Nearly every business running lead forms on Facebook eventually opens the CRM and finds names that go nowhere. Numbers that ring out, addresses with a character missing, people with no memory of filling anything in. At Growth Bully, a Malta performance marketing agency, we usually get asked to switch the fraud filters on. There are rarely any fraud filters to switch on, because what gets called a fake lead is three separate problems wearing the same coat.
Separating them matters, because the fixes point in opposite directions. Tightening a form does nothing to a lead who is simply too early. Adding qualifying questions does nothing to an automated submission. The first job is to read what you already have before you change anything in the account.
What counts as a fake lead from Facebook ads?
A genuinely fake lead is one no human submitted, or one a human submitted with deliberately false details. It is rarer than most advertisers assume. The large majority of what gets labelled fake is real: a real person who tapped a real button, and who either mistyped a digit or has no current intention of buying.
That distinction decides where you spend your time. Fraud tooling is sold against the rare case, which is why so many businesses buy a filter and see no change in the pipeline. The common cases are not a security problem at all. They are an instruction problem: the campaign was told to collect form fills and it collected form fills.
Why am I getting low-quality leads from Facebook ads?
Because submitting was effortless and the advert asked nothing of the reader. The platform reliably finds the cheapest available action, and a pre-filled form behind a broad offer is close to the cheapest action available. What arrives is volume from the top of the decision, not people ready to speak to you this week.
Cheap is the symptom people notice last. A falling cost per lead alongside a flat number of conversations is the clearest signal that quality has moved, and it reads as success on every dashboard until someone counts the calls that actually happened.
How do you tell the three problems apart?
Pull the last fifty leads and sort them into three piles. The pile they land in tells you which fix to apply, and it takes about twenty minutes. Most accounts find that the pile they assumed was largest is actually the smallest.
- Not a person, or deliberately wrong. Invalid number formats, mismatched country codes, nonsense names, several submissions from the same device inside a few minutes. If this pile is more than a handful, treat it as a delivery and placement problem.
- A real person, wrong details. One digit short, an old number, a typo in the address. These are recoverable and they are mostly an interface problem rather than an intent problem.
- A real person, genuinely interested, wrong fit. They answer, they are polite, they cannot afford it or they wanted something you do not sell. This is the pile almost everybody actually has, and no form setting touches it.
Record the split rather than remembering it, so you can tell whether a change worked. Our note on lead scoring for small teams covers how to keep that record without building anything elaborate.
How do you stop junk and automated submissions?
Add a small amount of friction and verify the contact detail you intend to use. A pre-filled instant form with no confirmation step is the easiest thing in the account to submit by accident. Asking the reader to retype the number they want you to call is enough to clear most of it.
Three changes do the bulk of the work. Switch the form from the higher-volume setting to the higher-intent one, so the reader has to confirm the submission on a review screen. Require the phone field to be typed rather than auto-filled. Then check your placements: a disproportionate share of junk tends to arrive from the broadest automatic placements rather than from the feed.
Do qualifying questions improve Facebook lead quality?
Yes, and more than any filter will. Two or three questions that a buyer can answer instantly, and a browser cannot answer comfortably, change who finishes the form. Ask about timing, the scale of the job, and the thing that would disqualify them.
The reason this works is self-selection rather than screening. You are not using the answers to reject people, you are using the questions to make the form feel like the start of a conversation. Expect the lead count to fall and the conversation count to hold. If conversations fall too, the questions are too long or too personal for a first touch. What a usable answer looks like is set out in what a qualified lead is.
Can the offer be the reason the leads are bad?
Often it is the whole reason. An offer with no cost of entry attracts people with no stake in the outcome, and a discount or giveaway attracts people shopping for the discount. Lead quality is set by what you asked somebody to raise their hand for, long before the form renders.
The practical test is to read the advert as a stranger and ask what the next five minutes look like. If the answer is nothing in particular, you have invited a browse. Moving the ask from a vague enquiry to something specific and dated, a quote on a defined job or a short call about a named problem, changes the pile the leads land in. There is more on building that ask in offer design for lead generation.
What should the campaign optimise for?
Whatever you actually want, which is almost never a form fill. If the only event you send back to the platform is a submission, that is the behaviour it will keep buying. Optimising for a qualified lead or a booked call, even on lower volume, changes who the delivery system goes looking for.
We have watched this in our own account. A small test budget on a lead advert produced sixteen leads at around ten euro each and not one of them turned into a conversation, which looked like a cheap success and was nothing of the kind. The full read of that account, including which signals did predict closed clients, is in the Facebook ads metrics that predict clients. One account is directional rather than conclusive, but the direction has held every time we have retested it.
Sending that event back means wiring the CRM to the platform so a screened lead reports as a conversion. It is a build rather than a setting, and the single change that most reliably moves quality on a Facebook ads account.
How long before lead quality changes?
Allow two to three weeks after a change before you judge it, and compare whole cohorts rather than days. Delivery takes time to settle after a form or event change, and lead quality can only be measured once the leads have had a chance to answer the phone. Reading it inside a week will show you noise.
Judge it on one number: the share of leads that reach a real conversation. Cost per lead will move in the wrong direction when quality improves, and anyone reporting only cost per lead will tell you the change failed. If the leads still go nowhere after two cohorts, the problem is upstream of the form and sits in the offer or the audience.
If you would rather have someone read the last fifty leads with you and say which of the three piles you actually have, tell us what the leads are doing and we will give you the answer and the order to fix it in.