Who This Is For / When to Use
Use this guide if:
You just launched a new ad or campaign
You are seeing higher costs early on
You are unsure whether to edit your ad
You want to understand how AI optimization works over time
You want to avoid resetting the learning phase
The Short Answer
You should wait at least 7 days before making significant edits to your ad.
Editing too early can reset the learning phase and delay optimization.
What Is the Ad Learning Phase?
The learning phase begins when:
A new campaign or ad is launched
A new ad set is created
Major changes are made to budget, targeting, or creative
During this phase, ad platforms like Google, Facebook, TikTok, LinkedIn, and others collect performance data to understand how users respond to your ads.
The goal is to determine who to show your ads to, when to show them, and which creatives perform best.
What Is Being Optimized During the Learning Phase?
Each of the following is actively optimized while your ad is learning.
Campaign Structure and Relevancy
AI automatically creates tightly themed campaigns that align with:
Your goals
Your landing pages
Your offer and messaging
This improves ad relevancy, which directly impacts performance.
Ad Creative A/B Testing
The system tests multiple variations of:
Headlines
Descriptions
Images
Videos
Primary text
Different users may see different versions of your ad based on context. Over time, the best-performing combinations receive more delivery.
Audience Targeting and Keywords
During the learning phase, the system:
Evaluates keyword performance
Tests audience interests
Identifies high-performing age and gender groups
Adjusts targeting based on conversion signals
This applies whether your goal is clicks, leads, calls, or sales.
Budget Allocation
Your budget is automatically shifted toward:
The best-performing creatives
The strongest audiences
The most effective placements
Poorly performing assets receive less spend over time.
Ad Placements and Timing
Ads are tested across multiple placements, including:
Google Search and Display
Facebook News Feed, Stories, and other placements
LinkedIn and YouTube (when connected)
The system also optimizes time of day and delivery windows to improve results.
How Long Does the Learning Phase Last?
In most cases:
The learning phase lasts about 7 days
Platforms typically need around 50 optimization events during that time
Performance may fluctuate early, including higher cost per lead or conversion. This does not mean the ad is failing.
Why You Should Not Edit Too Soon
Making significant changes during the learning phase can:
Reset the learning process
Increase costs
Delay performance improvements
Cause unstable results
Examples of changes that can reset learning:
Large budget changes
Major audience edits
Creative replacements
Frequent manual adjustments
Real Performance Example
In one example:
Day 1: Cost per lead was approximately $18
Day 3: Cost per lead dropped to around $5
Day 7: Performance continued improving without edits
Day 14: Cost per lead reached $2.87, the lowest recorded
This improvement happened by letting the algorithms work without interruption.
When You Should Make Changes
The system compares your ad performance to benchmarks.
If Your Ad Is Above Benchmarks
Do nothing
Let the algorithms continue optimizing
If Your Ad Is Below Benchmarks
Edit audience targeting
Adjust creative
Or create a new ad
Typically, less than 10% of ads fall below benchmarks and require changes.
Best Practices
Wait a full 7 days before editing
Avoid frequent changes early on
Monitor benchmarks instead of day-one costs
Let AI testing and optimization run uninterrupted
Make changes only when performance is clearly below benchmarks
FAQ
Why did my cost per lead spike after launch?
Early cost increases are normal during the learning phase while platforms gather data.
Can I make small edits during the learning phase?
Minor edits may be possible, but significant changes can reset learning and are not recommended.
What happens if I keep editing my ad?
Repeated edits can prevent the campaign from ever fully optimizing, leading to higher costs and unstable performance.
How do I know if my ad is performing well?
The system shows whether your ad is above or below benchmarks early on, guiding whether action is needed.