How Meta Ads targeting actually works for aesthetic clinics

If you've set up Facebook or Instagram ads yourself, you've probably spent time in the audience builder — choosing an age range, a radius around the clinic, and a list of interests like "beauty salon" or "skin care." It feels like the important part. It's presented as the important part.

It isn't, and hasn't been for some time. Meta's delivery system now does most of the targeting work itself. Your job has shifted from picking who sees the ad to telling the system what a good outcome looks like — and most clinic accounts never do that properly.

What the algorithm is actually doing

When you run a campaign, Meta isn't really showing your ad to "women 30–55 interested in skincare within 20km." It's running an enormous, continuous prediction problem: of all the people it could show this ad to right now, which ones are most likely to do the thing you said you wanted?

That last part is the hinge. If you told it you want reach, it finds cheap eyeballs. If you told it you want engagement, it finds people who habitually like and comment — a population that overlaps very little with people who book appointments. If you told it you want a conversion, and you gave it a working way to measure that conversion, it goes hunting for people who behave like people who convert.

The algorithm is extremely good at finding more of whatever you've told it to find. Most clinic campaigns are just telling it to find the wrong thing.

Why boosting a post isn't a strategy

The boost button is the clearest example. It's convenient, it's right there under the post, and it defaults to optimising for engagement.

So Meta does exactly what you asked. It finds people likely to react to a post. You get likes, a few comments, a satisfying little activity spike — and no measurable increase in enquiries, because "likely to tap a heart on a nice-looking clinic photo" and "considering booking a cosmetic appointment this month" are almost entirely different populations.

Boosting isn't useless. It's a reasonable way to put a genuinely good piece of content in front of more local people. It just isn't patient acquisition, and treating it as such is why a lot of clinic owners conclude "we tried Facebook ads and they didn't work."

The thing that actually determines performance

If the algorithm learns from conversions, then it can only learn if conversions are being tracked. This is the single biggest technical difference between clinic accounts that work and clinic accounts that don't.

In practice that means:

• The Meta pixel installed correctly on your site, firing on the pages that matter.
• A defined conversion event — a booking form submission, a completed enquiry, a click on the phone number. Something that represents real intent, not a page view.
• Enough volume for the system to learn from. A campaign generating two conversions a week gives the algorithm almost nothing to work with, which is one reason very small budgets spread across many treatments tend to fail.
• Consistency. Every time you significantly edit an ad set, learning effectively restarts. Constant fiddling keeps a campaign permanently in its least efficient phase.

When I audit a clinic's Meta account, tracking is where I look first. It's remarkably common to find a pixel that was installed once during a website build and has been silently broken since a theme update — which means every campaign since has been flying blind.

Where targeting still matters

Two places, mainly.

Geography

For a single-location clinic this is the constraint that genuinely matters. People will travel a certain distance for a consultation and not much further, and that distance is different in Newcastle than it is in inner Sydney. Setting the radius too wide is one of the most common ways regional clinics quietly waste budget — you're buying impressions in towns nobody is driving in from.

Audiences built from your own data

This is the highest-value targeting available to you, and most independent clinics never use it. A customer list uploaded from your practice management system, or an audience of people who visited a specific treatment page, gives Meta a real example of what your patient looks like. A lookalike built from actual patients consistently outperforms an interest stack assembled by guesswork.

If you go down this path, handle it properly — patient data attracts privacy obligations well beyond advertising rules, and it's worth being deliberate about what you upload and what consent covers it.

Interests: the part everyone over-thinks

Interest targeting still exists and still has a role, particularly for a new account with no conversion history to learn from. But it's a starting hint, not a strategy.

Stacking fifteen beauty-adjacent interests doesn't make targeting more precise — it usually makes the audience so narrow that delivery becomes expensive, or so broad that the interests cancel each other out. Broad targeting with strong conversion signals now frequently beats tightly-stacked interests, which is the reverse of the advice most clinic owners were given when they first set up an account.

What to check in your own account

1. Open Events Manager and confirm your pixel has fired recently. If it hasn't, nothing else on this list matters.
2. Check what each campaign is optimising for. If it says engagement, reach or traffic, that's your finding.
3. Look at your geographic radius against where your patients actually travel from.
4. Count your weekly conversions per ad set. Consistently low single digits means the system isn't learning.
5. Check whether you've ever uploaded a customer list or built a lookalike. Most clinics haven't.

None of this requires spending more. In most of the accounts I look at, the budget is adequate — it's the instructions that are wrong. Fix what you're asking the system to optimise for, make sure it can measure that thing, and the same money starts behaving very differently.

Written by Prince — Growth & Performance Marketing. Platform features and policies change frequently; verify specifics against current Meta documentation. General information about advertising practice, not legal or regulatory advice.