Meta has introduced a new metric, Diff Authoring Time (DAT), as part of its effort to better understand and improve developer productivity across its vast engineering teams. Unlike traditional measures such as commit counts or lines of code, DAT focuses on the actual time it takes for a developer to create and submit a code change.
The metric tracks the period from when a developer begins working on a “diff”, Meta’s term for a code change until it is submitted. This includes writing, debugging, testing, and context switching, but excludes waiting for code reviews. By measuring this “inner loop” of coding, Meta aims to uncover friction in workflows that slows down productivity.
According to Meta, DAT has already helped identify inefficiencies and guided improvements in developer tools. For example, introducing type-safe mocking in Hack reduced authoring times by about 14%, while optimizations in the React compiler delivered gains of more than 30%. Improvements in code sharing across teams have also saved thousands of hours annually.
Meta stresses that the metric is not designed to evaluate individual developers. Instead, DAT is aggregated to highlight trends and measure the impact of tooling or process changes at scale. Safeguards are in place to avoid misuse and to protect psychological safety, a critical factor in ensuring that developers do not feel pressured to “game” the system.
Industry observers see this as part of a broader shift toward more sophisticated, experiment-driven metrics in large engineering organizations. By focusing on developer experience rather than vanity measures, companies like Meta hope to build faster, more efficient, and less frustrating workflows. However, the success of such metrics ultimately depends on how responsibly they are applied. Used well, DAT could help reshape the way tech companies measure and improve the work that powers modern software.