Andromeda Creative Strategy: 21 Facebook Ad Angles to Test
A practical list of 21 genuinely different Facebook ad angles—plus a four-week testing plan, naming system, and scorecard for Meta’s creative-led delivery era.
Meta’s Andromeda retrieval system was designed to handle a much larger, more diverse pool of ads. The wrong response is to manufacture endless minor edits.
The useful response is to test different customer ideas.
An angle is the strategic reason someone should care. The hook earns the first seconds of attention. The format is how the idea is delivered.
Angle → hook → proof → offer → format
Do not confuse these layers.
21 Facebook ad angles
Problem-aware angles
1. Hidden cost: Show what the unresolved problem costs each month.
2. Common mistake: Explain the behavior keeping the customer stuck.
3. Symptom diagnosis: Connect visible symptoms to the deeper issue.
4. Risk prevention: Show what may happen when action is delayed.
5. Time drain: Quantify hours lost to the current process.
Outcome angles
6. Before and after: Contrast the starting state with the desired state.
7. Speed to outcome: Show a credible faster path without making unsupported promises.
8. Simplicity: Reduce a complex decision to a clear process.
9. Control: Help the buyer gain visibility, predictability, or confidence.
10. Identity: Show what capable owners, teams, or customers do differently.
Proof angles
11. Customer story: One problem, decision, process, and result.
12. Demonstration: Show the product or service solving the problem.
13. Expert breakdown: Diagnose a real example and explain the correction.
14. Data point: Lead with a verified number and show the mechanism behind it.
15. Objection answered: Address the strongest reason buyers delay.
Comparison angles
16. Old way versus new way: Contrast methods, not just features.
17. DIY versus specialist: Explain when each option makes economic sense.
18. Cheap versus total cost: Show how low upfront cost can create downstream expense.
Offer angles
19. Diagnostic: Audit, assessment, calculator, or checklist.
20. Low-risk first step: Trial, sample, consultation, or pilot.
21. Capacity or timing: Use a real deadline, service window, or availability constraint.
Turn one angle into three executions
Suppose a home services company chooses the hidden-cost angle.
| Execution | Hook | Proof | Format |
|---|---|---|---|
| Founder explanation | “That small leak may already be costing more than the repair.” | Cost breakdown | Talking-head video |
| Customer story | “They waited three months. Here is what changed.” | Job timeline | Testimonial montage |
| Visual diagnosis | “Three signs water is moving behind the wall.” | On-site footage | Annotated vertical video |
These are meaningfully related, but not duplicates.
A four-week testing plan
Week 1: Establish concepts
Week 2: Read the funnel
Review:
High attention with weak qualified conversion often means the promise attracts the wrong person or the page breaks continuity.
Week 3: Iterate the best concept
For the strongest angle, change one layer:
Do not change the audience, offer, page, and creative simultaneously.
Week 4: Add one new market hypothesis
Retain the winner and introduce one genuinely new angle. This creates a repeatable pipeline instead of a launch-and-panic cycle.
The creative naming system
Use:
Angle | Hook | Format | Persona | Offer | Version
Example:
Hidden Cost | Monthly Waste | Founder Video | Owner | Audit | V2
The name should let someone understand the test without opening the ad.
The creative scorecard
| Layer | Question | Metric |
|---|---|---|
| Attention | Did the opening earn interest? | Hold rate or early retention |
| Interest | Did people seek more information? | Outbound CTR |
| Message match | Did the page continue the promise? | Landing-page conversion |
| Qualification | Were the right people attracted? | Qualified rate |
| Commercial value | Did it produce customers profitably? | Cost per sale and profit |
Never promote a concept solely because it has the lowest CPM or highest CTR.
How many ads should you run?
There is no universal Andromeda number.
The practical limit depends on:
If the campaign spends $300 per week against a $150 target CPA, splitting that budget across 20 ads will not create useful learning. Run fewer, more distinct hypotheses.
If the campaign spends $30,000 per week, a larger concept portfolio may be justified.
Seven “new ads” that are not new ideas
Avoid counting these as major concept tests:
1. A different button color
2. The same video with new captions
3. A cropped version of the same image
4. A new first frame with the identical argument
5. The same testimonial in three aspect ratios
6. A minor headline rewrite
7. AI-generated scenes that repeat the same promise
They may be useful production variants. They are not seven market hypotheses.
A creative brief that improves retrieval inputs
For every concept, document:
Then produce only after the brief has a distinct answer.
The key Andromeda principle
Meta’s engineering explanation emphasizes its ability to process more candidate ads and model richer relationships between people and creative. That supports diversity—but does not guarantee that every additional asset improves performance.
Read the underlying source in Meta’s Andromeda engineering article.
Your job is to supply commercially different, evidence-backed ideas and a reliable conversion signal.
Need a systematic review of your angles, funnel, and qualified results? See my Facebook Ads service or request a Revenue Leak Audit.
Related:
Written by
Vince Servidad
PPC Strategist · Google Ads, Meta Ads & conversion systems
Filipino PPC strategist. A seven-figure Shopify brand and 10+ years across Google Ads, Meta Ads, stores, tracking, and content.
Need help with Facebook Ads?
Get strategic and hands-on support from a PPC strategist based in the Philippines.