Logo featuring a stylized bee, circuit lines, and the text BLUE AVISPA in bold turquoise font.
Google Reviews logo with five gold stars underneath, indicating a five-star rating.

Why you should NOT use Google Performance Max

Quick Note: WOW, I asked ChatGPT on Extended Thinking to make and image and the details just blew me away!!! It got video titles, click rates, spam etc in there! I’m going to put the image at the bottom. it NAILED the details. Ok, Back to the topic

Why I Stopped Using Google Performance Max for Lead Generation

Note: Google Ads are ok, and Google Local Service Ads are great! Google Performance Max in my opinion… is garbage.

There is a big difference between a bad lead and a fake lead. A bad lead is a real person who is not a fit. A fake lead is a bounced email, a fake number, or a form submission that was never a real sales opportunity at all.

That is what made this experience with Google Performance Max so bad. This was not just disappointing lead quality. It was a campaign that appeared to run across broad, low-relevance inventory while producing submissions that often looked invalid in the real world.

Illustration of a person with a magnifying glass examining a website search page on a computer screen.

The Core Problem

  • When we reviewed the submissions, many of the email addresses were invalid.
  • We called and emailed the leads, and many of the phone numbers were fake while many emails bounced.
  • When I checked the fake submissions, they were coming from this campaign.

So this was not just a case of weak leads. It was a case where a lead-generation

campaign was producing a meaningful amount of junk.

Illustration of a person fitting a puzzle piece onto a presentation board with images and a gear icon.

The Placement Problem

The placement export I reviewed was a one-day snapshot for August 28, 2025,
and even that single day was enough to show the issue.

  • 2,619 placements in total
  • 2,079 placements on YouTube about 79.4%
  • 335 placements on apps/app stores about 12.8%
  • 205 placements on websites/other inventory about 7.8%

That alone is a red flag for a local home-service lead campaign.

Using a conservative title-and-URL review of the placements, at least:

  • 326 placements about 12.4% showed obvious children’s-content signals
  • 532 placements about 20.3% showed obvious foreign-language or non-English signals
  • Only about 69 placements roughly 2.6%  showed obvious home-remodeling / cabinet / kitchen / bathroom relevance
Those numbers are not a perfect scientific classification, because they are based on visible placement names and URLs, but they are conservative enough to make the point:
this campaign was not staying tightly on relevant inventory.

The ads were showing on things like:

  • children’s videos
  • cartoons
  • foreign-language videos
  • random entertainment content
  • unrelated apps

Examples from the report included:

  • “¡A Cantar con los Animales! … Canciones Infantiles”
  • “¡Ambulancia, Ayúdame! … BabyBus en Español”
  • Mickey and friends content
  • Pocoyó and other children’s cartoons
  • Pinkfong / Baby Shark style content
  • Apps like ABC Games: Tracing & phonics, Azar, and Adam4Adam
For a cabinet refinishing campaign, that is not a small miss. That is a major relevance problem.

Why This Matters

Our campaign was location-restricted, so the issue was not simply that the ads went outside the target area. The issue was that, within the allowed area, Google was still distributing the campaign across children’s content, foreign-language content, apps, and other low-intent placements instead of keeping it tightly aligned to high-relevance home-service intent.

If the ads are being shown in weak environments, it becomes much easier for junk traffic and junk submissions to enter the funnel.

We Did Not Run This Carelessly

We did not just switch on Performance Max and hope for the best. We tried to prevent this.

  • We filled in the available branding and business information.
  • We chose the relevant category.
  • We applied every restriction we reasonably could.
  • We regularly blocked YouTube videos and placements.
  • We worked with Google Ads experts for 2 months or so.
  • We followed guidance from Google representatives.

So the problem was not that we ignored the tools or failed to tighten the campaign. The problem was that even after doing that, the campaign still showed in low-relevance places
and still produced fake leads.

Outline of a gear with a thumbs up and a check mark, symbolizing approval, quality, or successful settings.

Google’s Response

Google separated the issue into two categories:
  • invalid traffic / invalid clicks
  • invalid leads

In one message, Google said there were 0 invalid clicks recorded on the Performance Max campaign for part of the date range reviewed. Later, they said they did find suspicious traffic, but that it had already been filtered and credited so we were not charged for it.

But when it came to the fake form submissions, Google’s answer changed.
Instead of treating that as a placement-quality or supply-quality issue, they framed lead quality as a shared responsibility and pushed the solution back onto us.

Their recommendations were basically to:

  • use reCAPTCHA
  • validate emails and phone numbers
  • tighten form protections on our end

That is the part I cannot ignore. They did not meaningfully acknowledge the supply problem. They did not say they would stop showing ads on obviously poor-fit placements. They did not fix the spam-lead issue at the source.

Instead, the answer was: filter it harder on your end.

This experience changed how I look at Performance Max for lead generation.

I think this is just lazy advertising and I see it a lot. People set up perfomance max and let it go. There really is no good way to run that, it’s a a fundamental issue WITH Perfomance Max from what I saw.

  • a 2,619-placement one-day report
  • nearly 4 out of 5 placements on YouTube
  • 326 placements with obvious children’s-content signals
  • 532 placements with obvious foreign-language or non-English signals
  • only about 69 placements with obvious remodeling / cabinet / kitchen / bathroom relevance
  • fake phone numbers and invalid email addresses tied to the campaign
  • Google responding by telling us to improve filtering rather than fixing the underlying placement problem

That is why I stopped trusting Performance Max for this kind of lead generation.

If a campaign can show on low-relevance content, generate junk submissions, and then leave the advertiser to put up with fake leads to filter through because they cannot or will not.

Below is sample from a 2622 line CSV Report I exported directly from Google Ads Using their own Report tool on Aug 28, 2025.

This was an ongoing issue that repeated day after day after day for months. This is just the first few pages of it. I assure you, there was a lot more where this came from! I’m excluding the whole file in case there was any personal info or private campaign info in the report.

A website-style list of YouTube video titles, emojis, and URLs organized in a handy spreadsheet format.
A screenshot of a website's YouTube video list, displaying titles and their links in a neat table format.
A website-style list of YouTube video titles, emojis, and URLs organized in a handy spreadsheet format.
A stressed man looks at a laptop showing fake leads on his website and wasted ad budget, with warning signs and analytics.

Recent Posts

Share this: