Uber's engineering team shares how their approach to measuring AI's impact on developer productivity evolved through multiple phases: from adoption and engagement tracking to ROI measurement and agentic workflows. Key lessons include why 'developer years saved' failed as an ROI metric, how behavioral signals outperform demographic segmentation, the importance of causal analysis over correlation, and why traditional activity-based metrics like PR counts break down with AI agents. Uber's emerging framework centers on feature velocity as a North Star metric, supplemented by flow efficiency, quality, and capability expansion to capture meaningful business outcomes rather than raw engineering output.
309 Impressions