I built big audience lists for years. Now I delete them.

01What picking the audience by hand was for

In the old days the platform knew very little about who would buy from you. So you told it. You said, I think my buyer is 35 to 45, follows these 3 pages, and does these 2 things online. The system went and found those people. Then you learned if your guess was right.

That was real work and it was worth money, because your guess was usually better than the machine's guess. That has flipped. Once the machine has response data, who clicked, who stopped, who filled the form, its guess is much better than yours. Your guess has turned into a fence around a better process.

Your idea of your buyer used to be the best information in the account. Now it is the limit on it.

02What narrowing actually costs you

Every filter you add does 2 things at once. It removes people who would have bought but did not match your idea of a buyer. And it shrinks the pool, so you burn through it faster and the same faces see your ad again and again.

On a very big budget you can take that hit. On the budgets most service businesses actually run, narrowing is how you get to week 3 with an audience that has seen your ad 11 times, and you sit there wondering why the results fell apart. They did not fall apart. You ran out of people. And you did it on purpose, in the settings, on day 1.

I want to be blunt about this one, because it is the piece most owners get wrong. Small budgets need broad audiences more than big budgets do, not less.

03The 3 settings I keep

Only 3, and only where they are genuinely true for your business.

Now look at what is not on that list. No interests. No behaviours. No stacked lookalikes. No exclusions, except the ones you really need, like keeping your current clients out of a campaign meant to find new ones.

The age band is the one people fight me on. Fair enough. Ask yourself 1 question about it. Have you ever actually tested it, or did you type it in once in 2019 and never look again? Most age bands I see are a guess that got promoted to a rule.

04So where does the targeting go? Into the ad

This is the part that makes broad work, instead of making it lazy and cheap. The targeting does not disappear. It moves out of the settings panel and into the first line of your ad.

When your ad opens with, if you are a video editor using Premiere Pro, you have just targeted video editors who use Premiere Pro. Not with a setting. With a sentence. People who are not editors scroll past, and that is useful information. Editors stop, and that is stronger information. Within days the system knows who this ad is for, with more detail than any interest list could ever describe.

Here is the trap, and it is the single reason people try broad and decide it is nonsense. Broad targeting and ads written for 1 specific group of people are 1 move, not 2. They are halves of the same thing. Go broad while running 2 general ads and you have made your account worse, because you removed the only steering you had and put nothing in its place.

Broad targeting with 2 general ads is worse than the old way. Broad with 12 specific ads is much better.

This is also why so much of my work now sits upstream of the ad account. The ad writing, the voice, the knowledge about the business, all of it lives in the command center I build for a client, so a sentence written for 1 kind of buyer can be made in minutes instead of weeks. Same reason I keep saying the words matter more than the settings. The place people find you is shifting to the words too, and people now ask a chatbot who to hire instead of scrolling a list.

05The 1 case where narrow still wins

I would rather be honest than absolute, so here is the real exception.

If the total number of people who can legally or practically buy from you is genuinely tiny, and you can prove it is tiny, narrow wins. Say you only sell to people holding a specific licence in 1 small country. Broad would spend most of your money reaching people who cannot buy, and no clever first line filters them out fast enough.

The test is simple. Is your constraint a fact or a guess? Only a licence holder can buy this is a fact. My buyer is probably interested in business is a guess wearing a fact's clothes. The second kind is what costs you money.

If you cannot write a sentence that proves the limit, it is a guess, and it is costing you buyers quietly.

06And what about lookalikes?

Lookalike audiences were a smart answer to a problem that has mostly gone away. You asked the platform to find people similar to the ones who already bought. That is exactly what the delivery system now does on its own, from live data, faster, with fresher information.

When you build a lookalike from a list, you are freezing a photo of who bought in the past and asking the system to keep looking backwards at that photo. Broad lets it look at what is happening this week.

07How to move across without panicking

If your account runs on detailed targeting today and it is making money, do not blow it up on Monday morning because of an article. Not even mine. Do it in order.

One more warning on how you judge it. Do not use the platform's own reported return as your only scoreboard. Last click reporting will read both campaigns wrong. Keep your own count of calls booked and clients won, in your own sheet or your own dashboard. In the command centers I build this is the first screen the owner sees, because the number you trust is the one you own.

The summary is 1 line. Stop describing your buyer to the platform. Start describing your buyer to your buyer. The settings panel was never the interesting part of this job, and now it is barely part of it at all.

08Questions, answered