Audience Targeting: Cold, Warm, and Retargeting Audiences

The Three Audience Types

Every effective Meta Ads strategy uses three audience categories in different parts of the funnel:

Cold Audiences (Prospecting)

People who have no prior relationship with your brand. These are the largest audience pools but require your best creative to earn attention.

Cold audience types:

• Interest targeting: People who have shown interest in topics related to your product
• Behavioral targeting: People who have taken specific actions (online purchasers, frequent travelers, etc.)
• Demographic targeting: Age, gender, location, job title, education
• Lookalike audiences: People similar to your existing customers (covered below)

Warm Audiences (Engagement Retargeting)

People who have engaged with your content, followed your page, or visited your website but haven't converted.

Warm audience examples:

• People who watched 50% of a video ad
• People who engaged with your Instagram or Facebook page
• Website visitors (requires Pixel)
• People who opened but didn't submit a Lead Ad form

Hot Audiences (Purchase Retargeting)

People who demonstrated strong purchase intent.

Hot audience examples:

• Website visitors who viewed product pages
• People who added to cart but didn't purchase
• Past customers (for repeat purchase or upsell campaigns)
• App users who reached a specific in-app event

Lookalike Audiences

Lookalike audiences are Meta's most powerful targeting feature. You provide a source audience (your customer list, purchasers, or top website visitors), and Meta finds users with similar characteristics in the broader population.

Quality of the source audience determines quality of the lookalike:

• Best source: Customer purchase list (actual buyers are better signal than visitors)
• Good source: High-value website visitors (people who visited 3+ pages)
• Acceptable source: Email subscriber list

Lookalike percentages (1%-10%) indicate similarity:

• 1% lookalike = most similar to your source, smallest audience
• 10% lookalike = broader audience, less similar

Start with 1-3% lookalikes for highest relevance.

Audience Sizing Guidelines

The Audience Overlap Problem

Running multiple ad sets with overlapping audiences causes auction competition with yourself — you're bidding against your own ads. Meta's Audience Overlap tool shows where audiences overlap. Consolidate or exclude overlapping segments.

iOS 14+ Impact on Targeting

Apple's App Tracking Transparency (ATT) framework in iOS 14.5 significantly reduced Meta's data from iPhone users who opted out of tracking. This impacted:

• Pixel data completeness
• Retargeting audience sizes
• Attribution window accuracy

Mitigations:

• Implement Conversions API (server-side tracking)
• Use broader audience targeting rather than relying solely on Pixel-based audiences
• Extend attribution windows
• Focus on first-party data (customer lists, email subscribers)

Key Takeaways

• Build a three-tier audience strategy: cold prospecting, warm engagement retargeting, hot purchase retargeting
• Lookalike audiences based on actual customers outperform interest targeting
• Check for audience overlap before launching multiple ad sets
• Use Conversions API to mitigate iOS privacy impact on data quality