Fake Facebook Leads? Where They Come From — and the Pipeline That Filters Them Out

Every contractor, realtor, and med spa that runs Facebook Lead Ads eventually says a version of the same thing: “half these leads are fake — it’s like Facebook pays people to fill out the form.” The suspicion is understandable, the truth is more useful: junk leads come from four specific, fixable mechanics — and a lead pipeline built with those mechanics in mind produces a different quality of lead entirely. This guide covers both: where the junk actually comes from, then the setup, step by step.

Where Junk Leads Actually Come From

1. The two-tap submission. Meta’s Instant Forms pre-fill name, email, and phone from the user’s profile — that’s what makes them convert so well, and it’s also what lets a mildly curious scroller (or a misplaced thumb) submit a complete-looking lead in two taps without ever typing anything. These are real people with real contact info and near-zero intent. They are not fraud; they are friction working against you.

2. The cheap-inventory problem. By default, Meta can show your ads across its Audience Network — third-party apps and sites, including rewarded placements where users are compensated for engaging with ads, and where spam sites with documented bot traffic live. This is the grain of truth inside “Facebook pays people to fill out forms”: Meta doesn’t, but some of its cheapest inventory effectively does. Lead campaigns rarely belong there.

3. Optimization finds whatever you ask for. Meta’s delivery system optimizes for the outcome you chose — and if that outcome is “form submissions,” it will hunt down the people most likely to submit forms, who overlap heavily with sweepstakes-fillers and chronic clickers. Ask for cheap leads and the machine delivers exactly that, literally.

4. The lead was real when it was submitted. The response-time research is brutal on this: contact odds collapse within minutes. A homeowner who tapped your roofing ad at 12:40 and got a call Thursday reads as “fake number, never answered” in your notes — but the lead wasn’t fake; your pipeline was slow. In our experience this explains more “junk” than everything above combined, and it’s the one cause that’s entirely on your side of the fence.

The Pipeline That Filters Junk Out

Build the campaign in this order and each mechanism above gets a countermeasure.

Step 1 — Make the offer ask for a decision, not a prize. “Free Roof Inspection — takes 30 seconds” attracts homeowners with roof problems. “WIN a free roof upgrade!” attracts people who enter things. Your creative is the first qualifying question: name the service, name the area, imply a next step that involves talking to a human. Boring converts worse and closes better.

Step 2 — Use the Higher Intent form type. Meta offers two Instant Form types : More Volume (submits instantly, no confirmation) and Higher Intent (adds a review screen the person must confirm before submitting). That one screen kills most two-tap accidents. You’ll see fewer leads — commonly reported drops run 10–40% — and the ones that remain meant it. For any business that calls its leads, this is the default, and it’s a radio button, not a project.

Step 3 — Add one question they have to type. Pre-filled fields cost nothing to submit; a short-answer question (“What needs repairing, roughly?” / “What’s your timeline?”) costs three seconds of typing — an eternity to a bot or an accidental submitter, nothing to a homeowner with a leak. One or two qualifying questions is the sweet spot: multiple-choice for data you’ll filter on, one typed answer as the intent gate. Every answer also arrives with the lead, so whoever calls opens with context , not a cold “so, uh, you filled out a form?”

Step 4 — Take your ads out of the junk inventory. In Ads Manager, switch placements from Advantage+ (automatic) to manual placements and uncheck Audience Network — at minimum, rewarded video. Keep Facebook and Instagram feeds, Stories, and Reels: places where a real person is scrolling a real feed. Expect the raw cost-per-lead to tick up — you removed the cheapest, worst inventory — and cost-per-lead-that-answers to come down. Judge the campaign on the second number, never the first.

Step 5 — Point targeting at your actual market. A tight service-area radius, sensible age bands, and no interest-targeting cleverness beats a broad net for local services. Every mile outside your real service area is a mile where every lead is junk by definition — the FTC-documented misery of bought marketplace leads frequently starts with exactly this mismatch, and there’s no reason to recreate it inside your own campaigns.

Step 6 — Respond like the research says. The pipeline’s last mile is the one most builds skip: the lead must reach a human within minutes, or Steps 1–5 were for nothing. That means delivery to the on-call phone by SMS , the inbox, and the tracking sheet the moment it’s submitted — not whenever someone remembers Meta’s Leads Center exists . Meta notifies no one by default, holds each lead for 90 days, then deletes it — the storage is quiet, the decay is not.

Step 7 — Close the loop with outcomes. Add a Status column to your lead sheet (answered / booked / junk) and review it weekly per campaign. Two payoffs: you learn which ad and which question produce leads that book (not leads that are cheap), and if you graduate to Meta’s conversion-leads optimization later, that outcome data is exactly what teaches the delivery system to find bookers instead of form-fillers. The measurement habit is the difference between “Facebook doesn’t work for us” and knowing your cost per booked job to the dollar.

The One-Week Diagnostic

Before concluding your leads are fake, run this: for one week, respond to every lead within five minutes (SMS delivery to whoever’s on call makes this workable even mid-job). Track contact rate against the previous month. If it jumps — and for most businesses it jumps dramatically — the leads were never fake; they were perishable. If it doesn’t jump and the numbers ring dead, work Steps 2–4 above: you have a form-and-placement problem, and it’s fixable in an afternoon.

Either way, the answer was in the pipeline — not in the conspiracy.