Adapting Frameworks like DORA and SPACE

How to measure AI developer productivity in 2025 | Nicole Forsgren

Topic 5 of 18

Published Oct 19, 2025 · Clip 16:00–17:41 (1:40)

Nicole clarifies DORA's prescriptive speed/stability metrics (deployment frequency, lead time, MTTR, change fail rate) remain useful for pipeline assessment but must emphasize early AI feedback loops beyond traditional phases. SPACE, as a non-prescriptive framework, excels in AI contexts by evaluating satisfaction, outcomes, activity, collaboration, and flow/efficiency, incorporating trust in nondeterministic AI outputs like hallucinations and reliability.