Engineering Productivity at Google: History, Structure, and Metrics

Engineering Productivity, known in Chinese as “工程生产力”, was formally introduced by Google in March 2016 when the original Quality Assurance (QA) team was renamed the Engineering Productivity (EP) team.

Since the end of 2005, Google launched the “Engineering Quality Culture Project” aiming to shift from heavy manual testing by dedicated testers to making every software engineer responsible for quality, encouraging engineers to write automated tests and adopt continuous integration. This initiative continued until 2010.

As the team and application scale grew, the QA department expanded its scope and eventually evolved into the EP department, now led by Engineering VP Michael Bachman, who has been with Google for sixteen years and was one of the earliest members of the EP team.

The EP organization is structured such that, before 2008, there was a single unified department across Google. Today, each product area has its own EP department, typically led by a director or senior director, and a horizontal team coordinates foundational engineering productivity across the company, for example focusing on mobile infrastructure for Android.

Metrics used to measure developer productivity are diverse and include:

A change list (CL) submission time to the code repository.

Number of forced submissions (-f) that bypass normal review processes.

Release frequency.

Time required for a deployment package to reach production.

Number of cherry‑picks (hotfixes) and rollbacks.

Code test coverage.

Outage incidents caused by production problems.

Pager incidents.

Time spent on code reviews.

Number of defects deployed to production.

The goal is to obtain a holistic view that makes it easier to identify problems, locate root causes, and truly improve productivity.

Additional material includes a PPT from a 2018 Bay Area community talk by Michael Bachman (images shown) and a Q&A session covering EP’s organizational structure, cross‑company considerations, and productivity metrics.

For the full PPT, readers are invited to follow the WeChat public account “持续交付2.0” and comment “EP-PPT”.