it.

• Search Partners. Ads shown on third-party sites outside Google Search, where traffic quality varies widely. It is one setting to switch off.
• Performance Max without structure. Early on it chases volume over quality, and it can serve on low-quality placements that attract bots.
• Broad match with weak negatives. Loose matching pulls in research, DIY and job queries unless someone mines the search terms report and adds negatives regularly.
• Unprotected forms. A form with no bot protection collects automated spam.
• Display Expansion and loose location settings. Ads leak onto Display or into the wrong areas and pick up cheap, low-intent clicks.

How do you stop junk leads from Google Ads?

Work through five steps in order. Each step removes a different kind of junk.

1. Clean the traffic sources

• Switch off Search Partners.
• Switch off Display Expansion on Search campaigns.
• Audit Performance Max: give it clear asset groups, clean conversion data and exclusions, or pause it as a test.
• Tighten location targeting and the ad schedule.

2. Mine the search terms report

Export the last 30 days of search terms. Mark the researchers, DIYers, job seekers and out-of-scope queries, and add them as negatives. Group the negatives into themed lists, such as informational, jobs, DIY and competitors, so every campaign can reuse them.

3. Harden the forms

Layer the protection so real people never notice it:

• A honeypot field, invisible to people and filled in by bots, with no friction at all.
• reCAPTCHA v3, which scores behaviour in the background.
• A timing check that rejects any form submitted in under two seconds.
• One qualifying question about budget, timeline or the service needed.

4. Validate leads before they reach the CRM

Check email addresses and phone numbers in real time, at the moment of capture. Fake details then never reach the CRM and never count as conversions.

5. Bid on validated leads only

Count a lead as a conversion only after it passes validation, and ideally only once it becomes a customer. Smart Bidding then looks for customers instead of form fills.

Field result. A client in crypto tax law had nearly all recent spend going to informational and DIY queries such as “do you pay tax on bitcoin”. Form fills kept arriving and the pipeline stayed empty. After cutting that traffic and counting qualified leads only, cost per real lead dropped from €430 to €133.

What is offline conversion tracking, and why does it fix lead quality?

Offline conversion tracking sends the outcome of each lead back to Google. When a lead becomes a customer in your CRM, Google learns which sources, audiences and search terms produce buyers, and moves budget toward them.

• Cost per lead can rise while cost per customer falls. For high-ticket lead generation, that is the trade you want.
• Google stops chasing cheap volume once it knows what a good lead looks like.

It needs your CRM connected to your tracking, and roughly 30 to 50 conversions a month per campaign before Smart Bidding can use the data well.

The most common mistake is counting every form interaction, chat window and phone-number click as a lead. That pollutes the signal before it ever reaches Google.

Which fix matches your symptom?

Symptom	Likely cause	First move
Gibberish names, fake numbers, forms filled in seconds	Bots on an unprotected form	Honeypot, reCAPTCHA v3 and a timing check
Junk spiked suddenly with no account changes	Search Partners or low-quality Performance Max placements	Switch off Search Partners, audit or pause Performance Max
Real people, wrong fit: DIY, price shoppers, out of area	Loose broad match, weak negatives, wide location settings	Search terms report, themed negatives, tighter locations
Plenty of leads, nothing in the pipeline	Every form fill counts, so bidding chases cheap conversions	Add a qualifying question and count validated leads only
Good leads got pricier while junk grew	Google optimising toward the cheapest conversion	Offline conversion tracking, bidding toward customers

What are the most common mistakes?

• Blaming Google before reading the search terms report. Traffic quality is the cheapest thing to check and the most common cause.
• Putting a CAPTCHA on every form. It blocks good leads too. Start with a honeypot, which adds no friction.
• Counting every form fill as a conversion. It trains the algorithm toward junk.
• Pausing Performance Max for good. Pause it as a test, then restructure it.

How we handle junk leads for lead-gen accounts

We clean the traffic first: Search Partners off, the search terms report mined, themed negative lists in place. Then we set an honest conversion definition that counts only real leads, and work toward offline conversion tracking so bidding targets customers.

When the CRM data is incomplete, we set up a simple source of truth, often a Google Sheet with UTMs, so the back-end numbers can be checked against what Google reports.