Why Your Meta Ads Stopped Working -- And How We Fixed It (Andromeda Update Explained)

Why Your Meta Ads Stopped Working -- And How We Fixed It (Andromeda Update Explained)

Author

Patrick Driscoll, Co-Founder & CEO

Published Date

April 9, 2026

In September 2025, we scaled one of our e-commerce brand partners 70% month over month. The brand had been struggling with Meta ads performance for months. Their creative was not bad. Their offer was not bad. What was crushing them was Meta's Andromeda update, and they did not know it.

If your Meta ad performance felt like it fell off a cliff sometime in summer 2025 -- dropping CTRs, spiking CPMs, spend spreading thin with no clear reason -- this article is specifically for you. This is not a surface-level breakdown telling you to "add more creative diversity." This is the exact framework TVG built and deployed to help brands adapt and grow profitably through the Andromeda era.


Watch the Full Andromeda Breakdown on YouTube

Patrick walks through the complete creative diversity framework, campaign architecture, and a live client account showing 71% month-over-month growth post-Andromeda.

What Is the Meta Andromeda Update?

Meta announced the Andromeda update in a technical paper published in December 2024. The full rollout and its measurable impact on ad accounts began around July 2025. Even teams spending hundreds of thousands of dollars per month on Meta saw account volatility during this window.

Andromeda is an AI-driven retrieval system built into Meta's ad delivery engine. To understand why it changes everything, you need to understand one thing about the problem Meta was trying to solve.

With the explosion of AI-generated content, the volume of ads being uploaded to Meta every day increased dramatically. Meta needed a faster, smarter way to sort through millions of ads in milliseconds, identify which ones are relevant to which users, and distribute them efficiently. Andromeda is how they built that capability, using Nvidia's Grace Hopper superchips for real-time matching at unprecedented scale.

How Meta's Ad Delivery Works (Simplified)

Ad delivery has four stages: Retrieval, Light Ranking, Final Scoring, and Auction. Andromeda specifically changed the Retrieval stage -- the very first filter that determines which ads even make it into consideration for a given user. If your ads are not retrieved, they never reach the auction. This is where the impact is felt.

What Changed at the Retrieval Level

The old system was manual indexing. One creative tended to absorb most of the spend in an ad set. Small variations on the same underlying creative actually mattered because the system treated them as distinct.

Andromeda uses AI to cluster similar ads. If you take one UGC video and create five versions with different opening hooks, the system reads all five as the same concept. Budget gets distributed across concepts, not individual ads. The practical effect: if you have five nearly identical ads in an ad set, they are treated as one, and the spend that used to spread across them now pools toward a single representative of that cluster.

Variation Testing Is Dead

Running five versions of the same ad with different hooks, different thumbnail swaps, or minor copy changes used to be a reliable way to find incremental winners. Post-Andromeda, Meta sees those as identical. You are not getting five data points. You are getting one. This is the single biggest reason brands are seeing less spend distribution and flatter performance.

Creative Is Now Targeting

Andromeda scans every element of your ad creative: pixel-level visual content, skin tones, backgrounds, text overlay, copy, and even your landing page. It builds a data fingerprint for each ad and uses that fingerprint to determine who sees it.

This means your creative itself is now doing the audience targeting. The person featured in the ad, the problem being addressed, the format, the emotional tone -- all of it signals to the algorithm which user to show the ad to. Broad targeting and Advantage+ campaigns work because the creative carries the targeting information. The implications for how you build your creative strategy are significant.

CLIENT RESULT

71% month-over-month revenue growth post-Andromeda restructure

This brand came to TVG in July 2025 with performance that had fallen off. Most of their spend was concentrated in carousel ads that had previously performed well. When we audited the account, the core problem was a lack of creative diversity -- the algorithm had almost nothing new to learn from. We restructured the account, simplified the campaign architecture, and injected diverse new concepts across multiple formats. Revenue scaled from $180,000 to $263,000 in a single month.

The TVG Creative Diversity Framework

The response to Andromeda is not to produce more ads. It is to produce more conceptually distinct ads. Each unique concept creates a new data fingerprint that the algorithm can learn from, test, and distribute to a new audience segment. The brands winning post-Andromeda are not the ones with the highest volume -- they are the ones testing the most fundamentally different ideas.

