Building and Scaling High Performing Technology Organizations
summary
Six years of DevOps Research and Assessment data show software delivery performance predicts business outcomes and can be measured by four metrics: lead time for changes, deploy frequency, time to restore service, change fail rate. High performers deploy multiple times per day, elite performers (7%) deploy faster and recover under an hour, Amazon averages a deploy every 11.6 seconds, and A/B tests reveal about two thirds of features deliver no value. Only 22% of cloud users meet NIST criteria, but doing so makes teams ~23× likelier to be elite.
AI-written summary · about
chapters (21)
- 0:00Intro
- 0:45the enterprise
- 1:58enterprise projects
- 4:55Amazon May Deployment Stats
- 8:22software delivery as a competitive advantage
- 9:18software delivery performance
- 11:08elite performers
- 12:22capabilities that drive high performance
- 14:13what is continuous delivery?
- 16:08technical practices
- 17:19key finding: architectural outcomes
- 20:10key finding: doing cloud right
- 24:32continuous testing
- 25:29lean management
- 26:44lean product management
- 28:34culture impacts performance
- 28:55high trust culture
- 33:11effective teams
- 35:52disaster recovery testing
- 37:18climate for learning
- 37:38autonomy
related talks
47mWhy We Need DevOps Now: A Fourteen Year Study Of High Performing IT Organizations
52mSecrets of High Performing Teams: Science Edition
30mThe Data Behind DevOps: Becoming a High Performer
34mDevOps: Next
31mThe Data Behind DevOps: What Does It Take To...
45mDevOps: The Key to IT Performance
22mKeynote: Double the Awesome
05mInterview