Facebook Ads Targeting in 2026: How to Find the People Who Actually Buy

Meta gives you four ways to pick who sees your ads: Core Audiences, Custom Audiences, Lookalike Audiences, and Advantage+ automation. Each one works. Each one fails if you use it wrong. This guide covers all four, explains when to use each, and lays out targeting strategies based on what you actually spend per month.

Core Audiences: The Starting Point

Core Audiences let you define who sees your ad by demographics, interests, and behaviors. You pick the filters. Facebook finds people matching them.

Demographics

• Age and gender: Narrow these only when your product genuinely skews. A skincare line for women 25-45 makes sense. A SaaS tool for everyone does not need age limits.
• Location: Country, state, city, or radius targeting. International campaigns should separate tier-1 (US, UK, CA, AU) from tier-2 geos. CPMs differ by 3-10x.
• Language: Target by language when running multilingual campaigns in the same country. Spanish speakers in the US are a separate ad set.

Interest Targeting

Facebook tracks what people engage with and groups them into interest categories. Some interests work. Others are junk.

• Specific brands work: People who follow Shopify, Klaviyo, or specific industry tools have real purchase intent.
• Broad categories fail: "Business" or "Technology" are too wide. You pay for millions of irrelevant impressions.
• Competitor targeting works: Fans of your direct competitors already want what you sell.
• Stack 3-5 related interests: More than that triggers AND logic and shrinks your pool below useful size.

Behavior Targeting

Behaviors come from real actions: purchase history, device usage, travel patterns. They cost more per impression but convert better than interests alone.

• Engaged Shoppers: People who clicked "Shop Now" on Facebook ads recently
• Frequent Travelers: Pulled from location data and travel app usage
• Small Business Owners: Based on page admin status and business activity
• Early Technology Adopters: Based on device upgrade patterns

Combine one behavior filter with 2-3 interests. That gives the algorithm enough room while keeping your audience relevant.

Custom Audiences: Your Warmest Traffic

Custom Audiences retarget people who already touched your business. They convert at 3-5x the rate of cold traffic. The person saw your brand. Now you close them.

Website Custom Audiences

Built from your Meta Pixel data. You need the pixel installed and firing correctly before these work.

• All website visitors (180 days): Broadest retargeting pool. Use for awareness or gentle reminders.
• Product page viewers (30 days): High intent. Show them the exact product they viewed with a reason to buy now.
• Add-to-cart abandoners (14 days): Hottest audience. They wanted your product. Hit them with urgency or a small discount.
• Purchasers (180 days): Use for upsells, cross-sells, and exclusions from cold campaigns.

Customer List Audiences

Upload your email list, phone numbers, or customer IDs. Facebook matches them to profiles. Match rates run 40-70% depending on data quality.

• Clean your list before uploading. Remove duplicates, fix formatting.
• Include email, phone, first name, last name, and country for best match rates.
• Segment your list: buyers, leads, high-value customers. Each segment gets different ads.

Engagement Audiences

• Video viewers: People who watched 25%, 50%, 75%, or 95% of your videos. 75%+ viewers are gold for retargeting.
• Instagram engagers: Anyone who liked, commented, saved, or messaged your IG profile.
• Lead form openers: People who opened but didn't submit your lead gen form.
• Facebook Page engagers: Weaker signal than video or IG, but still warmer than cold.

Lookalike Audiences: Scaling What Works

Lookalike Audiences find new people who resemble your existing customers. The algorithm analyzes your source audience and finds similar profiles across Facebook's user base.

Building Effective Lookalikes

• Best sources: Purchasers, high-value customers, repeat buyers. The stronger the signal, the better the lookalike.
• Weak sources: Page likes, broad website visitors, video viewers. These create diluted lookalikes that burn budget.
• Minimum seed size: 100 people, but 1,000+ gives the algorithm enough data to find real patterns.
• Value-based lookalikes: Upload customer lifetime value with your list. Facebook prioritizes finding people who match your best spenders.

Choosing the Right Percentage

• 1% Lookalike: Tightest match. Best quality, smallest pool. Use when testing or on limited budgets.
• 1-3% Lookalike: Good balance of quality and scale. Standard choice for most campaigns.
• 3-5% Lookalike: Wider net. Lower match quality but enough volume for aggressive scaling.
• 5-10% Lookalike: Near-broad targeting. Only useful when spending $500+/day and feeding the algorithm heavy creative volume.

Start at 1%. Prove ROAS. Then expand the percentage as you increase budget.

Advantage+ Audience: Let the Machine Decide

Advantage+ is Meta's AI-powered targeting system. You provide "audience suggestions" instead of hard constraints. The algorithm uses your pixel data, creative signals, and conversion history to find buyers.