TVG's internal framework for building creative diversity is built around four pillars. Every ad concept is defined by its position on each of these four axes:

FORMAT

Static image: Single visual asset, often graphic or photography-based

Short-form video: 15 to 60 second video, native or UGC style

Long-form video: 60+ seconds, educational, founder, or documentary format

Carousel: Multi-image sequential format

UGC: Creator-filmed, unscripted or lightly scripted authentic content

Demo / testimonial: Product demonstration or direct customer testimonial format

ANGLE

Problem-focused: Opens with the pain point the customer is experiencing

Desire-focused: Leads with the outcome or transformation the customer wants

Proof-focused: Social proof, reviews, results, before and after

Us vs. them: Direct comparison with alternatives or status quo

PERSONA

New buyer: First-time customer, unaware of your brand, cold audience

Gift buyer: Purchasing for someone else -- different emotional drivers

Skeptic: Has considered your category but not converted -- needs proof

Returning customer: Already bought -- cross-sell or replenishment angle

AWARENESS

Unaware: Does not yet know they have the problem you solve

Problem aware: Knows the problem, does not know your solution

Solution aware: Knows solutions exist, has not chosen yours

Product aware: Familiar with your brand, has not purchased

Most aware: Ready to buy -- needs a reason to act now

By mixing these four pillars, every combination produces a conceptually distinct ad with a unique data fingerprint. A problem-focused static ad targeting a skeptic at the solution-aware stage is a fundamentally different signal to the algorithm than a desire-focused UGC video targeting a cold new buyer. Both need to exist in your account. TVG builds 20 or more unique concepts per month using this framework.

Use AI to Scale the Ideation

Load all four pillars -- format, angle, persona, and awareness -- into ChatGPT or Claude with your product context, and prompt it to generate ad concepts across every combination. You can generate a month's worth of briefs in a single session. The framework provides the structure; AI accelerates the ideation.

Campaign Architecture That Wins Post-Andromeda

Complex campaign structures with dozens of ad sets, granular audience targeting, and overlapping campaigns worked in a different era. Post-Andromeda, the algorithm rewards simplicity and clean data. Here is the architecture that works:

Campaign Type

Purpose

What It Does

Advantage+ Sales Campaign (ASC)

Primary scaling engine

Runs your proven winners at scale with broad targeting. Let the algorithm find the audience using creative signals.

Creative Testing Campaign

Weekly concept validation

Tests new concepts against your current benchmarks. Winners feed directly into the ASC.

Retargeting Campaign

Warm audience capture

Most relevant during high-intent periods like Q4 or product launches. Nurtures and closes customers who have engaged but not converted.

Interest Stack (optional)

Data diversity

Used selectively to introduce fresh audience data signals. Not a primary driver -- use sparingly.

The principle is simplicity over complexity. You do not need 15 campaigns. You need one clean scaling setup powered by a weekly creative testing engine that feeds winners in and retires losers. That is the entire structure.

Audience Segmentation: Teaching the Algorithm Who to Target

Even with broad targeting, your audience segmentation setup directly impacts how well Andromeda learns. Keep your new customer and existing customer lists updated in Meta Ads Manager. If you are using Klaviyo, connect it to Meta and upload a dynamic list that updates automatically -- this ensures the algorithm always has a clean, current picture of your existing customer base.

This matters because your primary scaling goal should be net new customer acquisition. If the algorithm does not have a clearly defined existing customer segment to exclude from prospecting, it will mix audiences and dilute your acquisition efficiency.

Strengthen Your Signal: The Foundation Andromeda Requires

Every AI system is only as good as the data feeding it. Andromeda is no different. If your tracking is weak, the algorithm is operating on incomplete information, which means it cannot optimize efficiently regardless of how good your creative is.

Before anything else, audit and confirm all of the following:

  • Advanced Matching: Turned on in your Meta pixel settings. Matches more website events to Meta user profiles, improving algorithm optimization.

  • Conversions API (CAPI): Installed via gateway or direct integration. Server-side tracking that fills the gaps left by iOS privacy changes. Non-negotiable for accurate data.

  • Event Match Quality Score: Should be 8 out of 10 or higher. Check this in your Meta Events Manager. Low match quality means the algorithm is guessing.

  • Pixel event verification: Confirm all key events (ViewContent, AddToCart, InitiateCheckout, Purchase) are firing correctly and without duplication.

Signal Quality Is a Competitive Advantage

Most brands have degraded tracking from iOS updates and neglected CAPI setup. A brand with clean, high-quality signal data gives Andromeda more to work with than a competitor spending twice as much with poor tracking. Fix your signal infrastructure before scaling spend.

