Welcome to the Meta Learning Phase, a pivotal period in your Meta campaign’s life where the intricate interplay of platform learning can elevate campaign results. In this phase, the Meta Ads Delivery System meticulously hones its understanding of the target audience, optimal display times, preferred placements, and the most effective creative assets associated with your campaign.

What is the Meta Learning Phase, Exactly?

While the Meta Ads Delivery System will never stop learning about your campaign, the learning phase is a period specifically dedicated to cultivating this understanding. The process begins with the creation of a new ad or ad set, then it’s off to the races. The learning phase plays a significant role in enhancing the system’s ability to fine-tune performance by delving into the most impactful presentation strategies tailored to your unique ad set. As your ads gain visibility, the delivery system evolves, progressively honing its optimization capabilities. Although the learning phase marks a highly advantageous chapter in your ads’ trajectory, it’s also important to recognize that it often yields less-than-efficient results, underscoring the importance of doing everything you can to transition ad sets out of the learning phase as quickly as possible

Exiting the Media Learning Phase 

You’re never going to look back and say your Meta campaign performed its best while it was in the learning phase – and that’s okay. The good news is it doesn’t last forever. Generally, you can expect to see your campaign’s ad sets exit the learning phase after they accumulate 50 optimization events (calls, form fills, etc.) within a seven-day span. 

If your campaign seems to be taking much longer than that (or doesn’t seem to want to exit the learning phase at all), consider the following tactics: 

Avoid having too many ad sets:  Running a lot of ad sets means each one delivers less often, so conversion volume per ad set will be distributed rather than concentrated in a manner that will provide the best learnings. Thus, the fewer ad sets the delivery system analyzes at once, the more resources can be allocated to each. Creating many ad sets will result in spending additional budget but still fewer sets exiting the learning phase. Instead, if you practice account simplification by consolidation, you’ll combine learnings! 

Avoid making too many edits: Some manual edits can reset learning and delay the delivery system’s ability to optimize – prolonging an exit from the learning phase. Try to wait to edit your ad set until it’s out. I know it can be tempting, but remember that performance is less stable in the learning phase, so your results at this point aren’t necessarily indicative of future performance.

Additionally, not all edits are created equal. It’s not an exact science, so while some edits might not meaningfully impact the direction of the campaign’s performance, some edits will change how your ad set will perform in the future. Only modify ads when you have reason to believe that doing so should improve performance so much that it would outweigh the consequences of re-entering the learning phase.

Edits by category that will cause an ad set to re-enter the learning phase:

• Campaign: Budget, Bid amount, Bid strategy
• Ads: Any change (When a new ad is put into the rotation, that new ad enters the learning phase, but thankfully, existing ads don’t)
• Ad sets: Targeting, Placement, Optimization event, Adding new creative, Pausing for 7+ days

Avoid low conversion volumes and constrained setups: Narrow audience targeting, low budget, low bid or cost cap, and infrequent conversion events can all lead to delays in exiting the learning phase. Here’s how:

Narrow audiences: Larger audience sizes are more likely to generate enough conversions for an ad set to exit the learning phase. Don’t limit your campaign!

Low budget: Setting a very small budget will provide the delivery system with an inaccurate indicator of the volume of consumers the campaign should be optimized to. Since ad sets need around 50 optimization events over a 7-day period to exit the learning phase, the ad set should have enough budget to allow for this optimization volume within the time frame. If your budget doesn’t reflect this, consider allocating more.

Low bid or cost cap: Another way of restricting your ad set from getting the desired optimization volume within a week is by using a bid cap, target cost, cost cap, or value optimization with minimum ROAS.

Infrequent conversion event: If your conversion event happens fewer than 50 times in a week, consider optimizing for a lead that occurs more frequently. For example, if you see fewer than 50 purchases in a week, consider optimizing for add-to-cart events instead.

The Necessary Evil

The Meta learning phase is necessary to help the delivery system best optimize ads, so you shouldn’t try to avoid it altogether. But by meticulously fine-tuning your campaign parameters, understanding your audience dynamics, and maintaining a vigilant eye on performance metrics, you can guide your ad set towards a stage of stability and optimal delivery. So embrace the Meta learning phase as a transformative journey, paving the way for sustained success for your campaign.