Solved: Best Approach for Uploading Creatives in Meta Campaigns?

Hi there, Thanks for reaching out! Happy to give you some initial thoughts and guidance on your situation with managing a large number of creatives on Meta Ads. It's a common challenge, especially with global brands supplying lots of high-quality assets.

Understanding Meta's Algorithm and Creative Delivery...

You've hit on exactly what happens when you feed the Meta algorithm a large number of creatives within a single ad set. It's built to find efficiencies and performance quickly. It will test a sample of your creatives initially, identify which ones are getting the best early signals (like high click-through rates or driving conversions), and then it pretty heavily weights the budget towards those winners. Think about it from Meta's perspective: its job is to deliver on your campaign objective (conversions, awareness, etc.) as efficiently as possible. If it sees a few creatives are working much better than others based on its machine learning, it would be counterproductive for *its* goal (and yours, in theory, for direct performance) to evenly distribute spend across 40-50 different assets where many are likely underperforming. This is why you see only 2-4 creatives getting most of the love, as you mentioned. It's not ideal for testing every single variant you have, but it's how the platform optimises for performance within that ad set. I've seen this happen across various niches, from B2B SaaS campaigns where we're testing different value propositions to eCommerce where different product angles are tested – Meta just finds the quick winners.

So, How Do We Test Effectively With Lots of Assets?

This is where it gets a bit tricky with large volumes of assets and shorter campaign durations. As you rightly pointed out, simply chucking everything in one ad set means most won't get tested. Batching them in the same ad set isn't ideal because, as you said, you reset the learning phase each time you make a significant creative change. Learning phase resets can make performance unstable for a bit, and doing that repeatedly over a 5-10 week flight isn't great for consistent performance or gathering reliable data. Using separate ad sets for different creative batches is also problematic. If the audiences are the same or very similar (which they likely are if they're for the same campaign and targeting), Meta's auction system will see these ad sets as competing against each other. This drives up costs, makes delivery less efficient, and can lead to the "poor delivery" warnings you're seeing. It undermines the platform's ability to optimise effectively. Given the constraints – high volume of assets, relatively short campaign flights, and wanting to avoid learning phase issues or ad set competition – there isnt a perfect, universally agreed-upon method to give *every single* creative variant a statistically significant test within one campaign flight while still optimising for performance.

Focusing on Strategic Testing & Learnings...

Instead of trying to force equal spend across all 40-50 assets in one go (which the platform resists anyway), you might need to shift the focus slightly. The algorithm *will* find the best 2-4 performers from the pool you give it. The learning you get isn't necessarily which of the 50 was *the absolute* best if given infinite budget, but rather which *types*, *formats*, or *messaging angles* resonated most strongly with your audience *at that time*, based on the ones Meta chose to show.

Here's a way to think about managing this volume:

Challenge	Recommended Approach	Why it helps
Meta focuses spend on few creatives	Focus on diversity within the single ad set	Ensures Meta has various formats/angles to choose from initially
Too many variants for one flight	Rotate full sets of creatives across *different* campaigns/flights	Gives different creative pools a chance over time without resetting learning mid-flight
Avoiding learning phase resets	Make creative updates only when necessary or between campaign flights	Keeps campaign stable and allows for better performance optimisation
Gathering meaningful insights	Analyse performance by creative *type* or *message theme* across campaigns, not just individual asset performance within one ad set	Builds long-term understanding of what resonates with the audience

You could group the 40-50 assets into smaller, diverse sets (maybe 10-15 creatives per set, ensuring a mix of images, videos, carousels, and different messaging points) and run one set per campaign flight. This way, each flight gets a fresh, diverse pool for Meta to work with, you avoid mid-flight resets, and over several campaigns, you get to test a significant portion of your asset library. Your analysis then becomes about which *sets* performed best, and digging into the top performers within those sets, rather than trying to get equal impressions for all 50 in one go. It’s about managing expectations with the platform's behaviour and shifting your testing strategy to look at performance trends over time and across different creative groupings, rather than expecting a perfect test of every single asset in a short window. Managing campaigns for global brands, especially with large budgets and high volumes of assets, adds layers of complexity. Strategic planning around creative testing, data analysis across multiple campaigns, and understanding how the platform behaves at scale are crucial. Sometimes having an external expert eye can help structure this process to ensure you're getting the most value and learnings from your significant creative investment. If you'd like to explore this further, we'd be happy to book in a free consultation to chat through your specific setup and goals. Regards, Team @ Lukas Holschuh