Blog: Creative is the New Targeting: How to Master “Entity IDs” and Unlock Scale

Under the Andromeda architecture, the definition of ‘targeting’ has expanded. While technical settings remain important for guardrails, the algorithm now uses the creative asset itself as a primary targeting signal.

For the marketing leader, this means the ‘Creative Feedback Loop’ is now as critical as the bid strategy. We are shifting from managing static audience lists to managing ‘Creative Signals’ where we use specific ad concepts to guide the AI toward specific customer pockets.

The New Math: Why Volume Equals Reach

To produce effective creative, we must first understand the system’s demands.

Meta’s new Andromeda engine works like a high-speed matchmaker. Meta’s $100 billion supercomputer scans billions of data points to find the perfect ad for a specific user at a specific moment.

Creatives act as “keys” unlocking different audience doors.

• Ad A (a fast-paced video) unlocks the “Impulse Buyer” door.
• Ad B (a detailed review) unlocks the “Skeptical Researcher” door.

If you only have 3 active creatives, you can only unlock 3 doors. Restricting volume physically limits the AI’s ability to find matches. Scaling in this ecosystem requires enough shots on goal to give the matchmaker options.

However, adding keys only works if they are distinct.

Understanding the “Entity ID” System

To maintain a high-quality user experience, Meta’s computer vision groups visually similar ads under a single ‘Creative Entity ID’.

If we launch 20 ads that look identical, the system logically groups them to prevent clutter, which limits our ability to test effectively. Our goal is to achieve Signal Diversity by creating ads that are visually distinct enough to be treated as unique entries in the auction, ensuring we get true ‘shots on goal’.

This creates two problems:

• The Duplicate Penalty: If a new ad resembles an old one, the AI blocks it from the auction to prevent clutter. You think the new creative failed. In reality, it never started.
• The “All-or-Nothing” Risk: Launching 50 lookalike ads places one bet 50 times. If that one angle fatigues, the entire account crashes.

The Fix: True Variety. Changing a background color fails to trick the AI. Unlocking new audiences requires ads that look, sound, and feel fundamentally different.

The Efficiency Frontier: Managing Frequency

Why does creative velocity matter? Because in the current ecosystem, efficiency is closely tied to freshness.

The ‘Rule of 2.5’ Research suggests a Point of Diminishing Returns: for broad audiences, efficiency is optimal up to a frequency of roughly 2.5. Beyond this, we often see a natural rise in CPMs and a dip in conversion probability. After seeing an ad 4 times, a user is 45% less likely to convert.

The Penalty

Meta charges a premium to show fatigued ads.

• Engagement Drop: A 35% drop in click-through rate is the final warning sign before performance collapses.
• CPM Penalty: The cost to show a fatigued ad can jump by 29%.

The Takeaway: You cannot “sweat” assets. You need a constant pipeline of fresh creative to replace winners before they hit the cliff.

The Blueprint: Triggers and Barriers

Beating the Entity ID penalty and the Frequency Cliff requires enforced diversity. You cannot guess these angles in a boardroom. You need scalable audience research.

This research must focus on uncovering micro-personas—specific customer archetypes defined by their unique pain points and motivations, rather than just broad demographics like “Women 25-34.”

Historically, researching micro-personas was slow and expensive. Today, AI tools allow you to analyze thousands of customer reviews, forum discussions, and support tickets in minutes. This research uncovers the specific psychological triggers and barriers that define your audience segments.

Use this data to build ads around specific micro-moments that force the AI to find new people.

1. The Trigger (Start the Journey)

These ads target the emotional moment a problem becomes unbearable.

• The Status Anxiety Trigger: “I went to my high school reunion and realized everyone had advanced further in their careers than I had. I went home that night and applied for my MBA.”
• The Immediate Pain Trigger: “I woke up with back pain for the third day in a row and realized I couldn’t pick up my toddler. That was the moment I bought a new mattress.”
• The Aspiration Trigger: “I saw a photo of myself from five years ago and realized I had lost my spark. I booked the trip to Italy the next morning.”

2. The Barrier (Close the Sale)

These ads target the specific logistical friction preventing the purchase.

• The Lifestyle Friction Barrier: “I desperately need a bathroom renovation, but I work from home and cannot have a construction crew making noise for three weeks.” (Solution: The “One Day Install” Offer).
• The Complexity Barrier: “I know I need to invest, but I’m terrified of picking the wrong stocks and losing my savings.” (Solution: The “Robo-Advisor” Simple Setup).
• The Trust Barrier: “I’ve been burned by three different skincare brands that promised results and delivered breakouts.” (Solution: The “90-Day Empty Bottle Guarantee”).

Strategy: When you vary your creative based on these specific stories, you force the AI to find different people. The person worried about “construction noise” is a different data cluster than the person worried about “price.”

Conclusion: The era of hacking the algorithm is over. Marketing leaders must now feed it. Success requires a creative pipeline that delivers enough variety to bypass Entity ID penalties and stay ahead of the Frequency Cliff.