The Weekly Creative Flywheel

Andromeda rewards fresh data. Static ad accounts with the same creatives running for months are exactly what the algorithm penalizes. The brands winning post-Andromeda have built a consistent weekly creative production and testing rhythm. Here is the cadence:

Week Step

Action

Output

Ideate

Apply the four-pillar framework to generate new concept briefs

20+ distinct concepts mapped to format, angle, persona, and awareness

Produce

Brief out to UGC creators, graphic designers, or video team

Net new creative assets with clear concept documentation

Launch

Upload new concepts to the creative testing campaign

New data signals entering the algorithm's learning cycle

Scale Winners

Promote best performers to Advantage+ Sales Campaign

Proven concepts scaling to broader audiences

Archive Learnings

Document what worked and why -- angle, format, persona, awareness level

Institutional knowledge that improves future briefing

This is not about volume for its own sake. Twenty genuinely unique, well-briefed concepts will outperform 100 variations of the same idea every time under Andromeda. Quality and conceptual diversity are the variables that matter.

Your 4-Week Post-Andromeda Action Plan

If your Meta performance has degraded, here is a structured week-by-week plan to diagnose, restructure, and rebuild performance:

Week 1 -- Audit and Simplify

  1. Audit signal quality: check Advanced Matching, CAPI installation, and Event Match Quality Score

  2. Simplify campaign structure: consolidate into ASC (scaling) + Creative Testing + Retargeting

  3. Audit existing creative: identify which concepts are actually distinct vs. which are variations of the same idea

  4. Upload new customer and existing customer lists to Meta; connect Klaviyo dynamic list if applicable


Week 2 -- New Concepts Into Testing

  1. Use the four-pillar framework to brief 12 to 16 new concepts across diverse formats, angles, personas, and awareness levels

  2. Prioritize concepts that are fundamentally different from what is currently running -- not incremental variations

  3. Launch all new concepts into the Creative Testing campaign with controlled daily budgets

  4. Identify the 3 to 5 strongest performers from existing creative and consolidate them into the ASC


Week 3 -- Analyze Signals and Scale Winners

  1. Review creative testing results: look at CTR, hook rate, and cost per acquisition by concept -- not just ROAS

  2. . Identify which format, angle, and persona combinations are producing the strongest signals

  3. . Promote top performers from the testing campaign into the ASC

  4. . Brief the next batch of concepts specifically in the formats and angles that are showing traction


Week 4 -- Establish the Rhythm

  1. . Lock in weekly creative briefing, production, and launch as a recurring process -- not a one-time fix

  2. . Document what is working in a centralized creative log: concept, format, angle, persona, performance, and notes

  3. . Monitor spend distribution across concepts: if one ad is hogging all spend, review for clustering and add more diverse concepts

  4. . Review MER and true CAC week over week to confirm system efficiency is improving alongside creative diversity

What to Measure Post-Andromeda

Stop measuring success by ROAS alone. The metrics that actually tell you whether your account is healthy in the Andromeda era:

Metric

What It Tells You

Target

Concepts Tested Per Week

Whether the creative flywheel is actually running

3 to 5 new concepts minimum; scale with budget

Spend Distribution

Whether budget is spreading across diverse concepts or hoarding to one

No single ad should dominate 60%+ of spend consistently

CPM Trend

Whether signal quality and creative relevance is improving

Declining or stable CPMs signal healthy creative diversity

Conversion Rate by Format

Which creative formats Meta is favoring for your audience

Track separately for UGC, static, video, carousel

True New Customer CAC

Whether acquisition efficiency is improving

Ad spend / new customers -- not Meta's reported CPA

Marketing Efficiency Ratio

Whether the whole system is profitable as you scale

Total revenue / total marketing spend; target 3x+

Common Post-Andromeda Mistakes to Avoid

The most frequent errors we see when brands try to adapt to Andromeda on their own:

Creating 10 Ads That Look Identical

If your creative team is producing minor variations -- same video, different hook text -- you are not creating diverse concepts. You are creating one concept in Andromeda's view. Every new ad should have a meaningfully different format, angle, or persona to generate a new data fingerprint.


Over-Segmented Campaign Structure

Multiple campaigns with overlapping audiences, granular interest targeting, and separate ad sets for every audience slice creates signal fragmentation. The algorithm needs consolidated spend to optimize efficiently. Fewer campaigns with clean structure outperforms complex setups in the current environment.


