Audience Targeting

Broad Targeting vs Lookalike Audiences

In short: Both aim to find new buyers rather than people you already know, but they start from different places. Lookalike audiences model a percentage of a country against a seed list you choose, so quality depends entirely on how good that seed is. Broad targeting has no seed at all - it relies on live conversion signal from the optimization event instead of a static model built at one point in time. Lookalikes still add value when your seed is small and high-quality; once an account has enough conversion volume, broad targeting frequently matches or beats them because it reacts to current data rather than a snapshot. Pick a lookalike when your seed is better than what broad would find on its own; pick broad when it wouldn't be.

By the AdFlint research team · Fact-checked against current Google and Meta platform behavior · Last reviewed July 2026

Broad Targeting

Runs a Meta ad set with little definition beyond location, age, and gender, leaving the delivery system to find buyers from conversion signal alone.

You leave interests and behaviors empty and let Meta's optimization use pixel and Conversions API events to locate people likely to trigger your chosen event. It has become the default posture on Meta because the system finds pockets no manual interest stack would ever describe. It needs enough conversion volume to learn, and advertisers frequently kill it mid-learning before delivery has stabilized.

Full definition

Lookalike Audiences

Meta audiences modelled on a seed list, matching users who resemble your source on the signals Meta holds, sized as a percentage of a chosen country.

You pick a source - purchasers, high-value customers, engaged visitors - and a percentage; smaller percentages mean closer resemblance and less reach, larger ones the reverse. A value-based source lets Meta weight by revenue. Their edge has narrowed as broad targeting and Advantage+ audience improved. The persistent error is seeding from a large low-quality list, which models the wrong person very precisely.

Full definition

Side by side.

The differences that actually change what happens in your account.

	Broad Targeting	Lookalike Audiences
What it needs to exist	Nothing beyond enough conversion volume for the optimization event to learn from.	A seed source - a custom audience of purchasers, high-value customers, or engaged visitors - and a percentage size.
How it stays current	Continuously - it reacts to live conversion events on every delivery.	Only as current as the seed; a static seed produces a lookalike that ages with it.
What quality depends on	The volume and cleanliness of your optimization event.	The quality of the seed list - a large low-quality seed models the wrong person precisely.
Size control	No size lever; the pool is the full eligible population from the start.	You choose the percentage - smaller means closer resemblance and less reach, larger the reverse.
Value-weighting	Not applicable directly, though the optimization event itself can be a value-based conversion.	A value-based seed lets Meta weight the model toward your highest-value customers, not just any converter.
Where its edge has narrowed	Not applicable - broad is the option that has been gaining ground.	Its advantage over broad targeting has shrunk as broad and Advantage+ Audience improved at finding buyers without a seed.
Failure mode	Killed before enough conversion volume accumulates for the pattern to stabilize.	Seeded from a large, low-intent list, which produces a confident but wrong model of who to target.

What actually separates them.

A lookalike is a static model built from a seed at build time; broad targeting is a live process that reacts to conversion events as they happen.

Lookalike quality is bounded by seed quality - a bad seed cannot be fixed by widening or narrowing the percentage - while broad targeting's quality is bounded by conversion volume and event cleanliness.

Lookalikes offer a size lever through the percentage setting that broad targeting does not have, since broad starts from the full pool with no dial to turn.

A value-based lookalike seed lets you weight toward your best customers specifically, which broad targeting cannot do unless your optimization event itself carries value.

As accounts accumulate more first-party conversion data, broad targeting tends to close the gap with lookalikes, because it is drawing on the same underlying pool of buyer signal without needing a separate model.

Which one should you use?

Use Broad Targeting when

You have enough conversion volume for the algorithm to work from live signal rather than a static model.
You do not have a seed list good enough to justify building a lookalike from it.
You want delivery that adapts as your customer base and product mix change, without rebuilding an audience.
Your lookalike tests have not outperformed an open ad set, so the modeling step is not adding value.

Use Lookalike Audiences when

You have a small, high-quality seed - your best customers or highest-value purchasers - worth modeling from directly.
You want to control audience size precisely with the percentage lever rather than leaving it fully open.
You have value data on your customers and want the model weighted toward the most valuable ones, not just any converter.
You are prospecting in a new market and want a starting point closer to your existing customer profile than a cold, broad launch.

Common questions.

Are lookalike audiences still worth building given how good broad targeting has gotten?

They can be, mainly when your seed is small, clean, and clearly better than average - a list of your highest-value repeat customers, for instance. If your seed is just everyone who ever converted, the resulting lookalike often performs close to what broad targeting finds on its own, so the extra step may not be worth maintaining.

What percentage should I use for a lookalike?

Smaller percentages model closer resemblance to the seed at the cost of reach, larger percentages trade resemblance for scale. There is no universally correct number - it depends on how much volume you need versus how tightly you want to match the seed, and it is worth testing more than one size against your actual conversion results.

Can I combine a lookalike with broad targeting?

You can enter a lookalike as an Advantage+ Audience suggestion, which gives it priority in early delivery before the system expands past it, effectively blending the two. That is different from running broad with nothing entered, since the lookalike still provides a starting point the algorithm leans on before it opens up.

Or stop choosing between them.

AdFlint picks the setting, writes the ads, and keeps optimizing inside the Google and Meta accounts you already own.

Try AdFlint free See how it works

Related comparisons

All comparisons