Nova's AI Tool - How it works?

Nova's AI Optimization is a powerful tool designed to enhance the performance of ad campaigns. By automatically testing multiple versions of call-to-action (CTA) text and button colors, this feature helps identify the most effective ad variants.

Understanding the Optimization Methodology

When creating a new ad, you can set up multiple versions of CTA text and button colors. Nova's AI will then display these different versions to your target audience. Over time, it learns which version performs best and prioritizes showing that variant more frequently.

Real World Example Below Illustrating CTR Lift

The image above shows two versions of an ad for "Travelrite". The main difference is in the CTA button text:

• Left version: "Discover Spain and Morocco"
• Right version: "Visit Spain and Morocco"

The AI system tested these variants and collected the following data:

• Best Performer: 1.17% average CTR
• Realized CTR: 0.94%
• Baseline CTR: 0.36%
• Performance Lift: 159%
• Worst Performer: 0.36% average CTR

Baseline CTR Definition: The CTR of the poorest performing variant.

Realized CTR Definition: The actual average click-through rate achieved across all variants during the optimization process.

Uplift Calculation:  = (0.94% - 0.36%) / 0.36% = 159% Lift

This example demonstrates how even a small change in CTA text can significantly impact ad performance. The optimized version achieved a 159% lift in CTR compared to the baseline.

Best Practices and Key Points to Understand (also outlined in AI Tool usage Tutorial)

1. Optimal Timing: Implement this feature when creating a new ad, not mid-campaign.
2. Best Implementation: Utilize both CTA text and button color variants for comprehensive testing.
3. Learning Period: Be patient and allow at least 10,000 impressions per variant for accurate results. Analyze the performance of all variants, including the worst performers, to understand the full impact of optimization.
4. Experiment with variants to find the best-performing versions. You can opt to turn off a variant from the Variant management page if you wish. 
5. Performance Metrics: Your ad's average (realized) CTR will typically be lower than the best-performing variant but higher than if you had used only one version. Therefore, it is important to remember that the realized CTR will be an average of all tested variants, not just the best performer.
6. Variant Testing: The original version (default) is also tested against new variants.
7. Ad Freshness: This feature helps prevent "ad fatigue" by keeping your content dynamic, fresh and engaging over time

Why Variants should not be added Mid-Campaign

Adding variants on a Live running ad can temporarily lower performance as the system tests new versions. It becomes challenging to accurately compare "before and after" results due to various factors such as timing, impression volume, and changing default settings. For optimal results, set up variants when initially creating the ad.

Benefits of AI Optimization

1. Improved Performance: As seen in the example, optimization can lead to significant improvements in CTR.
2. Automated Testing: The AI handles the testing process, saving time and resources.
3. Data-Driven Decisions: Decisions on ad creative elements are based on actual performance data.
4. Continuous Improvement: The system adapts to changing user behaviors over time.

By following these guidelines and understanding how the AI Optimization feature works, you can maximize your ad performance and achieve better results from your campaigns