The holiday surge is the perfect time to test new strategies that can turn increased demand into tangible results. One powerful opportunity this year lies in understanding and applying Meta’s Andromeda algorithm effectively.
But before diving into how to test it, let’s start with the big question:
What Exactly Is the Andromeda Algorithm?
In simple terms, the Andromeda algorithm marks a major shift in how Meta delivers ads. Instead of heavily relying on audience targeting, Meta is giving far more weight to creatives.
According to Meta’s official engineering announcement, the platform now allows advertisers to feed a library of diverse creatives into their campaigns while using broad targeting. Meta then uses its machine learning to decide which creative should be shown to which person, based on what it knows about their preferences and behaviors.
- Before: Advertisers picked the audience, and Meta optimized delivery.
- Now: Advertisers focus on creative variety, and Meta decides who sees what.
This isn’t an entirely new concept. Creative refresh and testing have always been essential. What’s changed is the weight the algorithm gives to creative itself, making it the key performance lever.
So, how can ecommerce brands engineer their approach to get the most out of it?
The answer – Structured testing.
The most common campaign structure we see among eCommerce brands is an Advantage+ shopping campaign targeted to a narrow 1% lookalike audience of purchasers, historically Meta’s best-performing segment.
However, with the arrival of the Andromeda algorithm, it’s time to rethink that approach. Here’s what we recommend as a stronger starting point moving forward.
Campaign 1: Advantage+ shopping campaign + broad lookalike audience testing
For most eCommerce businesses, an Advantage+ Shopping campaign is the natural starting point as it is the most effective way to connect your website catalog with real-time demand within Meta’s ecosystem. And if you’re not running one yet, consider this your cue to start.
Steps to structure campaign 1:
- Create an automated product catalog – Push your entire product inventory to potential shoppers.
- Enable advantage+ settings – Give Andromeda the flexibility to optimize delivery.
- Audience targeting – Use a broad 1% – 5% lookalike of past purchasers.
- Campaign flight – Run for a minimum of 30 days to collect meaningful data.
- This setup lets the algorithm automatically match the most relevant product creative to the most likely buyer.
If you are already running catalog campaigns with narrow targeting, try broadening your audience and let the algorithm work harder for you.
If your targeting is already broad but revenue growth has plateaued, consider shifting focus to a mutually exclusive audience segment you haven’t tested yet
Campaign 2: Interest based audiences + creative diversification
Once your product-driven campaign is in motion, layer on a second campaign designed to test creative performance across interest segments.
Steps to structure campaign 2:
- Create an interest-based audience – Build a mutually exclusive interest-based audience (combining broad interests with demographic filters) closest to your brand which has minimum intersection with the audience that you are trying in campaign 1
- Insert multiple creative themes – Use the below examples to create a library of creatives that are mutually exclusive to the product catalog creatives that we are running in campaign 1:
- UGC-style short video clips with voiceovers
- Influencer collaborations
- Holiday sale banners
- Limited-time offer creatives
(Tip: Keep creatives detached from specific product lines.)
- Campaign Flight: Run for a minimum of 30 days to collect meaningful data alongside campaign 1
By taking this approach, you can identify which creative formats truly connect with each audience segment.
Next, it’s essential to monitor key engagement indicators to determine when it’s time to rotate or retire creatives.
What to Measure:
Top of funnel engagement at each creative level – Click to Session Ratio, Average Session Duration, Bounce Rates, Session-to-Add-to-Cart Ratio.
Tracking these metrics allows for fast, data-backed decisions on when to update or replace underperforming creatives.
Bottom-of-Funnel Metrics at a campaign level – Transactions, Last Click Revenue, Assisted Revenue
Keeping an eye on these metrics at the campaign level makes it easier to spot which campaigns are ready to scale and which creatives are pulling the weight.
We’ve often observed that session-to-add-to-cart ratios perform on par with those seen from organic traffic, with overall return on ad spend (ROAS) trending close to 2.
Conclusion: Test Boldly, Scale Strategically
The holiday period isn’t just about spending more — it’s about learning faster.
The Andromeda algorithm gives ecommerce businesses a chance to pair the right creative with the right shopper at the right time.
By combining:
Advantage+ shopping campaign + broad lookalike audience testing
+
Interest based audiences + creative diversification
+
Clear performance measurement
…you’ll uncover what truly drives engagement and revenue.
Broad targeting + strong creative variety + smart measurement = winning strategy this holiday season.
