Should You Optimize for Form Fills or Real Conversions?

Real conversions — once your account can reliably send Meta that signal. Optimizing for form fills tells Meta to maximize submissions, so it finds the people most likely to tap "Submit," quality be damned. Optimizing for real conversions (the deeper event: a qualified lead, a booking, a sale) tells Meta to find people most likely to become customers. Same budget, same ads, completely different people. The one catch: Meta can only optimize for an event it can actually see — so the real answer is "optimize for conversions, and make sure Meta can measure them." This is the exact lever behind the cheap-leads-that-never-book problem (#).

What's the actual difference between a "form fill" and a "real conversion"?

They sit at different depths of your funnel:

• A form fill (Meta calls this a Lead) fires the instant someone submits your instant form. It's shallow, cheap to trigger, and easy to fake — a bored thumb counts.
• A real conversion is an event further down: a qualified lead, a booked appointment, a purchase. It fires only when something of value actually happened.

Meta will happily optimize for either. Whichever event you point it at, it goes and finds more of the people who trigger that specific event. Point it at form fills and it becomes an expert at manufacturing form fills. That's not a bug — it's the machine doing precisely what you asked.

Why does optimizing for form fills backfire?

Because "most likely to submit a form" and "most likely to buy" are different populations, and Meta's targeting is good enough to tell them apart. When your optimization event is the form fill, the algorithm learns the fastest, cheapest path to more submissions — which usually means lower-intent people. Volume goes up, cost-per-lead goes down, and your close rate quietly collapses.

You end up with 100 leads and 0 sales — a great dashboard and an empty calendar. The why behind that, and the money math, is the subject of the companion piece: Why are my Facebook leads cheap but nobody books? (#)

What are "Conversion Leads," and when should I use them?

Conversion Leads is a Meta optimization option for lead campaigns that shifts the goal from number of form fills to leads most likely to convert down the funnel. Instead of chasing raw submissions, Meta uses the outcome data you send back — which leads your CRM marked as qualified or won — to find more people like your actual customers.

It's a strong fit when:

• You run lead forms but only a fraction of leads are any good.
• You can send lead status back to Meta (via CRM integration or the Conversions API).
• You have enough conversion volume for Meta to learn from.

It's a poor fit if you can't feed outcomes back yet — because then there's nothing for Meta to optimize toward. Setup steps and eligibility for Conversion Leads change over time; reconfirm the current flow against Meta's own docs before following any specific instructions.

What's the exact setting to change?

Two layers decide what Meta chases, and both have to line up:

1. The campaign objective — Leads vs Sales vs Traffic vs Engagement. This is the coarse dial. (Which one fits your business is its own question: Leads vs Sales vs Traffic — which should I pick? (#).)
2. The optimization / conversion event — inside the objective, the specific event you tell Meta to maximize. This is where "form fill" vs "a deeper conversion" actually lives.

The move, in plain terms:

• If you're on a lead objective optimizing for Leads (form fills) and getting junk, switch toward Conversion Leads or a deeper conversion event, and connect your CRM/offline outcomes so Meta can see which leads mattered.
• If you sell online, optimize for the Purchase conversion, not add-to-cart or landing-page views.
• Either way, the principle is the same: point Meta at the event closest to money that you can reliably measure.

Won't optimizing for real conversions make my leads more expensive?

Usually yes — and that's the point. Fewer, better leads cost more each but close more often, so your cost-per-booked-job drops even as your cost-per-lead rises. Judging the switch by CPL will make a good decision look like a bad one; judge it by real customers in your CRM instead. (The full CPL-vs-cost-per-booked-job argument is in the companion piece (#).)

There's also a learning cost: deeper events happen less often, so Meta needs more time and volume to optimize well. If your conversion volume is very low, you may need to keep a shallower event temporarily while you build up signal — a real tradeoff, not a rule.

What has to be true before this works?

One thing above all: Meta has to be able to see the conversion. If your real outcomes never make it back to the platform, "optimize for conversions" is an empty setting — Meta has no data to chase. That's why this decision is inseparable from your tracking:

• Your Conversions API (or CRM/offline event upload) has to be live and sending the right events — see Do I actually need the Conversions API? (#).
• Meta's reported conversions should roughly line up with your pipeline — a big mismatch means the signal is broken (see Meta says 40 leads but my CRM shows 22 (#)).
• A human still owns the call on which event to chase and when. This isn't a set-and-forget switch, and no setting guarantees results — it aligns the algorithm with your business instead of against it.

Get the signal flowing and the event set correctly, and Meta stops optimizing for the wrong thing. That single alignment fixes more "my leads are terrible" problems than any creative or bidding tweak.