When Advantage+ Beats Manual

• You get 50+ conversions per week (enough data for the ML model)
• Your creatives are strong (the algorithm reads your ad to find the right people)
• You sell e-commerce products (Advantage+ Shopping Campaigns crush manual setups at scale)
• You've exhausted your manual audiences and need fresh reach

When Manual Targeting Still Wins

• New pixel with zero conversion data
• Niche B2B products with tight ideal customer profiles
• Testing phase when you need controlled experiments
• Geo-restricted offers where the algorithm might waste budget on wrong locations

Advantage+ Shopping Campaigns (ASC)

ASC removes ad set-level targeting entirely. You set a budget, upload creatives, and let Meta's system split traffic between prospecting and retargeting automatically. E-commerce brands spending $5,000+/month often see 20-40% better ROAS with ASC compared to manual campaign structures.

You lose control with ASC. No audience exclusions, no manual budget split between prospecting and retargeting. Thin pixel data means ASC burns through cash while it learns.

Targeting Strategy by Budget

$500-$2,000/month: Testing Phase

1. Create 3-5 interest-based ad sets with different audience angles
2. Add one broad (no interests) ad set with Advantage+ enabled
3. Run 3-5 different creatives in each ad set
4. Kill underperformers after $20-30 spent with no results
5. Move budget to winners. Scale them to $50/day before adding new tests.

At this budget, manual targeting with tight interest stacks gives you more control than Advantage+. You need to learn what works before handing the keys to the algorithm.

$2,000-$10,000/month: Scaling Phase

1. Build Custom Audiences from your pixel data (you have traffic now)
2. Create 1-3% Lookalikes from purchasers and high-intent visitors
3. Layer a retargeting funnel: visitors > add-to-cart > purchasers
4. Test 10+ new creatives per week across your best audiences
5. Increase budgets 15-20% per day on profitable ad sets

Your pixel knows who buys. Lookalikes find more of them.

$10,000+/month: Volume Phase

1. Launch Advantage+ Shopping Campaigns alongside your manual structure
2. Test fully broad targeting (zero interests, zero behaviors) with strong creatives
3. Create 5-10% Value-Based Lookalikes for volume
4. Rotate 20-50 creatives to fight ad fatigue across large audiences
5. Expand to new countries with separate campaign structures

At high spend, the algorithm needs room. Tight targeting limits delivery. Broad targeting with strong creatives and deep pixel data outperforms micro-targeted ad sets.

7 Targeting Mistakes That Burn Budget

1. Audience too small. Under 500K people gives the algorithm nowhere to go. It either spends slowly or overpays for every impression. Aim for 2-10M on cold campaigns.
2. Too many interest layers. Each interest you stack narrows the pool. Three interests with AND logic can shrink a 10M audience to 200K. Use OR logic (separate interests in the same field) or split them into different ad sets.
3. Skipping Advantage+ entirely. Media buyers who refuse to test AI targeting leave money on the table. Run one Advantage+ campaign alongside your manual structure. Compare results after $500 spent on each.
4. No exclusions on retargeting. If you run cold and retargeting campaigns without excluding purchasers from cold and cold audiences from retargeting, you pay twice to reach the same people. Always layer exclusions.
5. Lookalikes from weak sources. A 1% Lookalike from page likes finds people who click "Like." It does not find buyers. Build lookalikes from actual purchasers or at minimum from add-to-cart events.
6. Blaming targeting when creatives fail. Targeting determines who sees your ad. Creative determines whether they act. If your CTR sits below 1%, the problem is your ad, not your audience. Fix creatives first.
7. Running on personal accounts at scale. Personal ad accounts cap daily spend at $250-1,000. They trigger restrictions when you scale fast. Agency ad accounts remove spending limits and reduce ban risk. Scaling above $100/day on a personal account is asking for trouble.

The Pixel Is Everything

Your Meta Pixel feeds every targeting method above. Custom Audiences pull from pixel events. Lookalikes build on pixel data. Advantage+ optimizes from pixel conversion signals.

If your pixel fires wrong events, double-counts conversions, or misses key actions, your targeting degrades across the board. Audit your pixel setup before spending a dollar on audience testing.

Pixel Health Checklist

• PageView fires on every page load
• ViewContent fires on product pages with correct product IDs
• AddToCart fires once per cart addition with value and currency
• Purchase fires once per completed transaction with total value
• Conversions API (CAPI) runs server-side alongside browser pixel for redundancy
• Event deduplication prevents double-counting between browser and server events

Frequently Asked Questions