Although artificial intelligence has been used for years, there has been an explosion in the number of “artificial intelligence-based” solutions recently. Meta has been using artificial intelligence to support all its apps and services, including advertising, since 2006, from the early days of the news feed.
Sharing more details on how businesses are using AI to improve ad performance, Meta has announced its new tool. AI Sandbox, which tests productive artificial intelligence opportunities for advertisers, the new features of Meta Advantage, which is an ad automation tool, and Meta’s artificial intelligence infrastructure and modeling are among the prominent ones.
AI Sandbox
The AI Sandbox will serve as a testing ground for early versions of new tools and features, including generative AI-based advertising tools. Meta aims to learn which elements are effective for advertisers and to make these features easy to use in advertising tools.
Meta primarily develops tools such as text variation, background generation and image outcropping to make an ad text more interesting or to improve some aspects of images.
- Text variation: It produces multiple versions of the text to highlight the important points of the ad text, and offers advertisers the option to try different messages for specific target audiences.
- Background generation: By creating background images from text inputs, it enables advertisers to try different backgrounds and diversify images more quickly.
- Image outcropping: By adjusting the aspect ratio of images for different planes such as Stories or Reels, advertisers use less time and resources to use images in different ways.
Meta is currently working with a small group of advertisers to quickly get feedback and improve products. Meta plans to start giving access to more advertisers in July and add more features to its products later in the year.
Meta Advantage
Meta Advantage is a portfolio of automation products that optimize campaign results using artificial intelligence and machine learning to help advertisers save time and money.
It also contributes to personalization by matching ads with the right people at the right time. Last year, Meta combined automation products under this new portfolio so that businesses can find and benefit from automation products more easily.
Observing an increase in the use of these products in this process, Meta stated that the number of advertisers using Advantage+ shopping campaigns on a weekly basis increased three times compared to six months ago.
Meta today introduced Meta Advantage features that will enable advertisers to leverage AI faster, use images more flexibly, and help them measure. These features are:
- Converting manual campaigns to Advantage+ shopping with one click: Businesses will be able to convert their existing campaigns into Advantage+ shopping campaigns, thus benefiting from artificial intelligence quickly and easily. This feature will gradually roll out to all advertisers with Advatage+ shopping access over the next month.
- Using video in catalog ads: Catalog ads allow advertisers to upload a product catalog and advertise with those products. Now, advertisers will not only have to use static images in their catalog ads, they can upload content such as branded video or customer videos. Then, with the help of artificial intelligence and machine learning, the best videos on Stream, Stories, Watch and Reels will be shown to people. Advertisers will be able to use video in manual sales campaigns or Advantage+ shopping campaigns. This feature is currently being tested with a limited number of advertisers and is planned to increase reach later in the year.
- Performance Comparison: This new automated report helps advertisers better understand the positive impact of automation by comparing existing sales campaigns with Advantage+ shopping campaigns.
- Performance improvement with Advantage+ audiences: This new Meta Advantage product will offer new ways to help advertisers reach people and sell. Thanks to Advantage+, advertisers will be able to direct who will see the ad with suggestions by adding their own target audience inputs, instead of restricting their target audience inputs by adhering to strict rules. Thus, Meta’s advertising system will be able to find people who can be interested in the products and buy the products apart from the recommendations. “Audience Controls” option will be available for advertisers who need certain restrictions. This feature is currently being tested with a select number of advertisers, and Meta plans to give more advertisers access for testing in the coming months with the learnings gained during this testing process.
Artificial Intelligence Infrastructure and Modeling Investments
Investing tens of billions of dollars in its infrastructure every year, Meta dedicates a large part of these investments to artificial intelligence capabilities for advertisements. With these investments, Meta aims to ensure that artificial intelligence tools such as productive artificial intelligence, which require more operational capabilities for effective operation, reach their full potential.
Meta started using larger and more complex models in its ad system to improve AI modeling and improve performance and measurement using less data.
53438
