Adapting Meta Ads Campaigns to Andromeda in 2026: Strategies for E-commerce Brands

xplr. - agence ads · créa · e-commerce

Introduction [00:00]

• Meta's Andromeda is the most significant update to its advertising algorithm since iOS 14.
• This update has a major impact on e-commerce brands running ads on Meta platforms.
• The changes confirm existing trends: simplification of ad accounts, creativity as a new performance lever, and macro-indicator-based management (CAC, COS) over micro-indicators (ROAS).
• Media buying is evolving from a purely technical approach to a more strategic and creative one.
• Meta has heavily invested in Artificial Intelligence to transform how ads are delivered on Facebook and Instagram, aiming to display the right ad, with the right message, at the right time, to the right person, with hyper-personalization.

Why These Changes? [02:59]

• These changes are a response to three major structural mutations in the advertising ecosystem.
• 1. iOS 14 and Data Loss [03:10]:
  • Apple's iOS 14 update in 2021 allowed users to refuse ad tracking.
  • Meta lost a significant portion of data, leading to a drop in ad performance and revenue.
  • This forced Meta to rebuild its targeting precision by relying primarily on its own data, moving away from dependence on external data.
• 2. Emergence of New Competitors [03:40]:
  • The online advertising market has evolved with the rise of platforms like TikTok, Amazon, and Instacart.
  • These competitors pushed Meta to innovate and not rest on its laurels.
  • Platforms are now compared based on efficiency and ease of use. Meta's challenge is to remain the preferred platform for advertisers.
• 3. Explosion in Creative Volume [04:16]:
  • Advertisers are producing an increasing volume of visual content.
    • Creative content has become the main lever for optimization.
    • Generative AI tools (Midjourney, Sora) multiply production capabilities.
  • The algorithm must evolve to handle a much larger volume of ads while maintaining minimal latency and maximum relevance for the end-user.
  • Andromeda is Meta's technological response to these challenges.

Andromeda [05:21]

• Andromeda is a complete overhaul of Meta's ad delivery engine.
• Goal: To analyze a massive volume of ads in record time to display the most relevant ad to each user.
• Mechanism: It's a "retrieval system" that acts like a funnel, selecting the most pertinent ad in a fraction of a second.
• Hyper-personalization: It goes beyond instant signals (clicks, views, add-to-carts) to evaluate behavioral sequences over time (videos watched, influencers followed, content themes).
• Powered by powerful Nvidia chips, capable of processing 10,000 times more information than previous versions.

GEM (Generative Ads Recommendation Model) [06:38]

• A predictive model designed to anticipate user purchase intentions even before they are explicitly expressed.
• It's trained on billions of signals to predict what a user will want to buy next.
• Example: If a user starts engaging with content about home renovations, GEM might anticipate a future move and start showing ads for sofas or wallpaper.

Lattice [07:23]

• A model that learns simultaneously from all campaigns, regardless of their objective (traffic, conversion, notoriety), platform (Facebook or Instagram), or placement (Reels, Feed, Stories).
• Previously, campaigns, platforms, and placements were optimized independently.
• Now, every advertising experience enriches the next.
• Example: Engagement with a Reel on Instagram can improve the delivery of a static ad in the Facebook feed. Learnings from an awareness campaign can enhance a conversion campaign on the same account.

Sequence Learning [08:24]

• Allows AI to "think" in logical steps and focus on the next natural step in the customer journey.
• Example: After a user buys a baby sleeping bag, the AI will anticipate needs for winter boots and then spring clothes in appropriate sizes over the coming weeks or months.
• It sequences the progression, pushing complementary products and anticipating changes in needs (e.g., size changes, seasonal items).

In Summary (Updates) [08:56]

• Andromeda: Sorts and selects the best ad in real-time (Diffusion and Relevance).
• GEM: Anticipates purchase intentions (Prediction).
• Lattice: Unifies learning across all campaigns and placements (Coherence).
• Sequence Learning: Anticipates the next step in the customer journey (Purchase Path).
• Overall Impact: Meta can better target who, when, and how. Ad delivery quality improves. The path to automation continues. The focus of media buying shifts from technical audience building to strategy and creative content.

