Use Xnurta AI to optimize ad campaigns

1. Function Introduction

Xnurta AI can help you manage your campaigns. Utilizing AI algorithms and based on your optimization objectives and historical ad performance data, Xnurta intelligently optimizes the bidding, targeting, budget, and structure of the campaigns under its management. Overall, Xnurta AI will make the following optimization adjustments to campaigns based on your business goal and target ACOS:

• Bidding optimization: Includes dayparting strategies, base bid adjustments, and bid placement adjustments.

• Target optimization: Includes keywords and ASINs harvesting and negation. Ad groups with category and audience targeting are not currently supported.

• Budget optimization: budgets are adjusted to periods with a high level of efficiency, including budget dayparting, budget increases and budget reallocation across campaigns in a managed group.

• Structure optimization: Create or pause campaigns, ad groups or promoted ASINs based on performance data.

 

1.1 Optimization Objectives Supported by Xnurta AI

1.1.1 First Priority: Business Goals

Business Goals		Use Cases
Increase Sales	Drive Growth	New products or products in the growth stage, where sales volume has room for further growth.
	Promotion Sales Boost	During platform or seller events, traffic needs to be captured quickly and orders accumulated.
Control Cost	Maintaining Stable Orders	Mature or profitable products, where sales have approached peak levels, and the focus is on gradually reducing ACOS while maintaining stable orders to increase profits.

 

1.1.2 Second Priority: Target ACOS

The expected target ACOS to achieve the business objective.

• When the first priority objective is [Increase Sales]: The target ACOS should be relatively loose, prioritizing sales volume growth before switching to cost control.

• When the first priority objective is to [Control Cost]: The target ACOS setting should be stable or slightly lower based on historical data.

Xnurta will automatically generate a recommended target ACOS and range. Please set the target ACOS reasonably using the reference data. If the target ACOS differs too much from historical data, it may lead to significant adjustments and data fluctuations. When the performance of the campaign at its current stage differs significantly from your ultimate goal, it is still important to follow the above guidance when setting targets, achieving large-scale optimization results through staged target adjustments every 2-3 weeks.

 

1.1.3 Optimization Objective Setting Guidelines for Campaigns at Different Stages

Campaign Status	Status Characteristics	Setting Guidelines
Brand New ASIN	No data accumulation, AI cannot directly reference the product's historical data	Recommended to adopt Drive Orders Increase as the optimization objective. Set ACOS based on store data provided by the platform and keep it as high as possible to give AI more space for learning and early data accumulation. The data observation period should be at least 21-28 days.
Short Launch Time, Limited Orders	There are ad clicks data, but data volume is small	Recommended to adopt Drive Orders Increase as the optimization objective. ACOS setting can reference the product's historical data and store data and keep it as high as possible to give AI more space for learning and early data accumulation. The data observation period should be at least 21-28 days.
Launched for Some Time, Certain Amount of Orders	Large amount of data accumulated over a period of time	ACOS setting can reference the product's historical data and adjust the target ACoS based on business goals. The data observation period should be at least 14-21 days.

 

1.2 Optimization Actions Taken by Xnurta AI

Currently, the optimization strategies and adjustments supported by Xnurta AI include:

Optimization Category	Optimization Strategy	Description	SP	SB/SBV	SD
Budget Optimization	Budget Dayparting	For ad campaigns with sufficient data and over-budget situations, Xnurta uses AMS(Amazon marketing Stream) hourly data to identify performance differences throughout the day. By setting time-based budgets, it controls ad spending during high conversion rate periods.  The metric shows the percentage of time lots in high conversion periods where over-budget situations did not occur. The higher the better.	✅	/	/
	Budget boost	For campaigns that allow Xnurta to increase the daily budget, Xnurta considers ACoS, order volume, and expenditure levels to make a judgment. When the ad campaign performs well and the estimated budget is insufficient to support all-day spending, the budget is increased to seize the momentum and continue the success.	✅	✅	/
	Budget Reallocation	If campaigns in a managed group show significant performance differences and some campaigns are out of budget, AI will reduce the budget for campaigns with poor performance, while increasing the budget for campaigns with high performance.	✅	✅	✅
Targeting Optimization	Harvesting	If AI identifies potential keywords or ASINs that have not yet been added to the campaign for the same promoted product, and are found in the search term report, consider adding them if budget allows.	✅	✅ only keywords	✅ only  ASINs
	Negation	Based on the historical performance of search terms and combined with the TOPSIS method, root word analysis, and predictive models, Xnurta calculates the comprehensive contribution and cost of search terms to negate inefficient or ineffective search terms, reducing unnecessary ad spending.	✅	✅ only keywords	/
	Keyword/ASIN pausing&supplementing	Pause underperforming keywords/ASINs and add new targets	✅	/	/
Bid Optimization	Rapid Testing	For targeting and keyword harvesting campaigns created within 7 days, Xnurta conducts rapid bid tests to obtain clicks at minimal cost.	✅	✅	✅
	Bid Dayparting	Xnurta analyzes AMS hourly data to determine performance differences at different times of the day. Based on the performance levels, it increases or decreases bids at different times.	✅	/	/
	Base Bid Adjustment	Xnurta adjusts bids by analyzing different ad metrics and levels of data over various look-back periods, combining model predictions and a risk control system.	✅	✅	✅
	Placement Percentage Adjustment	Based on certain data foundations, when there are significant differences in CVR between ToS, PDP, and RoS, Xnurta adjusts placements by judging the CVR performance of these three positions.	✅	/	/
Structure Optimization	Promoted Products, Ad Gourps, Campaigns pausing	Xnurta pauses underperforming promoted products, ad groups, and ad campaigns, reallocating their budgets to more efficient targets to improve the overall efficiency of the managed group.	✅	/	/

