Meta Ads: Why Your Careful Targeting Is Costing You Leads

Ask a small business owner who their Meta ads are targeting and you usually get a detailed answer: age 45 to 65, interested in travel, cruises, Expedia, TripAdvisor, income bracket, three job titles. It sounds like precision. It is usually the reason the campaign underperforms.

Meta's targeting has quietly changed. The system now finds buyers mainly by watching who responds to your ads, not by reading the interest boxes you ticked. Every box you tick shrinks the pool it can learn from. The instinct to narrow, which feels careful, is often what starves the campaign.

Why narrow targeting backfires now

Interest targeting was built when Meta had far more declared data to work with and far less signal from behaviour. Today it works the other way round: give the system a large audience and a clear conversion event, and it will find the people who convert faster than your assumptions can.

There is also a mechanical problem with small audiences. Campaigns need a run of conversions to get out of the learning phase, and a tightly sliced audience starves that. The campaign stays expensive, results swing week to week, and the owner concludes Meta does not work for their business.

The exception worth respecting: if you genuinely serve one city and one service, geography and language are real constraints, not guesses. Narrow on those. It is the interest-stacking that usually needs to go.

What to give the system instead

A real conversion event. This matters more than every targeting decision combined. If Meta only sees link clicks, it optimises for clickers. Send it the WhatsApp conversation, the form submit, the booking. Without that, targeting is guesswork whatever you pick.

Broad, with location and age as the only hard limits. For most service businesses, start with your service area, a sensible age range, and nothing else. Let the system read the responses.

Creative that filters for you. This is the part owners miss. When targeting is broad, your ad does the qualifying: say the price range, name the customer, state the service plainly. An ad that opens "Senior fares to India, handled by phone" will be ignored by everyone it does not suit, which is exactly what you want. Creative is now a targeting tool.

The audiences still worth building by hand

Two are genuinely useful, and both come from your own data rather than Meta's guesses:

• Your customer list. Upload past customers and enquiries. Use it as an exclusion so you stop paying to reach people who already bought, and as the seed for a lookalike.
• Site and engagement retargeting. People who visited your site, watched most of a video, or messaged your page. Warmer than anything you can buy cold, and cheap to reach.

Set these up early even if you will not use them for months. A customer list you upload today is far more useful than one you start building the week you need it.

How to test it without gambling the budget

Run your current setup and a broad version side by side for two to three weeks, same budget, same creative, same conversion event. Then compare at the booking level, not the lead level. Cheaper leads that never answer are the trap this whole exercise exists to avoid.

Resist editing either campaign mid-test. Every meaningful edit restarts the learning phase, and a campaign that gets tinkered with twice a week never leaves it. Set it, leave it, judge it on the month.

The uncomfortable part

Broad targeting takes control away from the owner, and that is genuinely uncomfortable. You cannot point at a settings screen and explain who the ads are reaching. What you get in return is a system that finds buyers you would never have listed, and the honest measure of whether it worked is bookings per rupee, not how sensible the targeting looked.

FAQ

Does this mean interest targeting is useless?

Not useless, just usually unnecessary and often harmful at small budgets. It still has a place for genuinely niche products where the buyer is a tiny fraction of the population. A travel service or a call centre is not that.

My broad campaign brought terrible leads. What went wrong?

Almost always one of two things: the conversion event is wrong, so Meta optimised for the wrong action, or the creative promises something vague enough that anyone clicks. Fix the event first, then make the ad say the price and the service plainly.

How big should my audience be?

Big enough that Meta reports it as broad rather than warning you it is too narrow. For most local service businesses that means your service area plus an age range and little else. If the estimate looks alarmingly large, that is usually the point.