Quick Answer: Meta killed lookalike audiences in 2026. Andromeda's predictive ML now handles targeting automatically — your creative quality is the new targeting lever.

TL;DR: Meta fully phased out lookalike audiences in early 2026, replacing them with Andromeda-powered predictive models that match ads based on aggregated behavior signals — not audience settings you define. The Andromeda retrieval algorithm, running on NVIDIA GH200 chips, processes 10,000x more ad variants in parallel with 100x faster matching. For veteran dropshippers, this means one thing: creative is targeting now. Your ad images, hooks, and angles determine who sees your ads, not your saved audiences or seed lists. Advantage+ automation is the default for all new campaigns. Static images still drive 60-70% of conversions despite Meta pushing video. New opportunities include Threads ads (favorable early CPMs), WhatsApp Ads through Meta Ads Manager, and ad sequencing for awareness campaigns. Adapt or watch your ROAS crater.

The Targeting Model You Built Over Years? It's Gone.

If you've been running Meta Ads since the iOS 14 days, you remember building lookalike audiences from customer lists, pixel data, and purchase events. You probably had a 1% LAL from high-AOV buyers that consistently outperformed everything else.

That audience doesn't exist anymore.

Meta fully deprecated lookalike audiences in 2026. Not "deprecated but still working in the background." Gone. The entire targeting paradigm shifted from advertiser-defined audiences to algorithm-defined distribution.

---

Andromeda: The Engine Behind Everything

Meta's Andromeda retrieval algorithm is now fully deployed across all ad placements. The technical specs matter because they explain why your old playbook stopped working:

Andromeda doesn't care about your 1% lookalike seed. It processes behavioral signals across Meta's entire user base and decides — in real time — which users are most likely to convert on your specific creative.

The Before vs After

Generative Ad Model (GEM)

Meta is rolling out GEM — their Generative Ad Model. The pitch: provide a product URL, set a budget, write a prompt, and the algorithm generates and tests creative automatically.

For dropshippers running 50+ products, this could slash creative production costs. But early results are mixed. GEM-generated creative tends to be generic. It works for broad awareness but underperforms hand-crafted creative for direct response.

Our take: Use GEM for initial testing of new products. Switch to custom creative once you identify winners.

---

1. Creative Is Your Only Targeting Lever

This isn't a slogan — it's literally how the system works now. Andromeda reads your creative (image, copy, headline, CTA) and uses those signals to determine distribution. A fitness product ad with a gym background will reach gym-goers. A kitchen gadget ad shot on a marble countertop will reach home decor enthusiasts.

The creative itself defines the audience.

Practical implications for dropshippers:

• Angle = Audience. The same product photographed for "busy moms" vs "college students" will reach completely different users.
• First 3 seconds of video (or the image itself) carry disproportionate weight in distribution decisions.
• UGC-style creative still outperforms polished brand content for direct response — the algorithm reads authenticity signals.

2. Static Images Still Work

Despite Meta's aggressive push toward video (including a new image-to-video tool that converts up to 20 product photos into video ads), static images still drive 60-70% of conversions across e-commerce.

Don't let Meta's product announcements pressure you into abandoning what works. The image-to-video tool is worth testing, but static product shots with clear value propositions remain the highest-converting format for most dropshipping products.

3. Advantage+ Is Non-Negotiable

Advantage+ Shopping campaigns are the default for all new campaigns. Fighting this is wasted energy.

What Advantage+ actually does well:

• Consolidates learning across placements
• Finds pockets of demand you'd never target manually
• Reduces the "audience fragmentation" problem from running 20 ad sets

What it does poorly:

• Spend distribution is opaque
• You lose granular placement control
• Creative fatigue happens faster when the algorithm has fewer constraints

Tactical recommendation: Run one Advantage+ campaign as your main driver. Keep one manual campaign for testing specific angles before feeding winners into Advantage+.

4. Retargeting Requires Disclosure Now

Meta now requires all advertisers to declare the source of data used in remarketing campaigns. This includes pixel data, customer lists, and app activity.

For dropshippers, this means:

• Your privacy policy needs to explicitly mention Meta pixel usage
• Customer list uploads require documented consent
• Data source declarations are visible to users in ad transparency

Non-compliance risks ad account restrictions. Update your store's privacy policy and data collection disclosures.

---

Threads Ads: Early Mover Window

Threads ads are rolling out globally, and early CPMs are favorable compared to Instagram and Facebook feed placements. This is the typical pattern — new placement launches with low competition, CPMs spike 6-12 months later as advertisers pile in.

Early data suggests:

• CPMs 30-50% lower than equivalent Instagram feed placements
• Engagement rates higher (novelty effect)
• Best for top-of-funnel awareness and brand building
• Text-forward creative performs well (matches platform culture)

If you have the budget to test, allocate 5-10% to Threads placements now. Document your benchmarks before the window closes.

WhatsApp Ads via Meta Ads Manager

WhatsApp ad management is now centralized in Meta Ads Manager. For dropshippers selling to markets where WhatsApp is the primary messaging platform (Latin America, parts of Europe, Middle East), this opens a direct click-to-WhatsApp ad flow without third-party tools.

