The Complete Meta Ads Playbook: ABO → CBO → Cost Cap Framework

Meta rebuilt their entire ad system in 2024-2025. If you're still running campaigns the old way, you're leaving money on the table. This playbook covers everything you need to know about the ABO → CBO → Cost Cap framework that's driving results for DTC brands.

What's Running Your Ads in 2025

Before diving into campaign structure, you need to understand the three systems powering Meta's ad delivery.

What This Means for Your Strategy

• Run more ads (8-15 per adset) — The old "3-6 ads" rule is dead. More options = better matching.
• Strive for creative diversity — Different people, locations, formats, angles. Andromeda pattern-matches visuals.
• Same visuals = same ad to Andromeda — Same actor + same scene with different scripts? Andromeda sees one ad. But you still get conversion data per offer.

Your Key Numbers

Before running any campaigns, you need to calculate your key metrics. These numbers will drive every decision.

How to Calculate Your Break-Even ROAS

Break-even ROAS tells you the minimum return needed to not lose money on ad spend.

How to Calculate Your Target ROAS

Target ROAS is your profitability threshold—the return you need to hit your margin goals.

How to Calculate Your Target CPA

Target CPA is the maximum you should pay to acquire a customer while staying profitable.

Example Key Numbers Summary

Budget Allocation by Campaign Type

The 3-Campaign System

Three campaigns, three jobs. Here's how ads move through the system.

Campaign Structure Overview

Testing Campaign Rules

The Testing campaign is where you discover what works. Here are the exact rules for when to graduate, keep testing, or kill an ad.

Decision Framework

Scaling Campaign Rules

Once ads prove themselves in Testing, they graduate to Scaling for maximum volume.

Cost Cap Campaign Rules

Cost Cap is your profitability protection layer. Only use it after you've proven your ads work.

Step-by-Step: Your First Week

Follow these steps exactly to set up your three-campaign system.

Step 1: Create Your Testing Campaign

Campaign name: [Brand] - ABO Testing - Prospecting
Objective: Sales
Campaign budget optimization: OFF (this makes it ABO)
Adset budget: Ads × $12.50/day (example: 12 ads = $150/day)

Step 2: Create Your Scaling Campaign (Empty for Now)

Campaign name: [Brand] - CBO Scaling - Prospecting
Campaign budget optimization: ON
Campaign budget: $1,000/day (start here, scale up later)
Leave it empty — winners will graduate here

Step 3: Wait 7 Days — DO NOT TOUCH

Step 4: Day 7 — Review Results

Apply the rules:

• ROAS ≥ Target ROAS + 4 purchases + CPA ≤ Target CPA? → Graduate to Scaling
• ROAS between Break-Even and Target? → Keep testing 7 more days
• ROAS < 80% of Break-Even + spent ≥ 6× Target CPA? → Kill (pause adset)

Step 5: Graduate Winners

Go to winning adset → Click Duplicate → Destination: Scaling campaign

Step 6: Repeat Weekly

How to Organize Your Adsets

This is the most important section. The hierarchy is: Adset → Entity ID → Ad

Understanding the Structure

• Adset = Budget container, grouped by FORMAT
• Entity ID = Visual concept, defined by ACTOR + SCENE
• Ad = Individual creative with different SCRIPT + OFFER variations

Visual Example

Campaign: brand_ABO_Testing_Broad_Tier1_120624
ADSET: Video_UGC_18+_costcap-72_7dayclick_A+placements
    ├── EID01 = Sarah 60F + Kitchen
    │       ├── EID01_video_ugc-testimonial_sku-product-a_90off_PainPointA
    │       ├── EID01_video_ugc-testimonial_sku-product-a_BOGO_PainPointB
    │       └── EID01_video_ugc-testimonial_sku-product-a_30off_Quality
    ├── EID02 = Linda 29F + Living Room
    │       ├── EID02_video_ugc-unboxing_sku-product-b_90off_Transformation
    │       └── EID02_video_ugc-unboxing_sku-product-b_BOGO_PainPointA
    └── EID03 = Mike 48M + Office
            ├── EID03_video_ugc-testimonial_sku-product-c_60off_HomeDecor
            └── EID03_video_ugc-testimonial_sku-product-c_90off_PainPointC

Why Format-Based Adsets (Not Demographic)?

