Purpose-built ML models for Apple Ads optimization

We are certified under SOC 2 and ISO 27001 standards, which validate our commitment to rigorous data security and privacy practices.

We do not use any personal data for training our AI models. We train our AI models based on data obtained in aggregated form from Apple Ads and MMPs.

1. Campaign Data

We train our AI models using aggregated data from Apple Ads Management API. Importantly, we do not use the actual text of keywords or associate data with specific campaigns or apps. The model operates on keyword-level aggregates, ensuring that no unique or identifiable campaign settings are used.

2. MMP Data

We utilize aggregated data, such as the sum of specific types of events per day, without associating these aggregates with any particular app or campaign. The way we aggregate this data ensures that in-app events or any other information that could potentially identify a specific user, app or campaign are entirely excluded from our training processes.

3. Market Intelligence

We use aggregated market intelligence data (such as keyword popularity, rankings and other metrics) from third-party open sources to enhance our services. This data is not linked to any specific competitor insights or individual account data, and it is entirely separate from the data collected via Apple Ads or Mobile Measurement Partners (MMP).

Each account functions as a distinct environment, and all decision-making processes occur within isolated contexts. While we may use aggregated data to train our models, this data is anonymized and cannot be traced back to any specific campaign or user. The model applies its learned insights in a way that does not involve cross-referencing data between different ad accounts.

Samba is designed to operate across large accounts and organizations with the transparency and oversight enterprise teams require. Every optimization is traceable, while our Customer Success team continuously monitors model performance and helps investigate edge cases.