Most “fake” Facebook leads are not fraud. They are real people who tapped a form by accident, or who filled it in without meaning to speak to anyone — and the reason so many of them reach you is that Meta's Instant Form is designed to be as easy as possible to submit.
That matters, because the fix for accidental leads is completely different from the fix for fraud.
Why Instant Forms produce so many low-quality leads
An Instant Form opens inside Facebook or Instagram and pre-fills the person's name, email and phone from their profile. Submitting it can take two taps. There is no website visit, no page to read, and often no moment where the person consciously decides “yes, I want this company to call me.”
That is not a flaw — it is why Instant Forms produce a low cost per lead. But it means the intent bar is on the floor. Compare that with a landing page, where someone has to click through, wait for a page to load, read something, and type their details. Every one of those steps removes people who were never serious.
So a high volume of weak leads from Instant Forms is the expected outcome, not a malfunction.
Why the phone numbers are wrong
Meta pre-fills the phone number from the person's profile. That number was often entered years ago, for two-factor authentication, on a phone they no longer use. Nobody is lying — the data is simply stale, and the form did not ask them to confirm it.
This is the single easiest thing to fix, and most advertisers never do it.
Why people deny filling in the form
Two common causes, and neither is dishonesty.
Accidental taps. A large share of ad impressions are served in placements where mis-taps are common — scrolling through Reels, Audience Network, in-app placements. The form opens, a thumb lands on the pre-filled submit button, and the person scrolls on without registering what happened.
Time. If you call four days after the form was submitted, a person who filled in five forms that evening genuinely will not remember yours.
What actually fixes it
1. Switch the form to higher intent
In Meta's form builder, change More volume to Higher intent. This adds a review step before submission, so the person has to confirm. Your lead count will drop and your cost per lead will rise. Both of those are the point.
2. Add a qualifying question
One question that a casual tapper will not bother answering — budget, timeline, city, or what they actually need. A free-text or multiple-choice question is friction, and friction is what you are buying here.
3. Ask them to confirm the phone number
Add a phone field that is not pre-filled, or mark the pre-filled one as requiring confirmation. This alone removes most of the dead numbers.
4. Call within minutes, not days
Response time does more for Instant Form lead quality than any targeting change. The same lead that feels “fake” after three days is often perfectly real at three minutes — they were still holding the phone, still thinking about it.
This is where most businesses lose the leads they have already paid for, and it costs nothing to fix.
The measurement problem underneath all of this
Here is the part that quietly makes everything worse. Meta optimises for what you report back to it. If you never tell Meta which leads became customers, the only signal it has is “this person submitted a form.”
So it goes looking for more people who submit forms easily — which is precisely the audience producing your weakest leads. Month three is worse than month one not because Meta got worse, but because it got better at finding what you taught it to find.
Breaking that cycle means sending the outcome back: this lead became a customer, this one did not. Once Meta has that, it starts optimising toward people who buy rather than people who tap.
How Claudphic Ads helps
Claudphic Ads captures every Instant Form lead the moment it is submitted and reports the outcome back to Meta automatically when you mark it Won — so your campaigns learn from real customers instead of form fills. It also scores incoming leads for the signals that usually indicate a weak or automated submission, and shows response time per lead so you can see where the delay actually is.
None of that replaces fixing the form itself. Do the four steps above first; they are free, and they work whatever you use to manage the leads afterwards.