How to launch AI for better performance?

1. Before Enabling AI managed groups

• Use Smart Diagnostics to Filter Campaigns:

• Transition campaigns with strong audit results directly to AI management. Optimize or manually manage underperforming campaigns until stable, then transition to AI.

1.1 Better Suited for AI Management:

• Campaigns with stable orders;

• Campaigns for mature products;

• Campaigns with sufficient budget and sufficient data.

1.2 Not Suitable for AI Management:

• Low data volume: Campaigns with large changes in conversions and unstable spending;

• Identify campaigns with poor budget settings: Look for significant overspending or large discrepancies between budget and actual spending.

• Campaigns with specific ad goals, like targeting organic rankings, require manual management. AI may not accurately bid for desired rankings, as it optimizes based on overall performance.
• Campaigns with automation rules set in external tools or platforms. This includes time-based budget rules in ERPs or "Rule-based Bidding" strategies.

2. Setting up AI

2.1 Understanding AI Personality

  Key points: 

• Risk vs. Reward: A more aggressive AI Personality can achieve goals faster but with higher risk. A more conservative approach is slower but safer.
• Time vs. Cost: A conservative AI Personality might take longer but potentially cost less due to its cautious approach.
• Customization: Tailor the AI Personality to your specific needs. There's no universally "best" setting. Experiment to find what works best for your goals and risk tolerance.

2.2 Setting AI Goals

Match AI strategy with product positioning:

• Product Factors: Can be customized based on the aggressiveness/importance of the product, serving as a reference when setting target ACOS and budget.

  Personality Formula  

  AI Action Space Application Scenarios  

• Bid Control: For sellers with extensive experience in a vertical category and clear CPC preferences (e.g., a historical best range of $1-2 and not allowing bids above $3), they can set the AI bid range to $0.8-2.5.
• Budget Control: If you want each ad campaign to strictly adhere to the budget you set and do not want AI to reallocate the budget across managed campaigns, you can disable the "budget reallocation" feature.
• Budget Increase: If a campaign is performing well and the daily budget is forecasted to be insufficient, allow AI to automatically increase the budget within the budget space to reduce missed order opportunities due to budget constraints.
• Brand Traffic Management: In non-brand traffic campaigns, add brand-related keywords to the "Keywords to Avoid" list. AI will automatically avoid capturing these keywords in non-brand campaigns, protecting the purity of brand traffic.
• Ad Structure Protection: If you have a clear strategy and purpose for each ad when creating them and do not want AI to change your ad structure, you can choose to disable the "Structure Optimization" feature.
• New Product Promotion and Traffic Filtering: For new product promotions, if your goal is to quickly increase clicks and not immediately filter out ineffective traffic, you can choose to disable the "Add Negative Targeting" feature. Once sufficient data is collected, you can enable this feature to clean up ineffective traffic. For certain keywords that need to maintain exposure, even if they are temporarily underperforming, you can add them to the "Do Not Negate Keywords" list.

3. FAQ

3.1 How to Reasonably View AI Performance?

AI performs more confidently with larger datasets.

Extend the lookback window for campaigns with limited data, fluctuating spending, or high-priced products.

A well-structured ad campaign allows for better control over results (Smart Diagnostics).

Smaller daily budgets often lead to more volatile campaign data.

3.2 In What Situations is Manual Intervention Required?

If a campaign is performing well and exceeding its budget but AI has not yet acted, you can manually increase the budget immediately. After the manual increase, if the performance is still good and insufficient, AI will continue to increase the budget.

For campaigns that perform well but have a low average active time, consider increasing the base budget based on the ratio of natural orders to ad orders or profitability.

When the budget is sufficient but spending is insufficient, you can promptly add new targeting to obtain more order opportunities.

During significant market fluctuations (e.g., major promotions, flash sales, external traffic), when substantial bid adjustments are needed, you can manually increase or decrease bids.

3.3 How to Review AI Performance?

It is recommended to monitor daily, review weekly, and adjust ad goals and AI persona according to your promotion rhythm:

Achieving Staged Goals: If the goals are achieved and the overall performance is good, you can optimize or pause poorly performing campaigns and set the next stage goals.

Not Achieving Goals: Identify the key factors preventing goal achievement and manually intervene.

3.4 How Often Does AI Adjust Ads?

Based on the sufficiency of the corresponding ad data (AMS), Xnurta can optimize and adjust ads on an hourly basis. The specific frequency of optimization adjustments depends on ad performance and other factors.

3.5 How Long Does it Take to See AI Performance After Using Xnurta?

3.6 Considerations for Managed Group Settings:

• It is not recommended to place campaigns with particularly large recent ACOS performance differences in the same managed group.
• It is not recommended to place campaigns with different bid cap requirements in the same managed group.
• It is not recommended to mix primary campaigns with secondary campaigns in the same managed group.
• It is not recommended to set conflicting primary and secondary goals, such as pushing for growth with a persona of 5 while setting a target ACOS lower than the actual ACOS.
• It is not recommended to directly set strict ultimate ACOS targets from the beginning; a gradual approach is advisable.