Broad Targeting on Meta Ads: Why Removing Interests and Lookalikes Might Be Your Best Move

For years, the gold standard of Meta advertising was layered targeting: stack 15 interests, add behavioral filters, layer in a 1% lookalike, and let the algorithm find your perfect customer inside that narrow pool. In 2026, the best-performing campaigns we manage use almost no targeting at all. Broad targeting — age, gender, country, and nothing else — consistently beats interest stacks and lookalikes in head-to-head tests across ecommerce and lead gen. Here's why the old targeting playbook is dead and how to make broad targeting work without wasting budget.

What Broad Targeting Actually Means

Broad targeting on Meta means creating an adset with only these parameters:

• Location: Country or region (e.g., United States)
• Age: Wide range (e.g., 25-65+) or 18-65+
• Gender: All, or filtered to one if your product is gender-specific
• Language: Optional — only if needed for non-English markets

That's it. No interests. No behaviors. No lookalike audiences. No custom audiences. No detailed targeting expansion toggle (it's on by default with broad anyway). The potential audience is typically 100-200 million people, and you're trusting Meta's algorithm to find the converters within that pool based solely on your creative and conversion data.

Why This Feels Wrong (And Why It Works)

Every marketing textbook says to define your target audience as narrowly as possible. The logic: why show ads to 200 million people when only 2 million are potential buyers? You're paying for impressions on the other 198 million who will never convert.

That logic made sense when Meta's algorithm was primitive. In 2018, the algorithm needed you to pre-filter the audience because it couldn't identify converters on its own. In 2026, Meta's machine learning has gotten so good at predicting conversion behavior that your interest-based targeting is often worse than what the algorithm would find on its own. You're constraining it to a subset that you think is right, when it has billions of data points suggesting a different subset.

"We ran an A/B test for a skincare brand: interest-targeted (beauty, skincare, wellness interests, 4.2M audience) vs broad (women 25-55, US, 95M audience). After 14 days and $3,200 in spend per variant, broad had a 28% lower CPA and a 15% higher ROAS. The algorithm found buyers the interest targeting would have excluded."

Why Broad Targeting Outperforms in 2026

Three structural changes to Meta's platform have shifted the advantage from narrow to broad targeting.

1. Meta's Algorithm Has More Data Than Your Interest Stack

When you target "people interested in yoga," you're relying on Meta's categorization of who likes yoga based on pages they follow, content they engage with, and apps they use. But Meta's conversion prediction model uses thousands of signals beyond interest categories — purchase history, browsing patterns, ad engagement across millions of advertisers, device behavior, app usage patterns, and more.

By narrowing to "yoga interests," you're telling the algorithm to ignore all those other signals and only look at one dimension. Broad targeting lets the algorithm use every signal it has, which produces better predictions of who will actually convert — not just who "seems like" your audience.

2. Interest Categories Have Degraded Post-iOS 14

After Apple's ATT rollout, Meta lost access to much of the cross-app tracking data that powered interest categorization. The "interested in fitness" audience in 2024 is less accurate than the same audience in 2020 because Meta has fewer data points to categorize users. Interest targeting is built on a degraded foundation, while Meta's on-platform conversion prediction model has actually improved.

3. Broad Targeting Eliminates Audience Overlap and Self-Competition

When you run 5 adsets with different interest stacks, there's massive overlap between them. The "yoga" audience overlaps heavily with the "wellness" audience, the "meditation" audience, and the "fitness" audience. You're bidding against yourself in Meta's auction for the same users. Broad targeting eliminates this entirely — one adset, one audience, no self-competition.

When Broad Targeting Doesn't Work

Broad isn't universally better. There are specific situations where you still need targeting constraints.

1. Hyper-Local Businesses

If you're a restaurant, gym, or salon, you need geographic targeting — showing ads to people 200 miles away is waste. Use broad targeting within a tight radius (10-25 miles), but don't add interest layers on top of the geo filter.

2. B2B With Extremely Narrow ICPs

If you sell enterprise software to CFOs at companies with $50M+ revenue, broad targeting on Meta will waste most of the budget on consumers. Use LinkedIn for this audience, or if you must use Meta, layer job title and industry filters. But even here, a custom audience (email list upload) plus a lookalike often outperforms interest-based B2B targeting.

3. New Pixel With Zero Conversion Data

Broad targeting works because the algorithm has conversion data to learn from. On a brand-new pixel with zero events, the algorithm has no signal to optimize against. In this case, start with interest targeting or lookalikes for the first 100-200 conversions, then test broad once the pixel has enough data.