Ad Sequencing for Awareness Campaigns

Meta now supports ad sequencing for Awareness and Engagement campaign objectives. You can define the order users see your ads — useful for product launches or seasonal pushes where storytelling matters.

Example sequence for a product launch:

1. Problem-aware creative (day 1-3)
2. Solution intro (day 4-6)
3. Social proof / testimonial (day 7-9)
4. Direct offer (day 10+)

This doesn't replace direct response campaigns, but it's a useful addition for building demand around new products.

---

Webhooks mTLS Certificate Update (Deadline: March 31, 2026)

If you use Meta webhooks for order tracking, conversion API, or any server-to-server integration, Meta is switching mTLS certificates to its own Meta CA by March 31, 2026. Your existing certificates will stop working after this date.

Action required: Update your webhook configurations before the deadline. This affects Shopify apps using Conversions API, custom integrations, and any middleware connecting Meta to your fulfillment stack.

Graph API v25 Changes

The Page Viewer Metric is replacing legacy reach metrics by June 2026. If you're pulling reach data via API for reporting dashboards, update your queries.

EU Ad Personalization

EU users now have a "less-personalized ads" option following Meta's EUR 266M fine. This means your EU-targeted campaigns may see lower match rates and slightly higher CPMs. Account for this in your EU market budgets.

Stricter Ad Identity Verification

Professional services, financial products, healthcare, and political advertisers face enhanced identity verification. For dropshippers, this mostly doesn't apply — unless you're selling health/wellness products that make medical claims. In that case, expect verification requests.

---

Weekly Creative Cadence

With Andromeda consuming creative faster, you need more volume:

Creative Testing Framework (Post-Lookalike)

1. Start broad. One Advantage+ campaign, broad targeting, 3-5 creative variants.
2. Read the data. After 1,000 impressions per creative, identify which angles the algorithm favors.
3. Double down. Create 3-5 variations of winning angles (different hooks, same concept).
4. Iterate weekly. Replace bottom performers every 7 days.
5. Scale with budget, not audiences. Increase spend on winning campaigns by 20-30% every 3-4 days.

Organic Content Warning

Meta's algorithm is now actively demoting "unoriginal" content — repurposed TikTok videos, recycled images, and cross-posted content without modification. If you're using the same creative on your organic page that you run as ads, the organic reach will tank.

Keep ad creative and organic content separate.

---

None of this matters if your fulfillment can't keep up with ad-driven demand. When a creative hits and Andromeda pushes it to a wider audience, volume spikes fast — faster than the old LAL scaling model.

The old playbook: scale audiences gradually, volume increases predictably. The new reality: algorithm finds a pocket of demand, volume can 3-5x in 48 hours.

Your fulfillment partner needs to handle that volatility. Stock-outs during a winning creative run are the most expensive kind of inventory failure — you're paying for clicks that convert to out-of-stock pages.

For broader logistics and platform context, see our March 2026 Intelligence Report.

---

Are lookalike audiences completely gone from Meta Ads in 2026?

Yes. Meta fully phased out lookalike audiences in 2026. They've been replaced by Andromeda's predictive ML models that use aggregated behavioral signals across Meta's platform. You can no longer create, edit, or use lookalike audiences in any campaign type. Advantage+ automation with broad targeting is the default.

Should I switch all my campaigns to Advantage+ Shopping?

For most dropshippers, yes. Advantage+ Shopping consolidates learning and typically outperforms manual campaign structures after the learning phase. Keep one manual campaign for creative testing, but route proven winners into Advantage+ for scaling. Fighting the automation wastes time and budget.

Do I still need to create custom audiences for retargeting?

Custom audiences for retargeting still work, but Meta now requires you to declare the data source for any remarketing campaign. You need documented consent for customer list uploads, and your privacy policy must explicitly mention Meta pixel usage. Non-compliance risks account restrictions.

Are static image ads still effective, or should I go all-in on video?

Static images still drive 60-70% of e-commerce conversions on Meta. Don't abandon them because Meta is pushing video tools. The best approach is a mix: 50-60% static images, 30-40% video (including UGC), and 10-20% carousels. Test Meta's new image-to-video tool for product launches, but keep your proven static formats running.

What's the biggest mistake dropshippers make with the new Meta Ads system?

Still thinking in terms of audiences instead of creative. The most common failure is running broad targeting with generic creative and wondering why performance dropped. Under Andromeda, your creative IS your targeting. Every visual element, copy angle, and hook tells the algorithm who to show your ad to. Invest in creative volume and variation, not audience research.

Related Intelligence Reports

#eu#compliance#regulations

EU Compliance Deadlines for Dropshippers: 2026 Calendar & Action Plan

Mar 10, 2026•Read

#israel#shipping#disruption

Israel Shipping Disruption: What the Iran Conflict Means for Dropshippers (March 2026)

Mar 10, 2026•Read

#shopify#fulfillment#dropshipping

Shopify Fulfillment Service: What to Look for Beyond Oberlo

Jan 23, 2026•Read

#tiktok-shop#fulfillment#social-commerce

TikTok Shop Fulfillment: What Dropshippers Need to Know in 2026

Jan 23, 2026•Read