TL;DR
In 2026, your targeting inputs are suggestions, not instructions. Meta's delivery system does the finding; your job is giving it room to work and signal to learn from. Broad targeting is the default for most accounts because it gives the system the largest pool and the cleanest data. Advantage+ audience expands on your suggestions automatically and is right for most prospecting. Detailed targeting still earns its place for niche products, tight geos, and strict exclusions. And the biggest targeting lever isn't in the ad set settings at all: it's the creative, which qualifies the audience before they ever click.
We spend less time on targeting than on lunch
Here's our targeting process: set it broad, add sensible exclusions, move on to the creative, which is where the actual money gets made. Total time spent: minutes.
This used to be controversial advice. Now it's just what works. Meta's delivery system in 2026 is extraordinarily good at finding buyers inside a broad pool, provided you give it two things: enough budget to learn, and conversion signal it can trust. Every interest you stack, every narrow audience you carve out, shrinks the pool and fragments the signal. You're not helping the system. You're handcuffing it and calling it strategy.
That's not to say targeting is irrelevant. It's to say the skill moved. The advertisers still winning on targeting in 2026 aren't the ones with the cleverest interest stacks. They're the ones who know exactly when to get out of the way and when to intervene.
Broad is the default (yes, really)
Broad targeting, meaning no interests, no lookalikes, just geo, age, and gender guardrails, is the starting point for most accounts now. The reasons are mechanical, not philosophical.
First, pool size. A broad audience gives the system millions of people to test against. It finds patterns you'd never think to target: the buyers who don't follow any of the obvious interest pages, the ones whose behavior signals intent that no interest category captures.
Second, signal quality. Every conversion inside a broad ad set teaches the system about your actual buyers, unfiltered by your assumptions. Narrow targeting teaches the system about buyers filtered through your assumptions, which is a smaller and often wronger lesson.
Third, stability. Broad audiences don't exhaust the way narrow ones do. A 2-million-person interest audience burns out. A 40-million-person broad pool keeps delivering.
The objection we hear: "but our product is niche." Niche products often do fine on broad, because the delivery system is good at finding the niche inside the crowd. The exception is genuinely tiny addressable markets, a few hundred thousand people, where broad is just wasteful. For everyone else, start broad and let the data argue otherwise.
Advantage+ audience: what it actually does
Advantage+ audience takes whatever targeting suggestions you provide and expands beyond them automatically to find conversions. Think of your inputs as a starting hint, not a fence.
For most prospecting, leave it on. It does what a good media buyer wishes they could do manually: follow the signal wherever it goes, without waiting for someone to approve a new interest. The accounts that fight it, clinging to manual targeting "for control," are usually the ones with the most to prove and the least signal to prove it with.
When to switch it off or constrain it: strict geographic requirements (licensed services, local businesses), age-restricted products, regulated verticals, or any case where showing the ad to the wrong person is a compliance problem, not just wasted spend. Control matters when the cost of being wrong is legal, not when it's a few dollars of inefficient delivery.
One honest caveat: Advantage+ audience needs conversion volume to work with. On tiny budgets with a handful of conversions a week, there's not much signal to expand on, and your suggestions carry more weight by default. That's fine. It's still not an argument for a 14-interest stack.
When detailed targeting still earns its place
Detailed targeting isn't dead. It's just demoted from "the strategy" to "a tool for specific jobs":
Genuinely niche markets. If your total addressable market is small, broad wastes money finding that out. Narrow it deliberately.
Language and cultural targeting. Selling in Spanish to US Hispanics, or running separate creative per language, still benefits from explicit targeting. The system can't read your mind about language.
Exclusions, always. Exclude past purchasers from prospecting (unless you want repeat buyers, then segment them deliberately). Exclude email subscribers from cold campaigns when it makes sense. Exclusions are the highest-ROI targeting work you can do, and most accounts skip it entirely.
