Meta Ads Audience Targeting Best Practices (2026)

By Kristians, Founder, AutoAdy. Six years buying Meta ads — first for agency clients, then for his own ecommerce and lead-gen brands.

Last updated:

What changed: Made the opening H2 a question, cited Apple's tracking framework directly for the iOS 14.5 claim, and added the overlap and saturation tools.

Broad targeting (age, gender, location only) now outperforms detailed interest targeting for 60% of Meta ad accounts since iOS 14.5 reduced tracking data. Lookalike audiences remain effective at 1-3% sizes. Custom audiences from email lists and website visitors deliver the lowest CPA but limited scale.

Key Takeaways

  • ✓Broad targeting outperforms interest stacks for 60% of accounts post-iOS 14.5
  • ✓Lookalike audiences at 1% deliver 25-35% lower CPA than interest targeting
  • ✓Custom audiences convert at 3-5x cold audience rates but have limited scale

What targeting works best on Meta ads now?

iOS 14.5 changed Meta targeting fundamentally. With less tracking data, Meta's broad-match algorithm often beats manually stacked interests. Hierarchy: Custom Audiences (website visitors, email lists) for retargeting at 3-5x better rates. Lookalikes at 1-3% for prospecting at 25-35% lower CPA. Broad targeting for scale with strong creative and 50+ weekly conversions. Interest targeting for new accounts or niche B2B.

Post-iOS 14.5 Strategy

Before iOS changes, stacking 10-15 interests was standard. Now Meta estimates 30-40% of interest signals are degraded. The shift: rely on first-party data (email lists, pixel events), use broader audiences for algorithm room, and let creative do the targeting. Strong creative that speaks to your ideal customer self-selects the right audience with broad targeting. Refresh Lookalike seed audiences quarterly.

How AutoAdy Helps with Targeting

AutoAdy detects audience overlap between ad sets that causes self-competition. It monitors saturation (rising frequency + declining CTR) and recommends when to expand or refresh targeting.

What we see in AutoAdy accounts

Figures below come from Meta ad accounts AutoAdy has audited or monitors. They are directional, not a controlled study.

  • •Audience overlap detector prevents self-competition between ad sets
  • •Saturation alerts when frequency exceeds 3.0 and CTR drops below 50% of peak
  • •Targeting effectiveness comparison across broad, interest, and lookalike audiences

Sources

FAQ

Common questions

Is interest targeting dead after iOS 14.5?

Not dead, but degraded. It still works for new accounts and niche B2B. For established accounts with conversion data, broad or lookalike targeting outperforms interests by 15-30% on CPA.

What lookalike audience size should I use?

Start with 1% for best quality (25-35% lower CPA). Scale to 3-5% when 1% saturates. Above 5% performs similarly to broad. Use purchase-based seeds over page views.

How often should I refresh custom audiences?

Upload customer emails quarterly. Website visitor audiences should use 30-60 day windows. Refresh Lookalike seeds whenever your customer base grows by 20%+.

Ready to try audience overlap detection that actually works?

Start free. Connect your Meta account. See your wasted spend in 60 seconds.