¶ Best Practices for Creating AI Group
This document will help you understand the best practices you must follow while tagging campaigns/ad sets in the AI Group. We will learn about multiple parameters you can consider while defining the structure of an AI group along with a real-world example.
You can group campaigns or ad sets into AI groups based on their shared characteristics, giving Pixis AI better control. This allows Pixis AI to gain deeper insights into your campaigns/ad sets and provide Bid/Budget/Creative Rotation recommendations.
¶ Parameters to consider while tagging campaigns/ad sets
Now, to ensure that Pixis AI understands your objectives and recommends the best optimization actions, you can consider tagging your campaigns/ad sets based on the following parameters:
-
Campaign Objective
-
Audience type
-
Budget type
-
Bid type
-
Optimization KPI
-
Region
Note: This document provides general guidelines for creating AI Groups across all Pixis-supported ad platforms(Google, Facebook, & TikTok).
¶ Campaign Objective
All campaigns within an AI group should share the same objective. Ideally, objectives can be awareness, consideration, conversion, etc.
-
Awareness Campaigns: Aimed at increasing brand awareness or maximising reach.
-
Consideration Campaigns: Targeted towards increasing website traffic, engagement, app installations, video views, lead generation, or messaging.
-
Conversion Campaigns: Geared towards boosting conversions, catalogue sales, or in-store traffic.
¶ Audience type
It is important to note that all campaigns/ad sets placed in one AI group should target the same audience type. For instance, if you want to generate website sales for two types of shoes for two different audiences, such as teenagers and elderly people, it is not recommended to group these two campaigns/adsets. If you group them, the AI might lower the spending on the low-performing Campaign/Adset audience type, which can negatively impact your business objective. Therefore, it is advisable to have separate AI groups for different audience types to ensure that each group can reach its target audience with the assigned budgets.
¶ Bid type
To group your campaigns/ad sets under one AI group, you need to ensure that they have the same bid strategy type. A bid is the amount of money you are willing to spend to achieve a specific result from ad platforms. There are three different types of bid strategies on Meta: Highest Volume, CPR Goal, and Bid cap. For Google ads, bid strategies are TCPA, TROS, Maximising conversion, etc.
Note: Campaigns with different bid types should not be tagged under the same AI Group.
¶ Budget type
Apart from similar bids, campaigns/ad sets in one AI group must also share a similar budget range. Here, we need to explain that assets should have similar budget ranges to spend. If the amount to be spent for 1 asset is multiple times larger compared to the other asset then tagging them together is not recommended because when the budget is set on an AI Group level, it will redistribute the budgets and disrupt the spending of low-performing assets. Be it daily or lifetime. (Even if the lifetime budgets are different for 2 ad sets their daily budgets can be the same and budget allocation is taken care of by the budget planner on Optimize.
¶ Optimization KPI
Think of the optimization KPI as a metric that shows how well your bid strategy is meeting your campaigns/ad set goals with your current optimization criteria. The optimization KPI is chosen within your campaigns/ad sets. It can be different from your campaign objective. For example, you can choose conversions as your campaign objective but choose to optimise for link clicks within a campaign/ad set. You must choose your bid strategy with the optimization KPI in mind. The event works like a ground truth for the right bid strategies.
¶ Region
You can tag campaigns/ad sets based on region. Let’s say the goal is customer acquisition from two different geographies (state-wise): Delhi and Mumbai. We have to create two different AI groups for each one. But if the goal is overall customer acquisition from a country—irrespective of the state—we can place both of these campaigns under one AI group.
Note: The Pixis AI redistributes campaign/ad set budgets within the AI Group, scaling up successful campaigns/ad sets and scaling down underperforming ones. This means if you place the campaigns running in Delhi and Mumbai in the same AI Group and if the Delhi campaign is performing better than the Mumbai campaign, the budget for the Mumbai campaign will be reduced and increased for the Delhi campaign. This is undesirable, as no marketer would like to underspend on any geography.
¶ Edge cases to consider while tagging ad sets/ campaigns
Multiple factors are considered when tagging campaigns under an AI Group. Generally, the greater the similarity among campaigns or ad sets within an AI Group, the better. However, there are some important points to keep in mind:
-
Special Spend Requirements: Consider creating separate AI Groups for campaigns or ad sets with specific spend requirements, such as those needing to meet a certain daily expenditure.
-
Segmentation for Retargeting Campaigns: For retargeting campaigns, it's advisable to create distinct AI Groups to ensure effective management and optimization.
¶ Real-World Example for Creating an AI Group
MásMóvil is a prominent Spanish telecommunications company renowned for its extensive array of mobile, broadband, and television services.
In this example, we'll delve into the criteria employed to categorize MásMóvil's 35 Google campaigns into AI Groups. We'll specifically focus on the parameters pivotal for crafting these AI Groups. The following parameters were taken into account for crafting the AI Groups:
|
Campaign Type |
Business Objective |
Bid Strategy |
|
Search |
Leads & Purchase |
CPC(enhanced) |
|
Display |
Leads & Purchase |
Maximize conversion value |
|
Performance Max |
Leads & Purchase |
Maximize Conversion Value (Target ROAS) |
|
Discovery |
Leads & Purchase |
Maximum Conversion |
|
- |
- |
Target Impression Share |
|
- |
- |
Target ROAS |
|
- |
- |
Target CPC |
On the basis of the above campaign parameters, the below 11 AI Groups were created:
|
Campaigns |
Campaign Type |
Business Objective |
AI Group |
AI Group Name |
Bid Strategy |
|
4 |
Display |
Leads & Purchase |
AI Group 1 |
MM_Display_CPC_Google |
CPC(enhancded) |
|
2 |
Search |
Leads & Purchase |
AI Group 2 |
MM_Search_CPC_Google-LowEff |
CPC(enhancded) |
|
4 |
Search |
Leads & Purchase |
AI Group 3 |
MM_Search_CPC_Google-MedEff |
CPC(enhancded) |
|
6 |
Search |
Leads & Purchase |
AI Group 4 |
MM_Search_CPC_Google-HighEff |
CPC(enhancded) |
|
3 |
Display |
Leads & Purchase |
AI Group 5 |
MM_Display_CPA_Google |
Target CPA |
|
4 |
Search |
Leads & Purchase |
AI Group 6 |
MM_Search_Low-ROAS_Google |
Target ROAS |
|
4 |
Search |
Leads & Purchase |
AI Group 7 |
MM_Search_Medium-ROAS_Google |
Target ROAS |
|
2 |
Search |
Leads & Purchase |
AI Group 8 |
MM_Search_High-ROAS_Google |
Maximize Conversion Value |
|
3 |
Discovery |
Leads & Purchase |
AI Group 9 |
MM_Discovery_CPA_Google |
Target CPA |
|
1 |
Search |
Leads & Purchase |
AI Group 10 |
MM_Search_MaxConvValue_Google |
Maximize Conversion Value |
|
1 |
Performance Max |
Leads & Purchase |
AI Group 11 |
MM_Pmax_ROAS_Google |
Maximize Conversion Value (Target ROAS) |