Testing hypotheses. Want to know if an interest audience beats broad for your product? Test it properly, ABO, equal budgets, adequate time. Sometimes the answer is yes. The point was never "never use interests." The point is "stop assuming."
What doesn't count: stacking 14 related interests in one ad set because it felt thorough. That's not targeting, it's decoration.
Creative is the new targeting
Here's the real shift, and it's the one most advertisers still haven't internalized. In 2026, the ad itself does most of the qualifying.
A UGC-style video where a woman in her thirties talks about postpartum hair loss is targeting women in their thirties with postpartum hair loss, regardless of what the ad set settings say. The hook, the spokesperson, the language, the problem named in the first three seconds: that's your targeting now. The delivery system watches who stops, who watches, who clicks, and finds more of them.
This is why creative variety matters more than audience variety. Ten ad sets with one creative is ten ways to show the same message. One ad set with ten creatives is ten different doors into the market. The accounts we see scaling are the ones that stopped slicing audiences and started multiplying angles.
It also means your creative briefs need to think about qualification explicitly. Who is this ad for? Say it in the first three seconds. The system will handle the rest.
Lookalikes: the honest verdict
Lookalikes used to be the cheat code. Upload buyers, get more buyers, take the afternoon off. In 2026, the honest verdict is: sometimes useful, often just expensive broad.
They still work when the seed is large and high-quality. Ten thousand purchasers makes a decent seed. Two hundred email subscribers does not, and a 1% lookalike of a weak seed is just a random audience with extra steps.
Test lookalikes against broad head-to-head before committing budget. In many accounts we see, broad wins or ties, and ties mean broad wins on simplicity. Keep lookalikes where they demonstrably beat broad. Retire them where they don't. And please stop running seven lookalike percentages of five seeds in separate ad sets. That's the lookalike zoo, and the animals are all hungry.
Retargeting in 2026: smaller, still worth it
Retargeting pools are smaller than they used to be, thanks to attribution windows and privacy changes, and the glory days of printing money on 30-day website visitors are over. But retargeting still earns a dedicated slice of budget, usually 10-20%.
Keep it simple: one campaign, stacked engagers (video viewers, page engagers, site visitors), creative that acknowledges the relationship ("still thinking it over?"). Don't build six retargeting campaigns. Don't retarget with the same creative as prospecting and wonder why it underperforms. And cap frequency, because nothing burns goodwill like following someone around the internet for 45 days.
FAQ
Is broad targeting better than interest targeting in 2026? For most accounts, yes. Broad gives Meta's delivery system the largest pool and cleanest signal. Interest targeting still has uses for niche markets and hypothesis testing, but it shouldn't be the default.
What does Advantage+ audience do? It takes your targeting suggestions and automatically expands beyond them to find conversions. Leave it on for most prospecting. Turn it off when compliance, licensing, or strict geo requirements make wrong-person delivery costly.
Are lookalike audiences still effective? Sometimes. They need large, high-quality seeds to beat broad, and many don't. Always test lookalikes against broad before committing budget, and retire the ones that lose.
Should I use detailed targeting at all in 2026? Yes, for specific jobs: niche markets, language targeting, exclusions, and deliberate tests. What doesn't work is detailed targeting as the default strategy for every campaign.
What is "creative is the new targeting"? The idea that the ad itself qualifies the audience. The hook, spokesperson, and problem named in the first seconds determine who stops and engages, and the delivery system finds more of those people. Creative variety now matters more than audience variety.
How should I structure targeting for a new account? Start broad for prospecting with Advantage+ audience on, add a simple retargeting campaign with stacked engagers, exclude purchasers where appropriate, and put your effort into creative variety instead of audience slicing.
Written by the MetaMaxd team. We run Meta ads for brands that are tired of agency theater: no vanity metrics, no "hacks," just accounts engineered to make money. If your targeting strategy is a 14-interest stack from 2019, we should talk.