Zone 1: The Technical [10:53]

• Tracking and Data Collection [11:16]:
  • Data remains crucial for Meta's algorithm.
  • Ensure optimal tracking with Meta Pixel and Conversion API.
  • More numerous and precise signals lead to higher quality ad diffusion.
• Account Structure: Simplicity and Learning [11:44]:
  • A complex, scattered structure hinders learning and dilutes budgets.
  • A single campaign with a €500/day budget learns 5x faster than five campaigns with €100/day each.
  • Simple structures are essential for proper budget allocation and achieving objectives.
• Priority to Acquisition [12:10]:
  • Meta Ads should primarily serve acquisition, bringing new, qualified traffic to your site.
  • Meta is the most effective platform for acquisition, presenting products/services to millions of qualified non-customers daily.
  • For e-commerce, continuous new, qualified traffic is essential for sustainable growth.
• Simplified Account Structure Example [12:47]:
  • 1. Acquisition main campaign (70-80% budget):
    • Objective: Generate new, qualified traffic for continuous customer acquisition.
    • Setup: Advantage+ campaign with Purchase objective.
    • Audience: Broad, no specific targeting.
    • Exclusions: Site visitors, CRM clients, engaged users (to avoid retargeting existing audience).
  • 2. Acquisition testing (10-20% budget):
    • Objective: Test new messages, hooks, concepts, or personas to identify winning ads for the main acquisition campaign.
    • Setup: ABO campaign (Adset Budget Optimization).
    • Organization: 1 ad set = 1 test (one angle, one persona, one hook).
    • Audience: Broad, no specific targeting.
    • Budgeting: Invest at least 3x your average CPA before drawing conclusions.
  • 3. Retargeting (10% budget):
    • Objective: Convert warm audiences (visitors, leads, engagers) who haven't purchased yet.
    • These campaigns are less central but remain indispensable.
    • Audience: 180-day website visitors, Meta engagers, video viewers.
    • Exclusions: Purchasers (Pixel + CRM).
• Adsets to maintain control [15:59]:
  • Adsets no longer segment audiences but organize your ads.
  • Fashion brands: Adsets for best-sellers, new collections, discounted products.
  • Cosmetic brands: Adsets for sensitive skin, blemish-prone skin, current promotions.
  • The goal is to shift targeting from audience segmentation to creative content.
• Budget Management [16:37]:
  • Prioritize CBO (Campaign Budget Optimization) where the budget is managed at the campaign level.
  • Use ABO (Adset Budget Optimization) if more control is needed.
  • A new option allows Meta to automatically adjust budget towards best-performing adsets (up to 20% of original ad set budget).
• Other Objectives [16:55]:
  • While sales campaigns are prioritized, Meta's enhanced AI can create synergies between different campaign types.
  • Consider allocating a small portion of the budget to awareness (video views) or lead generation campaigns if the algorithm struggles to find new audiences or if you hit a performance ceiling.

Zone 2: The Creative [17:37]