Weak Tracking Signals

Running Andromeda without Advanced Matching and a properly installed CAPI is like asking someone to navigate with a broken compass. The algorithm cannot find the right audience if the event data feeding it is incomplete or mismatched. Fix the foundation before spending on creative.


Mass Market, Generic Messaging

Broad, non-specific creative does not give Andromeda enough information to find a relevant audience. The more specifically your creative speaks to a defined persona, problem, and awareness level, the more precisely the algorithm can target. Go narrow and specific, not broad and generic.


Watch the Full Andromeda Breakdown on YouTube

Patrick walks through the complete creative diversity framework, campaign architecture, and a live client account showing 71% month-over-month growth post-Andromeda.

Watch: Why Your Meta Ads Stopped Working And How We Fixed It (Andromeda Update Explained)


Get a Free Meta Ad Account Audit

Our team will audit your ad account structure, creative diversity, audience segmentation, and signal health against the Andromeda framework -- and hand you a personalized roadmap to fix what is holding performance back.

Book Your Free Core Growth Audit at thevisionarygrouptx.com

About TVG

The Visionary Group (TVG) is a full-service e-commerce growth agency helping 7 and 8-figure Shopify brands scale profitably through paid media, creative strategy, email, and analytics. TVG spends and manages millions in Meta ad spend monthly across active brand partners.

Our mission is to

Scale your brand with proven growth strategies.

Weekly breakdowns of what’s actually working in paid ads, email, and creative.

No spam, unsubscribe anytime.

OUR LATEST ON YOUTUBE

Free ad breakdowns, performance insights, and proven strategies you can implement to scale your brand.

OUR LATEST ON YOUTUBE

Free ad breakdowns, performance insights, and proven strategies you can implement to scale your brand.

We're an extension of your team.

Think of us as an extension of your team. If you succeed, so do we. Our digital marketing team works night and day to ensure you get the results you want. Everything is carefully planned out and strategized to make sure your brand scales profitably.

©2026 The visionary group. All rights reserved.

Home

Blogs

Why Your Meta Ads Stopped Working -- And How We Fixed It (Andromeda Update Explained)

Author

Patrick Driscoll, Co-Founder & CEO

Article Category

Meta

Published Date

April 9, 2026

Why Your Meta Ads Stopped Working -- And How We Fixed It (Andromeda Update Explained)

In September 2025, we scaled one of our e-commerce brand partners 70% month over month. The brand had been struggling with Meta ads performance for months. Their creative was not bad. Their offer was not bad. What was crushing them was Meta's Andromeda update, and they did not know it.

If your Meta ad performance felt like it fell off a cliff sometime in summer 2025 -- dropping CTRs, spiking CPMs, spend spreading thin with no clear reason -- this article is specifically for you. This is not a surface-level breakdown telling you to "add more creative diversity." This is the exact framework TVG built and deployed to help brands adapt and grow profitably through the Andromeda era.


Watch the Full Andromeda Breakdown on YouTube

Patrick walks through the complete creative diversity framework, campaign architecture, and a live client account showing 71% month-over-month growth post-Andromeda.

What Is the Meta Andromeda Update?

Meta announced the Andromeda update in a technical paper published in December 2024. The full rollout and its measurable impact on ad accounts began around July 2025. Even teams spending hundreds of thousands of dollars per month on Meta saw account volatility during this window.

Andromeda is an AI-driven retrieval system built into Meta's ad delivery engine. To understand why it changes everything, you need to understand one thing about the problem Meta was trying to solve.

With the explosion of AI-generated content, the volume of ads being uploaded to Meta every day increased dramatically. Meta needed a faster, smarter way to sort through millions of ads in milliseconds, identify which ones are relevant to which users, and distribute them efficiently. Andromeda is how they built that capability, using Nvidia's Grace Hopper superchips for real-time matching at unprecedented scale.

How Meta's Ad Delivery Works (Simplified)

Ad delivery has four stages: Retrieval, Light Ranking, Final Scoring, and Auction. Andromeda specifically changed the Retrieval stage -- the very first filter that determines which ads even make it into consideration for a given user. If your ads are not retrieved, they never reach the auction. This is where the impact is felt.

What Changed at the Retrieval Level

The old system was manual indexing. One creative tended to absorb most of the spend in an ad set. Small variations on the same underlying creative actually mattered because the system treated them as distinct.

