¶ FAQs for AI Grouping
-
What happens if we don’t tag the campaigns into different AI Groups once we change the campaign bid strategies?
-
Pixis AI will continue to optimise budgets even if campaigns/ad sets have different bidding strategies. However, bid optimization stops because the PAI(Performance AI) model identifies an error that campaigns/ad sets under this AI Group have different bid strategies.
-
-
Do we need similar budgets while grouping campaigns/ad sets with similar bid strategies?
-
It is not necessary in NUM (New Unified Model). In NUM, the target results are dynamic based on budgets. In legacy models, yes, we need to ensure we don't have assets with varying budgets.
-
-
Do we need to add a campaign-type column before adding campaigns to the AI groups?
-
Yes, we need the campaign-type column as it helps the Pixis AI plugin identify your campaign details. You will get an error message without this column when creating an AI Group.
-
-
Why is it important for campaigns within an AI Group to share the same objective?
-
Campaigns within an AI Group should share the same objective to ensure coherence in optimization strategies. For example, awareness campaigns focus on brand awareness or reach, while consideration campaigns target website traffic, engagement, app installations, and other metrics.
-
-
What is the significance of bid type in AI Group creation?
-
All campaigns within an AI Group should have the same bid type to ensure consistent bidding strategies. Different bid types, such as cost cap, bid cap, and lowest cap, require distinct optimization approaches and should not be mixed within the same AI Group.
-
-
What are the different phases we have, and what do they do?
-
Phase1 : Observation
-
During this phase, the Pixis AI analyses and compares data from every AI Group to identify the most suitable Optimization AI Model with the highest Model Confidence Score.
-
-
Phase2 : Recommendation
-
During this phase, the AI Optimization model recommends actions with a higher Action Confidence Score. This phase concludes when at least one action in each AI group is validated and executed with a high Confidence Score (~0.8)
-
-
Phase3 : Calibration
-
During this phase, the Pixis AI automatically executes all the recommended actions with an improved Action Confidence Score on every AI Group of your Ad Account.
-
-
-
What determines the AI group moving from Recommendation to Calibration?
-
You can review AI recommendations and, based on your expertise, move them to the calibration phase, where you will enable the selected recommendations to be automatically executed on your ad platform.
-
-
Does any spend typically move the account from one phase to another?
-
No, a phase change is based on the marketer’s decision to execute and generate recommendations.
-
-
Do we have results next to the recommendation made? (what impact has been achieved because of the specific recommendations being made)
-
We have the metadata/ Action Basis to support recommendations, and we are working on the performance achieved. However, we currently do not have a feature to determine the impact of the recommendation.
-