Locker Rooms | What Do You Need?
Guide To the [Meta] Learning Phase
Referenced from: https://www.facebook.com/business/m/one-sheeters/guide-to-the-learning-phase
Discover how following learning phase best practices can lower CPA and increase conversions.
OVERVIEW
The Learning Phase
About
The Meta ads delivery system uses machine learning to optimize for your results. Each time one of your ads is shown, the delivery system learns more about the best people to target, times of day to show the ad, and placements and creatives to use. The more an ad is shown, the better the delivery system becomes at optimizing the ad’s performance.
The learning phase is the period when the delivery system still has a lot to learn about an ad set. During the learning phase, the delivery system is exploring the best way to deliver your ad set – actively trying different audiences, placements, and more – so performance has not yet stabilized. The learning phase occurs when you create a new ad set or make a significant edit to an existing ad or ad set.
Exiting the Learning Phase
Ad sets exit the learning phase as soon as their performance stabilizes. Typically, performance stabilizes after an ad set receives around 50 optimization events within a 7-day period.
If your ad set isn’t getting enough optimization events to exit the learning phase (or if the delivery system predicts that it won’t receive enough optimization events in the future), the Delivery column reads “Learning Limited.”
LEARNING PHASE IMPACT
Why the Learning Phase Matters
During the learning phase, ad delivery has not yet optimized, so ad sets in the learning phase are less stable and usually have higher CPA. These graphs illustrate the impact of learning phase on CPA.*
A smaller share of budget spent in the learning phase typically results in a higher share of budget spent on stable performance and lower CPA. Advertisers with ~20% of spend in the learning phase (2nd decile) see 17% more conversions and 15% lower CPA than advertisers with ~80% of spend in the learning phase (6th decile).*
For this reason, advertisers should avoid behaviors that prevent ad sets from exiting the learning phase. If you are spending more than 20% of your budget during the learning phase or have ad sets with the “Learning Limited” status, review the recommendations below.
*This data is as of June 18, 2019 and represents performance in average. Individual ad sets may see different results.
OPTIMIZING YOUR SPEND
Avoid frequent edits
Manual optimizations (edits) are a leading contributor to ad sets not exiting the learning phase. Certain campaign, ad set and ad edits reset the learning phase. These are the edits that will cause an ad set to reenter the learning phase:
Campaign
– Budget (depending on magnitude)
– Bid amount (depending on magnitude)
– Bid strategy
Ads
– Any change
Ad sets
– Targeting
– Placement
– Optimization event
– Adding new creative
– Bid strategy
– Bid amount (depending on magnitude)
– Budget (depending on magnitude)
– Pausing for over 7 days
TIP #1
To reduce spend in the learning phase, avoid editing an ad set or ad until it has exited the learning phase. This ensures your optimization decision is based on results that are more indicative of future performance. If you have multiple edits to make, batching them all at once ensures that learning is only reset one time.
OPTIMIZING YOUR SPEND
Avoid having too many ad sets
High ad set volume is another leading contributor to ad sets not exiting the learning phase. When an advertiser runs too many ad sets at once, each ad set delivers less often. This means that fewer ad sets exit the learning phase and more budget is spent before the delivery system has fully optimized performance.
Our recommended solution is to practice account simplification by consolidating ad sets. When advertisers consolidate ad sets, they also consolidate delivery learnings for their ads.
TIP #2
Try these alternatives to high ad set volume to speed up your time in the learning phase:
- If creating many ad sets to advertise to many small geographical areas, combine similar ad sets into fewer, larger ad sets or use the store traffic objective.
- If creating many ad sets for different placements, use automatic placements and asset customization.
- If creating different ads or ad sets for different languages, set multiple languages in a single ad set instead. Doing so enables you to dynamically show ads in the right language for the audience reached.
OPTIMIZING YOUR SPEND
Avoid low conversion volume
and constrained setups
Other campaign and ad set settings can inhibit an ad set from receiving enough conversion volume to exit the learning phase. Diagnose learning phase performance for a given ad set using Delivery View.
Low budget
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 around 50 optimization events in a 7-day period.
If you set a very small or inflated budget, you provide the delivery system with an inaccurate indicator of the people for whom the delivery system should optimize.
Low bid or cost cap
If you are using a bid cap, target cost, cost cap, or value optimization with minimum ROAS that restricts your ad sets from getting about 50 conversions in a 7-day period, then the ad sets may not exit the learning phase.
Small audience size
Larger audience sizes are more likely to generate enough conversions for an ad set to exit the learning phase.
Infrequent conversion event
If your conversion event happens fewer than 50 times in a week, consider optimizing for an event that occurs more frequently. For example, if you see fewer than 50 purchase events in a week, consider optimizing for add to cart events instead.
Note: Ad sets with longer conversion windows need more time to exit the learning phase.
TIP #3
While avoiding behaviors that prevent ad sets from exiting the learning phase is important, you shouldn’t try to avoid the learning phase completely. Testing new creative and marketing strategies is essential for improving your performance over time, and the learning phase is necessary to help the delivery system optimize your new ads.