Stop Targeting. Start Disqualifying.

Your clicks are cheap. Your buyers aren't showing up. So you do the thing that feels responsible: you open the targeting settings and start narrowing. A tighter interest stack. A lookalike off your purchasers. Maybe an age band. It feels like control.

It's the wrong lever. Nine times out of ten, when the traffic is cheap but the sales aren't there, the targeting settings aren't the problem, and tightening them makes it worse. Here's what's actually happening, and where the real targeting lives.

Interest audiences are data you rent

When you pick "interested in skincare" or build a lookalike, you're not doing something clever the algorithm can't. You're paying a premium to rent a slice of data that Meta already has, and then handing the algorithm a smaller sandbox to work in.

Meta's machine is very good at one thing: finding the person who takes the action you optimized for. The more room you give it, the better it gets. Every interest you stack, every lookalike you layer, is a constraint you've put on the one system that's better at this than you are. Most of the time, the most effective audience is no audience at all. Broad, wide open, let it run. In our accounts, broad beats a hand-built audience something like three times out of four.

That sounds backwards until you see what "broad" is actually doing. It's not spraying your ad at everyone. It's reading who engages, who clicks, who buys, and steering toward more of them in real time. Your job isn't to describe your buyer to Meta in the targeting panel. Your job is to make sure the right signal is the one it learns from.

Which brings us to the actual problem.

An unqualified click is not free

Here's the part most founders miss. You look at a cheap CPC and think, great, cheap traffic, more is better. But a click from the wrong person isn't neutral. It's a lie you're telling the algorithm.

Every time someone who was never going to buy clicks your ad, Meta writes that down: this is what a good prospect looks like. Then it goes and finds more people like them. You've just taught the most powerful targeting system on earth to chase your worst-fit audience, and it will, efficiently, at scale, for as long as you let it. Cheap clicks that don't convert don't just waste today's budget. They pollute the data that decides tomorrow's delivery. You water down your own results over the long run.

So the goal isn't more clicks. It's fewer, better clicks. You want the wrong person to see your ad and keep scrolling. You want them disqualified before they ever cost you anything.

Move your targeting into the first line of copy

This is the shift. Stop targeting in the targeting panel. Start targeting in the ad itself.

The first line of your copy, the hook, is your real audience filter. Name your buyer out loud. Call out the exact person you want and, by implication, everyone you don't.

• Selling a high-end product? Lead with the price frame or the outcome that only your real buyer cares about. "If you've tried every drugstore serum and you're done wasting money…" quietly tells the bargain-hunter this isn't for them.
• Selling to a specific identity or situation? Say it in the first three words. Naming the person is the most reliable disqualifier there is.

We had to use this the hard way with an account in a niche Meta won't let you target anymore. The interest simply doesn't exist in the panel. So we put the identity as the first word of the ad. The people it was for leaned in. The people it wasn't scrolled past. The targeting moved out of the settings and into the sentence, and the account worked.

Read your current top ad. Does the first line let the wrong person off the hook, or does it pull everyone in with a vague "you deserve this"? If it's the latter, you're paying to teach the algorithm to find more of the wrong people.

Then add one filter in the form

Copy does most of the disqualifying. The lead form does the rest.

If you're running lead-gen, or any funnel with a form before the sale, add one conditional-logic question. Not five. One. Find the single question whose answer predicts whether someone actually buys, and ask it. Budget range. Timeline. "Do you currently have X?" The answer routes the good-fit person forward and quietly filters the tire-kicker.

The instinct is to keep forms short to lift opt-in rate. But opt-in rate is a vanity number if the opt-ins don't close. It is often better to have fewer opt-ins at a higher quality than a flood of leads your sales process has to wade through. One good filter question costs you a few opt-ins and saves you a pipeline full of people who were never going to buy.

What to actually do this week

1. Turn off your interest and lookalike stacks on one campaign. Go broad. Let Meta have the room. Give it a few days before you judge it, because you're changing what it optimizes toward.

2. Rewrite the first line of your best ad to name your buyer. Make the wrong person want to scroll past. That's a feature, not a bug.

3. Add one qualifying question to your form, the one that predicts a close, and stop optimizing for raw opt-in count.

The founders who win on Meta right now aren't the ones with the cleverest audience builds. They're the ones who stopped trying to out-target the algorithm and started feeding it clean signal. The right people in, the wrong people out, decided in the copy instead of the settings panel.

Your targeting was never in the targeting. It was always in the words.