Mozilla, like many other organizations, relies on data to make product decisions. However, unlike many other organizations, Mozilla balances its goal of collecting useful, high-quality data with giving its users meaningful choice and control over their own data. Our approach to data is most succinctly described by the Mozilla Privacy Principles. If you want to know what Mozilla thinks about data, the Principles will tell you that.
From those principles come Mozilla's Privacy Notices. They differ from product to product because the data each product deals with is different. If you want to know what kinds of data each Mozilla product collects and what we do with it, the Privacy Notices will tell you that.
From the Principles and the Notices Mozilla derives operational processes to allow it to make decisions about what data it can collect, store, and publish. Here are a few of them:
- Data Collection: Mozilla's policies around data collection
- Data Publishing: How Mozilla publishes (a subset of) of the data it collects for the public benefit
If you want to know how we ensure the data Mozilla collects, store, and publish abide by the Privacy Notices and the Principles, these processes will tell you that.
The data Mozilla collects can roughly be categorized into three categories: product telemetry, usage logs and website telemetry.
Inside Firefox, most Telemetry is collected via a module called "Telemetry". The details of our ping formats is extensively documented in the Firefox Source Docs under Toolkit/Telemetry.
In newer products like Firefox for Android, instrumentation is handled by the Glean SDK, whose design was inspired from what Mozilla learned from the implementation of the Telemetry module and has many benefits. At some point in the near future, Mozilla plans to replace the Telemetry module with the Glean SDK. For more information, see Firefox on Glean (FOG).
When ping submissions from our clients hit our end points, they are aggregated and stored into ping-level datasets. On a daily basis, the information in these pings datasets is summarized and transformed into derived datasets which are easier to reason about and faster to query. You can learn more about this in Guiding Principles for Data Infrastructure.
Some of our products, like Firefox Sync, produce logs on the server when they are used. For analysis purposes, we take this log data, strip it user identifiers and summarize it into derived datasets which can be queried with either BigQuery or Looker. As with product telemetry, this data can be helpful for understanding how our products are used. For example, it can tell us how many people from a particular locale are engaging with a particular service.
Mozilla uses tools like Google Analytics to measure interactions on our web sites like mozilla.org. To facilitate comparative analysis with product and usage telemetry, we export much of this data into our Data Warehouse, so that it can viewed with Looker and other tools.