Andromeda uses AI to cluster similar ads. If you take one UGC video and create five versions with different opening hooks, the system reads all five as the same concept. Budget gets distributed across concepts, not individual ads. The practical effect: if you have five nearly identical ads in an ad set, they are treated as one, and the spend that used to spread across them now pools toward a single representative of that cluster.

Variation Testing Is Dead

Running five versions of the same ad with different hooks, different thumbnail swaps, or minor copy changes used to be a reliable way to find incremental winners. Post-Andromeda, Meta sees those as identical. You are not getting five data points. You are getting one. This is the single biggest reason brands are seeing less spend distribution and flatter performance.

Creative Is Now Targeting

Andromeda scans every element of your ad creative: pixel-level visual content, skin tones, backgrounds, text overlay, copy, and even your landing page. It builds a data fingerprint for each ad and uses that fingerprint to determine who sees it.

This means your creative itself is now doing the audience targeting. The person featured in the ad, the problem being addressed, the format, the emotional tone -- all of it signals to the algorithm which user to show the ad to. Broad targeting and Advantage+ campaigns work because the creative carries the targeting information. The implications for how you build your creative strategy are significant.

CLIENT RESULT

71% month-over-month revenue growth post-Andromeda restructure

This brand came to TVG in July 2025 with performance that had fallen off. Most of their spend was concentrated in carousel ads that had previously performed well. When we audited the account, the core problem was a lack of creative diversity -- the algorithm had almost nothing new to learn from. We restructured the account, simplified the campaign architecture, and injected diverse new concepts across multiple formats. Revenue scaled from $180,000 to $263,000 in a single month.

The TVG Creative Diversity Framework

The response to Andromeda is not to produce more ads. It is to produce more conceptually distinct ads. Each unique concept creates a new data fingerprint that the algorithm can learn from, test, and distribute to a new audience segment. The brands winning post-Andromeda are not the ones with the highest volume -- they are the ones testing the most fundamentally different ideas.

TVG's internal framework for building creative diversity is built around four pillars. Every ad concept is defined by its position on each of these four axes:

FORMAT

Static image: Single visual asset, often graphic or photography-based

Short-form video: 15 to 60 second video, native or UGC style

Long-form video: 60+ seconds, educational, founder, or documentary format

Carousel: Multi-image sequential format

UGC: Creator-filmed, unscripted or lightly scripted authentic content

Demo / testimonial: Product demonstration or direct customer testimonial format

ANGLE

Problem-focused: Opens with the pain point the customer is experiencing

Desire-focused: Leads with the outcome or transformation the customer wants

Proof-focused: Social proof, reviews, results, before and after

Us vs. them: Direct comparison with alternatives or status quo

PERSONA

New buyer: First-time customer, unaware of your brand, cold audience

Gift buyer: Purchasing for someone else -- different emotional drivers

Skeptic: Has considered your category but not converted -- needs proof

Returning customer: Already bought -- cross-sell or replenishment angle

AWARENESS

Unaware: Does not yet know they have the problem you solve

Problem aware: Knows the problem, does not know your solution

Solution aware: Knows solutions exist, has not chosen yours

Product aware: Familiar with your brand, has not purchased

Most aware: Ready to buy -- needs a reason to act now

By mixing these four pillars, every combination produces a conceptually distinct ad with a unique data fingerprint. A problem-focused static ad targeting a skeptic at the solution-aware stage is a fundamentally different signal to the algorithm than a desire-focused UGC video targeting a cold new buyer. Both need to exist in your account. TVG builds 20 or more unique concepts per month using this framework.

Use AI to Scale the Ideation

Load all four pillars -- format, angle, persona, and awareness -- into ChatGPT or Claude with your product context, and prompt it to generate ad concepts across every combination. You can generate a month's worth of briefs in a single session. The framework provides the structure; AI accelerates the ideation.

Campaign Architecture That Wins Post-Andromeda

Complex campaign structures with dozens of ad sets, granular audience targeting, and overlapping campaigns worked in a different era. Post-Andromeda, the algorithm rewards simplicity and clean data. Here is the architecture that works:

Campaign Type

Purpose

What It Does

Advantage+ Sales Campaign (ASC)

Primary scaling engine

Runs your proven winners at scale with broad targeting. Let the algorithm find the audience using creative signals.

Creative Testing Campaign

Weekly concept validation

Tests new concepts against your current benchmarks. Winners feed directly into the ASC.