4. Products With Genuine Gender/Age Restrictions

Menopause supplements shouldn't target 18-year-old men. Children's toys shouldn't target retirees. When your product has a genuine demographic boundary, filter for it. But don't confuse "my customers tend to be 30-45" with "I should only target 30-45." The algorithm might find profitable customers at 28 or 52 that you'd never have targeted manually.

How to Set Up Broad Targeting Correctly

Going broad isn't just "remove all targeting and pray." There's a specific setup that makes it work.

Step 1: Make Sure Your Pixel Has Data

Check Events Manager. You need at least 50 optimization events (purchases, leads, whatever you're optimizing for) in the last 7 days for broad targeting to work well. If you have fewer, build up the data with interest-targeted campaigns first.

Step 2: Use the Right Campaign Objective

Broad targeting with a "Traffic" or "Engagement" objective will send your budget to people who click or engage but never buy. Only use broad targeting with conversion-focused objectives: Sales (for ecommerce) or Leads (for lead gen). The algorithm needs a clear conversion signal to optimize against.

Step 3: Set Broad Demographics Only

In the adset, set location (country or radius for local), age range (as wide as reasonable), and gender (all, unless product-specific). Leave detailed targeting completely empty. Don't add interests "just in case" — even one interest filter constrains the algorithm.

Step 4: Let Creative Do the Targeting

This is the critical shift. When you remove audience targeting, your creative becomes the targeting. The ad itself filters who engages. A video about post-pregnancy fitness will naturally attract new mothers even if your audience is "all women 18-65." A headline about "restaurant owners" will naturally attract restaurant owners from a broad audience. The creative self-selects the right people.

This means your creative needs to be specific about who the product is for, even though your targeting is broad. Generic creative + broad targeting = wasted budget. Specific creative + broad targeting = the algorithm finds your people.

Step 5: Give It Budget and Time

Broad audiences are larger, so the algorithm needs more budget and more time to explore and find the converting pockets within the pool. Budget should be at least $50-100/day per adset for broad to work. If you're spending $20/day, the algorithm can't explore a 100M-person audience fast enough to optimize.

Give it 5-7 days before judging. Broad targeting has a longer ramp-up than interest targeting because the initial exploration phase covers more ground. CPA will be high for the first 2-3 days, then drop as the algorithm narrows in on converters.

Broad Targeting vs Advantage+ Audiences

Meta's "Advantage+ Audience" (formerly "Advantage Detailed Targeting Expansion") is different from true broad targeting. Advantage+ lets you add "audience suggestions" — interests or lookalikes that serve as starting hints — and then the algorithm expands beyond them if it finds better performance elsewhere.

In practice, Advantage+ performs similarly to broad but gives the algorithm a starting direction. Our recommendation:

• If your pixel has 500+ conversions: use true broad (no suggestions). The algorithm doesn't need hints.
• If your pixel has 50-500 conversions: use Advantage+ with 2-3 interest suggestions as starting hints.
• If your pixel has fewer than 50 conversions: use traditional interest targeting until you build the data.

The Creative-as-Targeting Framework

Since broad targeting shifts the filtering job from the audience settings to the creative, here's how to build creative that targets effectively.

Call Out Your Audience in the First 2 Seconds

"Attention restaurant owners" or "If you're a new mom struggling with..." immediately self-selects the right viewers. People who aren't in that group scroll past. People who are stop and watch. This is free targeting built into the creative.

Show the Product in Use by Your Target Customer

A video of a 30-something woman using your skincare product targets 30-something women without any audience settings. The viewer pattern-matches ("she looks like me, this might be for me") and the algorithm learns from who engages.

Use Language Your Audience Uses

Industry jargon, slang, and insider references naturally filter the audience. An ad that says "tired of your CAC climbing every quarter?" will only resonate with marketing professionals and founders. The language is the targeting.

Test Multiple Creative Angles Within the Same Broad Adset

Run 3-5 different creative angles in one broad adset. Each angle will naturally attract a different pocket of the broad audience. The algorithm will discover which pockets convert best and route delivery accordingly. This is creative-driven audience discovery — the most powerful combination of broad targeting and the algorithm's optimization capabilities.

The Bottom Line

Interest targeting was a crutch that made sense when Meta's algorithm was weak. The algorithm is no longer weak. In most accounts with sufficient conversion data, removing all interest and lookalike targeting and going broad produces equal or better results at lower CPA — because you're letting the algorithm use its full intelligence instead of constraining it to your assumptions.

The shift requires a mental model change: targeting moves from the audience settings to the creative. Your ad is the filter. Make the creative specific enough to attract the right people, and trust the algorithm to find them within a broad pool. Test it against your current interest stack for 2 weeks. The data will tell you which approach wins for your specific account.