🥸 First, keep in mind that some information is not disclosed by the Meta group and we simply do not have access to it. In this article, you will find an answer based on Meta's open source documentation as well as our own observations due to our business.

Refers to how far back in time the data goes. Meta generally uses data from the past 90 days for ad targeting, although some types of data may have longer or shorter retention periods.
👉 The value "lifetime" means the time period for which the insights data is available. By default, this time period is 2 years or shorter.

For interests, Meta uses user engagement data from the past 90 days to determine which interests to include in its targeting options. This means that interests based on recent user activity, such as interests related to current events or popular culture, may be updated frequently.

For behaviors, Meta uses a combination of observed and user-provided data to determine user behavior, and retention periods for different types of behavior data can vary. For example, some behavior data, such as recent online purchases, may have a retention period of 90 days, while other behavior data, such as historical travel habits, may have a longer retention period.
​
Meta has also implemented changes in recent years to improve user privacy and limit the retention of certain types of data. For example, Meta has limited the retention period for certain types of sensitive data, such as health information and financial information, to only a few days.

If you create a saved audience based on the interest "Starbucks", Meta will include people in your audience who have engaged with Starbucks-related content or activity on Meta or across the web within the past 90 days. This could include people who have liked the Starbucks Meta page, commented on Starbucks posts, visited the Starbucks website, or made a purchase at a Starbucks location using a registered Starbucks card, among other activities.

It's worth noting that the interest targeting options on Meta are based on a variety of data sources, including user engagement data, demographic information, and third-party data providers. Meta uses a combination of machine learning algorithms and human review to determine which interests to include in its targeting options, and these interests may be updated or changed over time.

Overall, if you create an audience based on the interest "Starbucks", you can expect that the audience will include people who have engaged with Starbucks-related content or activity within the past 90 days, but the exact composition of the audience may vary based on a variety of factors.

📚 Learn more from the Meta knowledge base and other sources: