Fake Leads from Google and Meta Ads: Where They Come From and What Helps
CTO, lead-metrics.com
The campaign is running, the cost per lead looks good, and then sales calls: half the numbers don't exist, the email addresses read asdf@asdf.com, and the people who do pick up don't remember any form. Fake leads cost you twice: once in ad spend and once in sales time.
There is a third, often overlooked cost: if every submitted form counts as a conversion, the algorithm learns from the spam and finds you more of it. This article covers where fake leads come from, why Google's and Meta's filters don't stop them and what actually helps. As of November 2026.
Key points
- Platform filters are not enough: Google itself says its invalid traffic filters don't prevent invalid leads on their own, because they only run after an ad is served.
- Fake leads have four sources: bots, people entering false details, accidental submissions in prefilled forms, and click farms.
- Protect the form: reCAPTCHA, server-side validation and double opt-in on your website; the "Higher intent" form type and SMS verification for Meta Instant Forms.
- The biggest lever: don't optimize for submitted forms but for qualified leads and deals from your CRM. According to Google, such lead goals trigger additional protections.
- Measure, don't guess: only when you see per ad how many leads turn into deals do you know where the fake leads come from.
Where fake leads come from
Not every bad lead is fraud. For choosing countermeasures it helps to separate four types:
| Type | Typical pattern | What helps |
|---|---|---|
| Bots and form spam | Many entries in a short time, made-up names, repeating patterns, often at night | reCAPTCHA, server-side validation |
| Real people, false details | Plausible name, but wrong number or disposable email | Double opt-in, SMS verification |
| Accidental taps and autofill | Person doesn't remember, profile data is outdated | Review step in the form, autofill off, qualifying questions |
| Click farms and incentivized traffic | Formally valid data, zero purchase intent | Optimizing for CRM stages, checking placements |
The first two types can be caught technically. The last two only show up once you see what happens to the leads in sales.
Why Google's and Meta's filters are not enough
Google filters invalid traffic such as bots, click farms and accidental clicks before billing, and credits what it detects afterwards. Since November 2026, Google has been gradually rolling out a new card on the Overview page of the Campaigns tab that shows both: the estimated savings from automated filtering and the credited cost (PPC Land, Nov 12, 2026). Alongside it, Google shared a number: in 2026 it reduced advertisers' exposure to invalid leads by an average of 30%. How Google defines an invalid lead and what the 30% refers to is not stated.
More important is what Google writes in its own help center: invalid traffic filters don't prevent invalid leads on their own. They run after an ad serves and therefore can't block traffic from your website. And real people sometimes enter false contact details too (Google Ads Help: Prevent invalid leads, About invalid traffic).
In practice: a bot that clicks your ad and fills in your form ends up in your CRM first. Whether Google credits the click later doesn't change that. Protecting the form is your job.
The real problem: spam becomes a training signal
Smart Bidding at Google and Meta's optimization learn from the conversions you report back. If the conversion is "form submitted", a bot entry counts just like a real inquiry. If the algorithm finds a placement where forms get filled out particularly cheaply, it moves more budget there. Cost per lead goes down, and so does lead quality.
That's why sorting out fake leads in your CRM alone doesn't help much. As long as the platform is told they are a success, it keeps optimizing for them. What this looks like with Performance Max and how to check where leads come from in the channel report is covered in the article on Performance Max for lead generation.
Protecting your website form
Google recommends three technical measures for website forms (Google Ads Help). Together with a hidden field, they make a solid baseline:
Add reCAPTCHA
reCAPTCHA checks in the background whether a human is filling in the form. The invisible version costs hardly any conversions. Make sure your consent banner covers the service.
Add a hidden field
A field that humans can't see (a honeypot) gets filled in by simple bots. If it contains a value, discard the entry. It costs nothing and catches part of the simple spam.
Validate on the server
Better bots bypass browser checks and post data straight to your form. Google therefore recommends verifying on the server that the email domain is valid and the phone number matches a legitimate regional format. You can filter out disposable addresses in the same step.
Use double opt-in
With double opt-in, a lead only counts as valid once the person clicks the link in a confirmation email. It costs volume but confirms the address is real and read. For leads you want to call right away, checking the phone number is often enough.
If your forms run in funnel tools like Heyflow or Typeform, check which of these features the tool provides itself.
Protecting Meta Instant Forms
Instant Forms are especially prone to accidental submissions because name, email and phone number are prefilled from the profile. Meta offers several settings:
- "Higher intent" form type: before submitting, people review and confirm their details (Meta Business Help).
- SMS verification: people confirm their phone number with a code, otherwise the form isn't submitted (Meta Business Help). The phone number must be a required field for this.
- Turn off autofill: autofill can be disabled for email and phone so outdated profile data doesn't slip through (Jon Loomer).
- Qualifying questions: one or two multiple-choice questions, for example on timing or budget, filter out the merely curious.
Each of these measures lowers the number of leads. Whether that pays off is decided by cost per closed deal, not CPL. How to compare both routes fairly is covered in the article instant forms or landing page.
Optimize for real leads, not forms
The most effective protection is to stop reporting every form as a success to the platform and report only leads that move forward in sales.
Google Ads: Google recommends lead goals such as "Qualified lead", "Converted lead", "Book appointment" or "Request quote". According to Google, these goals automatically trigger invalid traffic protections built specifically for lead generation campaigns. As guidance, Google names at least 15 conversions in the last 30 days, uploads ideally daily, and a conversion that happens within 7 days of the ad interaction (Google Ads Help: Best practices for generating high-quality leads). The CRM data reaches your account through offline conversions with Enhanced Conversions for Leads.
Meta: for Instant Forms there is the Conversion Leads performance goal, which optimizes for CRM stages instead of the submitted form. Requirements and setup are covered in the article on bad leads from Meta Ads. For leads from your website, send later stages as events through the Conversion API.
Which stage you pick depends on volume: the closer to revenue, the less spam, but also the fewer conversions for the learning phase. The qualified lead is often the best compromise. How to choose stages, assign values and meet deadlines is explained in the guide sending pipeline stages as conversions to Meta and Google.
Measuring fake leads
Without numbers, all you have is a feeling that "lead quality got worse". Three things make the problem visible:
- Label spam in your CRM: create a dedicated lost reason for invalid leads, such as "spam / invalid data". That way the share can be analyzed instead of the leads simply disappearing.
- Compare leads and deals per ad: an ad with many leads and not a single deal is suspicious, especially if other ads in the same campaign close normally.
- Look for patterns: clusters at unusual hours, many leads from one placement, identical answers to qualifying questions, or leads without a prior website visit.
Common mistakes
- Looking only at CPL: a falling CPL with flat deal numbers is often the first sign of more spam.
- Silently deleting spam: removing invalid leads from the CRM instead of labeling them destroys the data you need for analysis.
- Reporting every lead as a conversion: then every bot entry teaches the algorithm to find more of them.
- Tightening everything at once: if you add reCAPTCHA, double opt-in, SMS verification and new questions at the same time, you won't know what worked and what only cost volume.
- Counting on credits: Google credits invalid clicks, not the time your sales team spent on made-up numbers.
How LeadMetrics helps
LeadMetrics doesn't replace spam protection in your form. But it shows you where bad leads come from and makes sure the platforms optimize for the good ones:
- Leads and deals per ad: lead tracking attributes every lead to its campaign, ad set and ad. With the CRM integration for HubSpot and Close you see deals, cost per closed deal and revenue next to it. Ads with many leads and no deals stand out immediately.
- Journey per lead: in the lead list you see the journey timeline and click history for every lead, so whether and how often the person visited your website before the form.
- Only good signals back: in the Conversion API you decide per project which LeadMetrics event goes to which platform event, for example won deals instead of every lead. Custom metrics such as a qualified lead, set for example by a HubSpot workflow, can also be sent as events to Meta, Google Ads, TikTok and LinkedIn.
That way you only report leads to the platforms that are worth something in sales, and you see per ad whether your measures against fake leads have worked.
Ready for better tracking?
Discover how LeadMetrics helps you understand the true ROI of your marketing channels.
Try free for 10 daysTerms in this article
Related guides
How to Send Sales Pipeline Stages as Conversions to Meta and Google
Which pipeline stages to send to Meta and Google Ads as conversions, how to map them to events, assign values, meet upload deadlines and avoid duplicates.
Read guideMeasure Real ROAS in Lead Generation: From Click to Closed Deal
How to measure real ROAS in lead generation: attribute CRM revenue to campaigns, read CPL, CPA and ROAS correctly and send deal values back to Meta and others.
Read guide