• Creative is Your New Targeting [17:53]:
  • Every interaction with an ad (view, click, engagement) becomes an audience signal.
  • Meta analyzes these signals to identify the profile of users most likely to purchase.
  • A good creative attracts the right people, their attention generates good data, and this data fuels the algorithm to target similar profiles.
  • Your creative content and hook are your new targeting tools.
  • Hooks for Filtering: Start videos with precise, filtering questions to engage only relevant audiences.
    • Example (Happy Kids): "Is your child struggling to stop the pacifier?" If not, users scroll past, preventing bad data collection. If yes, engagement provides valuable signals.
    • This allows Meta to understand who the target persona is, analyze their behavior, and refine ad delivery.
• Diversify Content to Broaden Targeting [20:31]:
  • No single content type (UGC, FGC, unboxing, static, AI-generated) is a miracle solution.
  • The key is Creative Diversity: vary formats, messages, angles, and concepts to appeal to different audiences.
  • Diversifying content allows you to address various user segments with different needs and consumption habits on Meta platforms.
• Audience Segmentation for Content Strategy [22:12]:
  • 1. Bottom of Funnel (BOF):
    • Objective: Convert purchase intent into actual sales.
    • Audience: Currently in need of your products or considering competitors.
    • Recommended Content: Product-centric ads with photos/videos, benefits, technical details. Promotional codes, free shipping offers, exclusive bundles. Seasonal offers (e.g., Black Friday). Comparative arguments (product vs. alternatives). Urgency/scarcity (e.g., "last chance").
  • 2. Middle of Funnel (MOF):
    • Objective: Reassure and guide prospects; show social proof.
    • Audience: Have a latent need for your products, but haven't expressed it.
    • Recommended Content: Client testimonials (before/after). Authentic client reviews. Micro-influencer collaborations. Press mentions, TV features, co-branding, whitelisting.
  • 3. Top of Funnel (TOF):
    • Objective: Introduce the brand, explain its mission, purpose, and values; attract new audiences.
    • Audience: Cold audience, not actively seeking your product, but within your broader target.
    • Recommended Content: Founder/team interviews (vision, mission, DNA). Product demonstrations in a usage context. "How to choose...", "3 indispensable for..." advice videos. Behind-the-scenes, factory tours, company values, lifestyle content.
• Evaluating Creative Diversification [24:28]:
  • Create a table with three columns (TOF, MOF, BOF).
  • Classify all your ads from the past few months.
  • Often, there's an imbalance with many BOF creatives but few TOF/MOF.
  • Mission: Fill the other columns by producing creatives adapted to each stage of the customer journey.
• True Diversity, Not Just Variations [24:55]:
  • Diversifying creatives means creating genuinely distinct concepts, not just minor variations (e.g., changing CTA color).
  • Meta now distinguishes true diversity from simple variations and penalizes accounts that duplicate overly similar ads.
• New Creative Indicators [25:10]:
  • Creative Fatigue: Measures global ad weariness, even between visually similar ads across all variations. Helps determine when to inject truly new creatives.
  • Creative Similarity: Groups highly similar ads (same concept, visual, or structure). Helps avoid saturating audiences with almost identical messages.
  • Top Creative Themes: Categorizes ads by dominant theme (comparative, product demo, testimonial). Helps identify the most effective concepts to replicate.

Zone 3: The Piloting [26:31]

• ROAS: An Indicator to Relativize [26:36]:
  • ROAS is still useful but should not be your sole primary compass.
  • Focusing solely on ROAS can skew your analysis, especially when comparing different funnel stages (e.g., a high-ROAS BOF promo ad vs. a lower-ROAS TOF awareness ad).
  • ROAS is platform-specific and short-term biased, making it unreliable for overall strategic guidance.
• CAC: Your North Star [27:18]:
  • CAC (Customer Acquisition Cost) is the true indicator of profitability.
  • It measures your ability to acquire new customers profitably.
  • Formula: CAC = Total Marketing Budget / Number of New Customers during the period.
  • Marketing Budget: Includes media spend, creative costs, agency fees, and other relevant costs.
  • A good CAC is calculated off-platform using real data (Shopify, CRM, accounting tools), independent of tracking systems and attribution models.
  • CAC provides a medium-to-long-term vision and global company data, making it the essential "guardian of truth" for your growth.
• The New Role of the Media Buyer [28:07]:
  • The role of performance managers and agencies is no longer limited to managing campaigns.
  • New focus:
    • Budget planning (monthly, weekly, daily, seasonal).
    • Setting coherent growth and profitability objectives.
    • Leveraging data for strategic decisions.
    • Ensuring every euro invested is used optimally.
  • Conversations with clients now revolve around:
    • How to drive healthy and profitable growth.
    • How to reach the right people with the right message at the right time.
  • Andromeda marks the end of the technical media buyer era. The future belongs to creative strategists and performance managers who can orchestrate data, storytelling, and campaign structure for sustainable, measurable growth.

Conclusion [28:27]

• Andromeda's updates demand a shift in Meta Ad strategies.
• Advertisers must prioritize data tracking, simplify account structures, adopt creative diversity, and focus on macro-level performance indicators like CAC.
• Success in 2026 and beyond will hinge on adaptability, embracing AI-driven tools, and moving towards more strategic and creative campaign management.