Three Classic Deconstruction Models for Meta Ads 2025
1.[1-1-10]: The Universal Structure for Beginners, Viral Hits, and Low Budgets (In-Depth Analysis)
I.Why is 1-1-10 the “golden cold start” in Meta’s ad world?
At its core, 1-1-10 is the cleanest, most algorithm-friendly cold start structure.
It solves three core pain points:
**① Enables the system to quickly pinpoint “your target audience”
(strongest goal constraint)**
Campaign = Telling the system “What I want”
A single Campaign + Adset configuration means the system receives an extremely straightforward instruction:
— “Find me the people most likely to buy.”
This maximizes the objective function and precision, avoiding dilution from multiple Adsets.
**② Provide the system with a controllable creative pool
(Creatives determine everything)**
Newcomers often fail because:
- Too few creatives
- Too limited dimensions
- Slow learning phase
10 Ads = The system has sufficient samples to run a “creative elimination tournament” within the same audience.
Meta automatically performs three actions:
* Local ad diffusion (targeting different users)
* Rapidly identifying the top 10–20% high-quality clickers
* Locking onto micro-audiences most likely to purchase
This is why many achieve ad optimization using the 1-1-10 approach even with daily budgets of $50–150.
③ One Adset = No splitting, no model fragmentation
Common beginner mistake:
- 3 interests
- 2 LLA
- 1 broad
Placing multiple creatives in each → All get stuck in learning phase
However: Budget is already insufficient, and the model is artificially fragmented → A death sentence.
Thus, 1-1-10 is the only model enabling beginners to achieve “multiple creatives + clean structure + strong learning” simultaneously.
II.Structure Breakdown: Why 1-1-10, Not 1-1-3 or 1-1-20?
The optimal number of creatives is 10, for the following reasons:
Fewer than 6 → Insufficient coverage of user preferences
More than 12 → System learning becomes diluted, making creatives harder to optimize
10 is the system’s optimal quantity:
It provides sufficient space while preventing assets from competing for budget.
Recommended Asset Types:
- 3 Hook-focused (Strong Hook)
- 3 Lifestyle
- 2 Product Long-Selling Points
- 2 Comparison/Pain Point Solutions
III.Applicable Scenarios
① New Accounts / Cold Starts
Low budget, weak models, requiring strong signals.
② Single-Product Bestsellers (E-bikes, chairs, knives, gun safes, etc.)
Single-product positioning, more reliant on creative performance.
③ Testing Initial Audiences / Identifying Attraction Points
You don’t know who will buy, nor why users click.
④ Running ASC (Adv+ Audience)
1-1-10 is the ideal companion for ASC.
IV. Budget Recommendations
- Daily budget: $50–150 (low budget)
- or $150–300 (medium budget)
High-ticket items like e-bikes, furniture, and gun cabinets can reach $200–300.
V. When to Upgrade from 1-1-10?
Structure must be changed when:
- 2–3 stable creative assets are developed
- You plan to scale up (budget increases from $100 → $300 → $600)
- You need to test different audience segments
- You need to reduce CPA
The next step is:
👉 Upgrade to 1-3-N (the golden structure for mid-to-high average order values)
2.[1-3-N]: Core Structure for High Average Order Value and Steady Sales Volume Growth (In-Depth Analysis)
I. Why is 1-3-N the “Strongest Structure for Mid-to-High Average Order Value Products”?
You’ve already run initial creative assets. Next steps:
- Stabilize volume
- Expand audience layers
- Control CPA
- Build a global model
At this stage, the system requires more “diverse delivery logic” rather than relying solely on creative assets.
Three Ad Sets accomplish three critical tasks:
① Audience tier differentiation → Traffic quality control
This signals to the system:
Which segment represents high-quality traffic (interest-based)
Which segment is ultra-broad (Adv+)
Which segment indicates high intent (VC/IC)
This is crucial for enabling the model’s “multi-dimensional learning.”
② More stable creative allocation
Each Adset runs 3–6 creatives, preventing creative cannibalization.
③ Better suited for high-value items (chairs, e-bikes, fitness equipment, furniture)
High-value items require more touchpoints and stable audience structures.
II. 3 Common Adset Combinations (Golden Pairings)
1) Adv+ Audience (Main Volume Expansion)
System freely finds users, widest coverage.
