[1/2] How Dora Reports Are Collected (DevOps Research and Assessment)?

If you’re an experienced developer or technical executive, you’ve probably heard of magic DORA metrics, which are often the baseline for evaluating development effectiveness. But have you ever wondered where they came from and how the benchmark for performance groups is assembled? So today I decided to talk about it.

DORA (DevOps Research and Assessment) This is the largest long-term study in the field of assessing the effectiveness of software development and operation processes. (Last year they celebrated their tenth anniversary.). Each year, the DORA team sends out a questionnaire to professionals in technical and related fields. In 2024 year 3 Thousands, and the top three represented areas are technologies (36%)fintech (16%) retail (9%). For greater sample representation, DORA focuses on industry research bases. (Stack Overflow Developer Survey) and tries to reach the most diverse teams, types of organizations and geography (The distribution of DORA participants is similar to the distribution of a survey from Stack Overflow.).

In the report 2024 a detailed description of the methodology, the models being tested, and questionable. By the way, 2023 writer gone From one large model to several small ones, where it is easier to build relationships between factors and check their correlations. Cluster analysis is used to analyze the results: respondents’ responses are grouped according to four key production metrics: change lead time, deployment frequency, change failure rate, failed deployment recovery time. The clusters themselves are formed on the basis of collected data without predetermined “ideal” values: for example, the boundaries between groups. (low, medium, high, elite) They change every year because they are determined by the statistical processing of fresh data. Continuation in the next post.

#AI #DevOps #DevEx #Metrics #Processes #Management