Building expanded targeting for Pinterest ads

Pinterest Engineering
9 min readintermediate
--
View Original

Overview

The article discusses the development and implementation of Expanded Targeting for Pinterest ads, a feature designed to enhance ad performance by automatically connecting Promoted Pins to relevant audiences. It highlights the importance of contextual targeting and the techniques used to optimize advertising strategies on the Pinterest platform.

What You'll Learn

1

How to enable Expanded Targeting during Pinterest ad campaign setup

2

Why understanding Pinterest's unique audience interests is crucial for ad performance

3

How to utilize Textual-based and Embedding-based Retrieval techniques for ad targeting

Key Questions Answered

What is Expanded Targeting in Pinterest ads?
Expanded Targeting is a feature that automatically connects Promoted Pins to the most relevant Pinners without requiring advertisers to specify keywords or interests. This approach simplifies campaign setup and enhances ad performance by leveraging Pinterest's understanding of user interests.
How does Pinterest determine relevant keywords for ads?
Pinterest uses two main techniques: Textual-based Retrieval, which identifies the best keywords based on engagement data, and Embedding-based Retrieval, which utilizes machine learning models to match user context with relevant ads. This ensures that ads are shown to the most interested users.
What are the benefits of using Expanded Targeting for advertisers?
Advertisers benefit from Expanded Targeting as it simplifies campaign creation by eliminating the need to specify keywords or interests. Additionally, it maximizes return on investment (ROI) by delivering ads to the most relevant Pinners, thereby improving overall ad performance.
What techniques are used in Expanded Targeting?
Expanded Targeting employs Textual-based Retrieval and Embedding-based Retrieval techniques. Textual-based Retrieval focuses on generating keywords from engagement data, while Embedding-based Retrieval uses machine learning models to find relevant user contexts for displaying ads.

Key Statistics & Figures

Percentage of advertisers using Expanded Targeting
85%
As of August 2019, more than 85% of advertisers on Pinterest have opted into Expanded Targeting.
Percentage of ad groups using Expanded Targeting
60%
As of August 2019, 60% of ad groups on Pinterest have utilized Expanded Targeting.

Technologies & Tools

Machine Learning
Pinsage
Used for embedding-based retrieval to find the most relevant user context for displaying ads.

Key Actionable Insights

1
Advertisers should enable Expanded Targeting during campaign setup to streamline the process and improve ad relevance.
This feature allows advertisers to leverage Pinterest's algorithms to connect with the right audience, enhancing the effectiveness of their advertising efforts.
2
Understanding the differences between keywords and interests on Pinterest can significantly impact ad performance.
By recognizing which terms resonate more with Pinners, advertisers can tailor their strategies to align with user behavior, leading to better engagement and conversion rates.
3
Utilizing machine learning models for ad targeting can optimize the delivery of Promoted Pins.
By implementing techniques like Embedding-based Retrieval, advertisers can ensure their ads reach users based on contextual relevance, improving overall campaign success.

Common Pitfalls

1
Many first-time advertisers on Pinterest use the same targeting profiles from other platforms, which can lead to poor ad performance.
This occurs because Pinterest has unique user interests that may not align with those on other platforms. Advertisers should take the time to understand Pinterest's audience to avoid this mistake.

Related Concepts

Contextual Targeting
Keyword Targeting
Interest Targeting
Machine Learning In Advertising