User-initiated budget changes to campaigns with AI enabled take effect immediately, without waiting until the next day: (1) Budget changes made in Xnurta take effect in real-time;

(2) Changes made in the Amazon backend may cause budget fluctuations for a short period (up to 1 hours) after the change due to the time difference in Xnurta pulling data from Amazon, after which it will stabilize to the user-modified value. At 23:00 site time each day, based on the daily budget status (whether out of budget) and the sufficiency of AMS data, the budget dayparting plan for the next day is calculated and updated daily.

 

1.3 More Configurations of Xnurta AI

In addition to configuring the two prioritized objectives, Xnurta also provides more personalized configurations:

Configuration Item	Description
AI Personality	Sellers can control the aggressiveness of AI through the AI Personality setting, which is reflected in multiple aspects such as the magnitude of AI adjustments and reference values. There are 5 levels from conservative to aggressive. The more aggressive the AI, the higher the adjustment rate, and the more likely you are to reach your goals faster. At the same time, you also need to bear more potential risks, such as higher spend and some degree of order decline. The AI Personality can be freely combined with the two main AI optimization objective directions of increasing sales and controlling costs. For example: A more aggressive AI Personality paired with the goal of driving growth can achieve the effect of rapid scaling; a more aggressive AI Personality paired with the goal of maintaining stability may reach the target faster when the target ACOS is set reasonably.
AI Action Space	In section 1.2, sellers can freely choose which AI optimization actions to enable. Currently, the AI optimization actions that can be freely turned on or off include: bid optimization, budget dayparting, budget increase, budget reallocation, harvesting, negation, keyword/ASIN pausing&supplementing. Additionally, sellers can incorporate more of their operational experience into the AI: During bid optimization, control the AI's bid range to reasonably control CPC. During harvesting, set keywords that should not be harvested. When adding negative targeting, set keywords that should not be negated. This feature is available to Lead and Enterprise customers Know more about AI action space, please click AI Action Space Settings: Helping You Get Started with AI Faster
Campaign Name Tag Status	When enabled, the system will modify the campaign name based on the AI on/off status to prompt users whether the campaign is being managed by Xnurta AI. When AI is turned on, the campaign name will start with "Xnurta AI"; when AI is turned off, "Xnurta AI" will be removed. Assuming sellers operate ads through the Amazon backend or other systems, they can also know which campaigns are being optimized by Xnurta AI, reducing misoperations.

Too few actions enabled will limit Xnurta AI' potential performance. Therefore, Xnurta recommends the following settings:

• Enable as many action spaces as possible.
• Do not set a bid range, or avoid setting a bid range that is too narrow or lower than the CPC of the past 14 days.
• For ad campaigns with special strategy considerations, set keywords that should not be harvested and keywords that should not be negated.

 

2. Use Cases

• Campaigns and promoted products are in the growth stage and need to increase sales volume.

• Campaigns and promoted products are in the stage where sales have peaked, and there is a desire to control ad spend and increase profits.

• Sellers do not have enough time to spend on granular ad management.

 

3. Usage Conditions

• All marketplaces authorized by Xnurta are supported.

• AI can be enabled for SP, SB, and SD advertising campaigns. SP campaigns already support all the optimization strategies above, while SB and SD ads are gradually being expanded.

 

4. How to Use

4.1 Add a New Managed Group

There are three entry points for creating a new managed group, including:

  Entry Point 1: In AI Copilot - Creation Page  

Click the Add Managed Group button. You can directly create a new managed group and add existing campaigns to it. If you have used the smart creation feature before, you can click Add Campaigns to combine other campaigns with the smart creation campaigns.