2) Broad Interest Bundles (CPA Control)
Example:
- E-bikes: cycling, commuter, outdoor, hybrid bike
- Chairs: office worker, gaming, home office
- Gun Cabinets: hunting, firearm accessories, security
Broad interest bundles aren’t precise targeting but control “traffic quality tags.”
3) VC/IC Cold-to-Hot Conversion Logic
(Suitable for mid-to-high AOV)
- ViewContent 180
- ATC 14
- IC 30
- Engaged 365
The system favors this “weak retargeting” approach due to cleaner data.
III. Why N instead of a fixed number?
Adjust creative volume by stage:
- Medium budget: 3–4 creatives per Adset
- Scale up: Increase to 5–6 creatives
- Stable volume: Maintain 3–4 creatives
The value of N lies in flexible control:
Fewer creatives → Faster learning
More creatives → Faster audience expansion
IV. Budget Recommendations
Suitable daily budgets:
- $150–800
- Medium AOV: $300–500 (chairs, tool cabinets)
- High AOV: $500–800 (e-bikes, treadmills)
V. When to Upgrade from 1-3-N?
- When you’ve identified 1–2 absolute champion creatives
- ROAS is stable
- Budget is ready to scale from $500 → $2000/day
👉 At this stage, transition to the mature scaling structure: 1-N-1
3.[1-N-1]: High-Budget Scaling Framework for Established Brands (In-Depth Analysis)
This is the ultimate structural destination for all high-budget brands (annual spend $200K–$2M).
1-N-1 is the ultimate model for “matrix scaling,” “distributed budget expansion,” and “precision creative activation.”
I. Why is 1-N-1 the most powerful scaling structure?
Because it leverages Meta’s underlying mechanisms:
① Each Adset = Independent learning unit
Multiple Adsets = Multiple “black boxes” learning simultaneously → Scaling across distinct audience spaces.
② Each Adset contains only 1 creative → Extremely high learning efficiency
Creatives do not compete for budget.
Creative weight = 100%
The system remains unaffected by other creatives.
③ Leverage volume to build “variability” → Eliminate reliance on single audiences
N Ad Sets = Provide the system with N distinct scaling pathways.
**④ No single strong creative ≠ perpetual dominance
But the small-unit model naturally identifies its peak audience**
In high-budget eras, the greatest fear is “creative stickiness failure.”
The 1-N-1 structure ensures other ad sets in the matrix automatically fill gaps when creative performance fluctuates.
II. Suitable Products
Highly recommended for:
- High average order value ($1000–$3000)
- Established brands
- Accounts with stable conversions and multiple high-performing ads
- Accounts possessing “star creatives” (where creative quality trumps all)
III. What Should N Be?
Determined by budget:
|
|
|
|---|---|
|
|
|
|
|
|
|
|
|
Each Ad Set = 1 creative (champion-level creative)
IV. Typical 1-N-1 Structure Combination
**① Broad Ad Sets (60–70% allocation)
Multiple broad Ad Sets running concurrently**
As the system has accumulated substantial behavioral signals at this stage.
② Region-Specific Scaling Ad Sets
- US West
- US East
- US+CA
- US+EU
Targeting peak audiences in each region within the matrix.
③ Broad Interest Packs / Broad Tags
Not “precise interests,” but “tag augmentation.”
Examples:
- Outdoor Pack
- Tech Pack
- Home Lifestyle
- Shopping Heavy Buyers
V. Budget Recommendations
Suitable for daily budgets of $1000–$10,000+.
VI. When to Maintain / Reduce N?
Maintain N:
- Stable ROAS 1.5–2.5
- Steady traffic
- CPA within target range
Reduce N:
- Creative fatigue
- Budget contraction needed
- Seasonal demand decline
Your Campaign Lifecycle:
Phase 1: Cold Start (Low Budget) → Use 1-1-10
Core: Generate creative assets, capture initial conversion signals
Phase 2: Testing & Stabilization (Medium Budget) → Use 1-3-N
Core: Build multi-dimensional models, stabilize audiences, maintain controllable CPA
Phase 3: Maturity & Scaling (High Budget) → Use 1-N-1
Core: Construct distributed learning matrices, maximize scaling potential
