¶ Sandbox TAI: Pmax/Demand Gen/Youtube
In this article, you will learn how to generate audience signals & audience segment recommendations for Pmax, Demand Gen, and YouTube campaigns.
This feature provides in-depth insights into the interests, behaviors, and demographics of your target audience. By giving Pixis your brand information, it can analyze your data and create specific audience signals and segments, leading to more focused and efficient marketing campaigns. This data-driven approach assists in tailoring your ad content, maximizing ad spend, and ultimately achieving a greater return on investment.
¶ How to generate recommendations?
Follow the steps below to generate recommendations.
¶ 1. Filling Out Brand Information:
The following information will help Pixis AI generate recommendations specific to your brand that help in optimizing campaigns.
-
Brand Name: Enter the brand name.
-
Brand Website: Provide the URL of the brand's homepage.
-
Country: Select the country for the target audience from a dropdown menu.
-
Context Description: Describe the target audience, including their age range, preferences, and behaviors.
-
Brand Description: Detail the brand's history, values, and the type of products/services offered.
Once all fields are filled, click the "Submit" button to proceed.
¶ 2. Request Processing:
After submission, the platform processes the request. A JSON response appears indicating success and providing a request ID.
-
The request ID can be used for future reference or tracking.
¶ 3. Know Requests Status
To check the status of the submitted request, go to the "Know Requests Status" tab. This panel shows a list of all submitted requests with columns for the request ID, status, error descriptions (if any), and results. You can also use the search bar to find specific requests by entering the request ID. You can download the results as a CSV file for detailed analysis.
You can download the CSV file to view recommendations. The CSV file you download will have two worksheets:
-
Audience signals
-
Audience segmentation
You can view a sample CSV file Here
¶ Audience signals
Generating audience signals involves collecting and analyzing data points that help you understand your target audience’s interests, behaviors, and demographics. Creating audience segments involves grouping your audience based on shared characteristics or behaviors. This segmentation allows for more targeted and effective marketing. The sheet provides you with audience signal recommendations. Here’s a breakdown of the column headers that define audience signal recommendations.
|
Column Header/Index |
Description |
|---|---|
|
ICP Name (B) |
Contains the names of different Ideal Customer Profiles (ICPs) or target audience segments such as Marketing Professionals, E-commerce Entrepreneurs, Tech Startups, Digital Agencies, Data Analysts, etc. |
|
ICP Name |
Lists the names of each ICP, matching the columns to specific audience segments. Each column under this row represents a distinct target audience. |
|
Short Description |
Provides a brief description of each target audience, explaining who they are and what they are interested in. This helps in quickly identifying the characteristics of each ICP. |
|
Interests |
Lists specific interests relevant to each target audience segment, likely to be used for interest-based targeting. This includes hobbies, activities, or professional interests. |
|
In-Market |
Identifies the market categories that each target audience is actively searching in or purchasing from. This indicates their buying intent or current market engagement. |
|
Life Events |
Details significant life events that might be relevant for targeting, such as job changes or business creation. These events can be crucial for timing ad campaigns. |
|
Detailed Demographics |
Provides more detailed demographic information, such as age ranges and employment status. This helps in creating more specific and effective ad campaigns. |
|
Affinity |
Lists categories and interests that the target audience has an affinity for, indicating deeper engagement or preference. This helps in identifying broader interests related to the main target. |
|
Similar Websites |
Names websites that are similar to those the target audience might visit, which can help in contextual targeting and finding related audiences. |
|
Keywords |
Lists specific keywords relevant to each target audience, useful for keyword-based targeting. This helps in refining the ad content to match audience searches. |
|
Demographics |
Includes basic demographic information such as gender and age, providing a foundation for demographic targeting. This ensures the ads reach the right age and gender groups. |
¶ Audience segments
Audience segments are detailed groupings of potential customers based on shared characteristics and behaviors. They help you target your marketing efforts more precisely. The sheet provides you with audience segment recommendations. Here’s a breakdown of the column headers that define audience segment recommendations.
|
Column Header/Index |
Description |
|---|---|
|
type |
Specifies the type of audience signal, such as Life Events, In-Market Audiences, or Affinity Audiences. |
|
descr2 |
Provides a more detailed description of the audience signal, combining relevant categories and segments. |
|
child_segment |
Describes the more specific sub-segment within the broader parent segment, providing finer granularity. |
|
parent_segment |
Identifies the broader segment that the child segment belongs to, giving context to the more specific targeting. |
|
child_cosine_score |
Represents the cosine similarity score of the child segment, indicating how closely related it is to the target audience. |
|
parent_cosine_score |
Represents the cosine similarity score of the parent segment, indicating how closely related it is to the target audience. |
|
score |
The overall relevance score combining both child and parent segment scores to indicate the match quality. |
|
ICP Name |
Contains the names of the Ideal Customer Profiles (ICPs) or target audience segments, such as Marketing Professionals. |