Retargeting Campaign

Warm audience capture

Most relevant during high-intent periods like Q4 or product launches. Nurtures and closes customers who have engaged but not converted.

Interest Stack (optional)

Data diversity

Used selectively to introduce fresh audience data signals. Not a primary driver -- use sparingly.

The principle is simplicity over complexity. You do not need 15 campaigns. You need one clean scaling setup powered by a weekly creative testing engine that feeds winners in and retires losers. That is the entire structure.

Audience Segmentation: Teaching the Algorithm Who to Target

Even with broad targeting, your audience segmentation setup directly impacts how well Andromeda learns. Keep your new customer and existing customer lists updated in Meta Ads Manager. If you are using Klaviyo, connect it to Meta and upload a dynamic list that updates automatically -- this ensures the algorithm always has a clean, current picture of your existing customer base.

This matters because your primary scaling goal should be net new customer acquisition. If the algorithm does not have a clearly defined existing customer segment to exclude from prospecting, it will mix audiences and dilute your acquisition efficiency.

Strengthen Your Signal: The Foundation Andromeda Requires

Every AI system is only as good as the data feeding it. Andromeda is no different. If your tracking is weak, the algorithm is operating on incomplete information, which means it cannot optimize efficiently regardless of how good your creative is.

Before anything else, audit and confirm all of the following:

  • Advanced Matching: Turned on in your Meta pixel settings. Matches more website events to Meta user profiles, improving algorithm optimization.

  • Conversions API (CAPI): Installed via gateway or direct integration. Server-side tracking that fills the gaps left by iOS privacy changes. Non-negotiable for accurate data.

  • Event Match Quality Score: Should be 8 out of 10 or higher. Check this in your Meta Events Manager. Low match quality means the algorithm is guessing.

  • Pixel event verification: Confirm all key events (ViewContent, AddToCart, InitiateCheckout, Purchase) are firing correctly and without duplication.

Signal Quality Is a Competitive Advantage

Most brands have degraded tracking from iOS updates and neglected CAPI setup. A brand with clean, high-quality signal data gives Andromeda more to work with than a competitor spending twice as much with poor tracking. Fix your signal infrastructure before scaling spend.

The Weekly Creative Flywheel

Andromeda rewards fresh data. Static ad accounts with the same creatives running for months are exactly what the algorithm penalizes. The brands winning post-Andromeda have built a consistent weekly creative production and testing rhythm. Here is the cadence:

Week Step

Action

Output

Ideate

Apply the four-pillar framework to generate new concept briefs

20+ distinct concepts mapped to format, angle, persona, and awareness

Produce

Brief out to UGC creators, graphic designers, or video team

Net new creative assets with clear concept documentation

Launch

Upload new concepts to the creative testing campaign

New data signals entering the algorithm's learning cycle

Scale Winners

Promote best performers to Advantage+ Sales Campaign

Proven concepts scaling to broader audiences

Archive Learnings

Document what worked and why -- angle, format, persona, awareness level

Institutional knowledge that improves future briefing

This is not about volume for its own sake. Twenty genuinely unique, well-briefed concepts will outperform 100 variations of the same idea every time under Andromeda. Quality and conceptual diversity are the variables that matter.

Your 4-Week Post-Andromeda Action Plan

If your Meta performance has degraded, here is a structured week-by-week plan to diagnose, restructure, and rebuild performance:

Week 1 -- Audit and Simplify

  1. Audit signal quality: check Advanced Matching, CAPI installation, and Event Match Quality Score

  2. Simplify campaign structure: consolidate into ASC (scaling) + Creative Testing + Retargeting

  3. Audit existing creative: identify which concepts are actually distinct vs. which are variations of the same idea

  4. Upload new customer and existing customer lists to Meta; connect Klaviyo dynamic list if applicable


Week 2 -- New Concepts Into Testing

  1. Use the four-pillar framework to brief 12 to 16 new concepts across diverse formats, angles, personas, and awareness levels

  2. Prioritize concepts that are fundamentally different from what is currently running -- not incremental variations

  3. Launch all new concepts into the Creative Testing campaign with controlled daily budgets

  4. Identify the 3 to 5 strongest performers from existing creative and consolidate them into the ASC


Week 3 -- Analyze Signals and Scale Winners

  1. Review creative testing results: look at CTR, hook rate, and cost per acquisition by concept -- not just ROAS

  2. . Identify which format, angle, and persona combinations are producing the strongest signals

