DORA Report 2024 (Category Engineering)
I'll continue the story. DORA report 2024 yearThe methodology I described earlier in two posts 1 and 2. Here I will try to tell you about the interesting results of this survey, which was conducted by 3K respondents from different professions from different countries. The study itself covered the following areas: Classic DORA metrics: throughput and stability (lead time, change failure rate, time to recover). On the basis of these metrics, the teams are clustered by levels: elite/high/medium/low. Impact of AI on productivity, delivery metrics, documentation and code credibility Impact of internal development platforms (IDP, internal developer platform) Team productivity and process stability Developer experience, leadership and well-being: job satisfaction, burnout, flow, user-centricity
Interesting points noted by the authors
1. The impact of artificial intelligence (+) Increased flow (flow)Productivity and job satisfaction (+) Improvement of code quality 3,4% (+) Acceleration of code review 3,1% (+) Improving the quality of documentation 7,5% (+) Increasing the productivity of teams and organizations (-) Reducing the delivery capacity of software 1,5% (-) Reducing the stability of delivery 7,2% Interestingly, AI has a negative impact on software delivery performance. The authors explain the possible increase in batch size and associated risks. (larger size - greater likelihood of problems) Another interesting effect led the study authors to hypothesize about the “vacuum effect.” (vacuum hypothesis). The bottom line is that AI simultaneously improves performance and reduces time for important work without reducing the administrative burden and burden of toil – a paradox attributed to AI helping to complete meaningful tasks faster. (meaningful tasks)However, it does not reduce unnecessary duties.
2. Platform engineering (+) Increasing individual productivity by 8% (+) Improving the performance of teams on 10% (+) Increased organizational productivity by 6% (-) Reduced capacity by 8% (-) Reducing the stability of changes to 14% Here, the authors’ hypothesis is that different gates within platforms slow down the delivery of value, as they require compliance with quality/reliability/security requirements, and so on. As a result, it is often easier and faster to roll code past the platform:)
3. Management & Leadership (+) User focus improves product quality and reduces burnout (+) Transformational leadership improves all aspects of productivity (+) Sustainable priorities are critical to productivity and well-being In general, the right approach to management through goal-setting, working with priorities and changes yields results:)
If you look at the results of the survey, it seems that When integrating AI, you should separately control that large changes are broken down into smaller ones + monitor the impact on delivery. Invest in documentation and self-service systems – report points to direct link between document quality and software delivery performance Platform development requires a balance between control and software delivery performance metrics. (for example, lead time) Focus on user-centric design of internal tools and a leadership culture that supports vision and autonomy – this reduces burnout and improves the efficiency of local engineers Do not pay much attention to the clusters where the commands fall - it is better to look at the first and second derivatives. (speed of change for the better and acceleration)
#AI #DevOps #DevEx #Metrics #Processes #Management #PlatformEngineering #Software