You might think: "I should create adsets by demographic—Female 45-55, Female 55-65—so I can show each age group actors that look like them."

1. You're using broad targeting anyway. Andromeda already knows who responds to what.
2. Splitting by demographic fragments your signal.
3. Andromeda is 10,000x smarter. Let it figure out which user responds to which actor.
4. FORMAT actually matters for delivery. Video vs static get different placements and costs.

Recommended Adset Structure

What You Learn at Each Level

• Adset level — Which FORMAT wins (UGC vs polished vs static)
• Entity ID level — Which VISUAL (actor + scene) wins
• Ad level — Which MESSAGE (script + offer) wins
• Andromeda learns — Which USER responds to which actor demographic—automatically

Entity ID Rules + Naming

When to create a new Entity ID vs. add an ad to an existing one.

The Rule: Entity ID = Actor + Scene

Naming Convention

Ad Name Format: [EID]_[type]-[style]_sku-[sku]_[offer]_[angle]_[a+settings]_[initials]

Example Breakdown:

EID01_video_ugc-testimonial_sku-product-a_90off_PainPointA_a+copy_pk
EID01 = Entity ID 01 (Sarah 60F + Kitchen)
video = ad type
ugc-testimonial = style
sku-product-a = product being advertised
90off = offer (90% off)
PainPointA = messaging angle
a+copy = using Advantage+ copy
pk = media buyer initials

Targeting & Retargeting

How GEM/Lattice/Andromeda changes the retargeting game.

The 2025 Reality

• Andromeda already knows who visited your site—it pattern-matches behavior across platforms
• Broad targeting naturally prioritizes high-intent users
• Excluding site visitors fragments signal
• Lattice transfers learnings across campaigns

Recommended Targeting for All 3 Campaigns

Second Chance Protocol

Don't kill products prematurely. Separate CREATIVE failure from PRODUCT failure.

The Problem

The Rule: Separate Creative Failure from Product Failure

When an ad fails in Testing:

1. Kill the AD—not the PRODUCT
2. Log the product in a "Retest Queue"
3. Choose a retest path

Retest Option A: New Creative Approach

Retest Option B: Cost Cap Validation

Use Cost Cap as the "final arbiter" for products that failed in Testing. Cost Cap only delivers when Meta believes it can convert at your target, making it a pure efficiency filter.

Cost Cap Retest Rules (After 14 Days)

Product Death Rule

A product is only declared "dead" after:

• 2+ distinct creative approaches failed in ABO Testing

• 1 creative failed in Testing AND no Cost Cap delivery after 14 days

Quick Reference Cheat Sheet

Frequently Asked Questions

Q: Same video, different offers (25% off vs BOGO)—same adset or different?

A: Same adset. Andromeda sees visually identical ads as one ad anyway. Put offer variations together, graduate them together. You'll still see which offer converts best at the ad level.

Q: When exactly do I remove an ad from Testing?

A: After 2 weeks of solid performance in Scaling. Not before. Keep it in Testing as a safety net while it proves itself.

Q: What if I only have budget for one campaign?

A: Start with Testing only. Add Scaling when you have winners and more budget. Cost Cap comes last—only after $1,500+/day total spend.

Q: Won't I compete with myself running the same ad in multiple campaigns?

A: No. Meta's auction picks the most efficient bid. You don't pay twice. During the 2-week proving period, keeping ads in both provides a safety net.

Q: What Cost Cap bid should I set?

A: Your actual Target CPA (calculated as AOV ÷ Target ROAS). Don't bid lower hoping for a deal. Meta's auction rewards honest bids.

Conclusion

This ABO → CBO → Cost Cap framework provides a systematic approach to Meta advertising that works with the platform's AI systems rather than against them. The key principles:

1. Let the algorithm learn — Don't touch campaigns during learning periods
2. Graduate winners systematically — Use clear rules, not gut feelings
3. Protect profitability — Cost Cap is your safety net for proven ads
4. Separate creative from product failure — Give products multiple creative chances