  3. . Promote top performers from the testing campaign into the ASC

  4. . Brief the next batch of concepts specifically in the formats and angles that are showing traction


Week 4 -- Establish the Rhythm

  1. . Lock in weekly creative briefing, production, and launch as a recurring process -- not a one-time fix

  2. . Document what is working in a centralized creative log: concept, format, angle, persona, performance, and notes

  3. . Monitor spend distribution across concepts: if one ad is hogging all spend, review for clustering and add more diverse concepts

  4. . Review MER and true CAC week over week to confirm system efficiency is improving alongside creative diversity

What to Measure Post-Andromeda

Stop measuring success by ROAS alone. The metrics that actually tell you whether your account is healthy in the Andromeda era:

Metric

What It Tells You

Target

Concepts Tested Per Week

Whether the creative flywheel is actually running

3 to 5 new concepts minimum; scale with budget

Spend Distribution

Whether budget is spreading across diverse concepts or hoarding to one

No single ad should dominate 60%+ of spend consistently

CPM Trend

Whether signal quality and creative relevance is improving

Declining or stable CPMs signal healthy creative diversity

Conversion Rate by Format

Which creative formats Meta is favoring for your audience

Track separately for UGC, static, video, carousel

True New Customer CAC

Whether acquisition efficiency is improving

Ad spend / new customers -- not Meta's reported CPA

Marketing Efficiency Ratio

Whether the whole system is profitable as you scale

Total revenue / total marketing spend; target 3x+

Common Post-Andromeda Mistakes to Avoid

The most frequent errors we see when brands try to adapt to Andromeda on their own:

Creating 10 Ads That Look Identical

If your creative team is producing minor variations -- same video, different hook text -- you are not creating diverse concepts. You are creating one concept in Andromeda's view. Every new ad should have a meaningfully different format, angle, or persona to generate a new data fingerprint.


Over-Segmented Campaign Structure

Multiple campaigns with overlapping audiences, granular interest targeting, and separate ad sets for every audience slice creates signal fragmentation. The algorithm needs consolidated spend to optimize efficiently. Fewer campaigns with clean structure outperforms complex setups in the current environment.


Weak Tracking Signals

Running Andromeda without Advanced Matching and a properly installed CAPI is like asking someone to navigate with a broken compass. The algorithm cannot find the right audience if the event data feeding it is incomplete or mismatched. Fix the foundation before spending on creative.


Mass Market, Generic Messaging

Broad, non-specific creative does not give Andromeda enough information to find a relevant audience. The more specifically your creative speaks to a defined persona, problem, and awareness level, the more precisely the algorithm can target. Go narrow and specific, not broad and generic.


Watch the Full Andromeda Breakdown on YouTube

Patrick walks through the complete creative diversity framework, campaign architecture, and a live client account showing 71% month-over-month growth post-Andromeda.

Watch: Why Your Meta Ads Stopped Working And How We Fixed It (Andromeda Update Explained)


Get a Free Meta Ad Account Audit

Our team will audit your ad account structure, creative diversity, audience segmentation, and signal health against the Andromeda framework -- and hand you a personalized roadmap to fix what is holding performance back.

Book Your Free Core Growth Audit at thevisionarygrouptx.com

About TVG

The Visionary Group (TVG) is a full-service e-commerce growth agency helping 7 and 8-figure Shopify brands scale profitably through paid media, creative strategy, email, and analytics. TVG spends and manages millions in Meta ad spend monthly across active brand partners.

Our mission is to

Help you create and earn on your terms.

No spam, unsubscribe anytime.

OUR LATEST ON YOUTUBE

Free ad breakdowns, performance insights, and proven strategies you can implement to scale your brand.

We're an extension of your team.

Think of us as an extension of your team. If you succeed, so do we. Our digital marketing team works night and day to ensure you get the results you want. Everything is carefully planned out and strategized to make sure your brand scales profitably.

©2026 The visionary group. All rights reserved.

We're an extension of your team.

Think of us as an extension of your team. If you succeed, so do we. Our digital marketing team works night and day to ensure you get the results you want. Everything is carefully planned out and strategized to make sure your brand scales profitably.

©2026 The visionary group. All rights reserved.

We're an extension of your team.

Think of us as an extension of your team. If you succeed, so do we. Our digital marketing team works night and day to ensure you get the results you want. Everything is carefully planned out and strategized to make sure your brand scales profitably.

©2026 The visionary group. All rights reserved.