Panel Discussion: Let's Improve - Leverage DORA to Im...- Nathen Harvey & Amanda Lewis & Dave Stanke

### Overview At cdCon, Google Cloud DORA team members Nathen Harvey, Amanda Lewis, and Dave Stanke introduce the DevOps Research and Assessment (DORA) framework. Through an opening skit, the speakers illustrate common developer pain points—such as narrow perspectives (IDE-only myopia, ops blindness, or restrictive security "stops")—before demonstrating how the DORA model unites technical, process, and cultural capabilities to predict software delivery performance and organizational outcomes.

### Key Topics - **The DORA Framework & Four Core Metrics**: The speakers explain how technical, process, and cultural capabilities drive software delivery performance, which in turn predicts organizational performance and practitioner well-being. Performance is evaluated across velocity (lead time for changes, deployment frequency) and stability (change fail rate, failed deployment recovery time). - **The DORA Quick Check**: The team demonstrates a live assessment where a sample team's delivery practices yield a 5.7/10 score. The speakers explain that stability metrics were solid, but velocity suffered from long cycle times (two-week to one-month sprint delivery cycles). - **Ongoing Research Areas**: The panel highlights focus areas in current and upcoming DORA annual surveys, including developer experience (DevEx), platform engineering adoption, artificial intelligence experimentation, code review speed, and documentation. - **Team-Level Usage vs. Leadership Anti-Patterns**: The speakers address the common pitfall of leaders weaponizing DORA metrics to compare teams, emphasizing that metrics are experimental guardrails intended solely for local team improvement.

### Technical Details - **DORA Metrics Definition**: During the walkthrough, the speakers examine the four software delivery metrics: lead time from code commit to production, deployment frequency, change failure rate (percentage requiring hotfix or rollback), and recovery time from failed deployments. - **AI and Code Review Findings**: Research from the previous year's report revealed that teams with faster code review speed achieved 50% better software delivery performance. Early AI findings indicated positive practitioner well-being and experimental engagement, though broad impact across all software delivery tasks had not yet materialized. - **Measuring Lead Time**: In response to audience questions regarding measuring lead time from business ideation versus code commit, the panel explains that while ideation involves the "messy front end," starting from code commit offers consistent, rigorously optimizable pipeline boundaries. Teams are advised to track what is practical to measure and focus on trends and bottleneck elimination. - **Identifying Bottlenecks**: Rather than trying to optimize every capability at once, cross-functional teams should identify specific constraint areas. For example, pushing version control coverage from 98% to 100% yields minimal value if automated testing is completely absent.

### Takeaways - DORA metrics are not meant for cross-team comparison or management scorecards; they should guide internal team experimentation and continuous improvement. - Organizations should connect capability investments (e.g., automated testing, cultural practices, documentation) to broader business KPIs using narrative context rather than raw numbers alone. - Survey data and capability catalogs on dora.dev provide baselines, guidance, and common pitfalls for iterative engineering improvements.