When choosing your campaigns, you can filter campaigns by portfolio/promoted product/ACOS range.

After selecting the campaigns and clicking Confirm, you will be redirected to the managed group editing page, where you can configure AI management-related settings.

 

  Entry Point 2: Product Center - Product Analysis - Ads Page 

Select the corresponding campaigns and choose to add them to a managed group or create a new managed group in Bulk Operation.

 

  Entry Point 3: Ads Management - SP/SB/SD Campaign page  

Select the corresponding campaigns, and choose to add them to a managed group or create a new managed group in Bulk Operation.

 

4.2 Modify Managed Group Settings

Click the pencil icon to enter the managed group editing page.

On the settings page, you can:

• choose to enable/disable AI
• perform basic and advanced settings for the managed group, such as setting business goals, target ACOS, and daily budget
• modify AI action space

You can also modify AI action space on the managed group list.

 

4.3 Add New Promoted Products to the Managed Group

On this page, you can add promoted products to the managed group. The selected products will be added to all campaigns within the group.

 

4.4 Manage Campaigns in the Managed Group

Click the AI icon to enter the managed group details page, where you can add campaigns to the managed group or remove selected campaigns from the managed group.

 

4.5 Delete Managed Group

Click the trash can button to delete the managed group. You can choose to delete the managed group directly or add the campaigns in this managed group to other managed groups. Managed groups with AI running do not support deletion!

 

4.6 Bulk Operation

Select managed groups to perform relevant batch operations on multiple managed groups. Please enable AI management first before modifying other settings.

 

4.7 View AI Performance

4.7.1 View Overall Performance

Please refer to the AI Copilot --> Summary

 

4.7.2 Analyze the Effects of a Managed Group or Campaign: AI Details Page

Click the AI icon visible throughout the system to specifically analyze data for a single managed group or campaign.

At the top of the page, you can search for managed groups or campaigns.

View the ad performance trend chart after each AI objective setting or adjustment through the AI Management Details page to observe ad performance and provide a decision-making basis for subsequent adjustments.

Overall performance of multiple campaigns within a managed group.

View detailed data for each campaign within the group and manage the campaigns in the managed group.

 

5. FAQ

  How often does AI adjust ad campaigns?  

Xnurta can optimize and adjust ads on an hourly basis, depending on the sufficiency of the corresponding ad data (AMS). The specific frequency of adjustments also needs to consider the ad's performance and other factors.

 

  How can I view AI's changing history?  

In the Ads Management --> Sponsored Products --> Log page

 

  How long does it take to see the performance of AI after using Xnurta? 

Depending on the expenditure and data situation of the corresponding ad, it generally takes 2-3 weeks to observe concrete optimization effects. However, if the data volume is very sparse, it will take more than 4 weeks to determine the effect. 

 

  How keyword harvesting works? 

Harvesting logic: Extract efficient search terms based on the number of orders and ACoS (influenced by AI personality; the higher the AI personality, the lower the harvesting threshold) and use them as targets for advertising.

Harvesting Sources: The Ad ASIN has corresponding broad targeting types such as auto, phrase, broad, category, and ASIN expanded.

👉🏻 Requirements for Harvesting Destinations:

• The promoted products in both the harvesting source and destination must be the same. 
• The destination ad group must has sufficient budget.
• The destination ad group must have only one match type. If more than one, then AI won’t harvest.

Flow Method: Match type will be from broad to narrow. keyword: auto → broad → phrase → exact

 

  What to be aware for budget dayparting? 

• Customers are advised to only intervene with the budget through Xnurta.
• It is not recommended for customers to make adjustments in the Amazon console.
• Avoid making adjustments around the hour or between 22:30-23:00.
• The budget intervention value should not be the same as the original budget value.

 

  How does promotion sales boost differ from drive growth?  

Promotion sales boost will attempt to add even more traffic entry points than drive growth, and will try to keep campaigns in budget. Here are specific diffrences for each action:

1. Base bid: Increase bids for targets with good long-term performance but declining recent performance, and for targets with good short-term performance and low budget consumption rate.
2. Bid dayparting: In general, only increase bids during high-efficiency time periods, and won't decrease bids lower than base bids. During Prime Day, if there is a risk of insufficient inventory, the bid will be decreased.
3. Budget dayparting: Won't do.
4. Budget Increase: In general, more likely to trigger a budget increase to ensure sufficient budget compared to non-promotion sales boost goals. During Prime Day, If there is a risk of insufficient inventory, the budget will be reduced.
5. Negative Keywords: Reduce the frequency of adding negative keyword. During Prime Day, won't add negative keywords.
6. Keyword harvesting: More likely to add new high-performing keywords.
7. Structure: Won't pause keywords, ad groups and campaigns.