Your Developers Are Coding Faster. So Why Is Delivery Still Slow?

The article discusses how AI coding assistants are accelerating the development process, but overall delivery time has not improved at the same pace. The reasoning behind this is that other stages in the workflow, such as code review, QA, testing, and release, remain unchanged. While coding time has been cut in half for a team, the total delivery time has only decreased by 10%, indicating that the delays are not solely occurring in the coding stage.

The article suggests that teams should not focus solely on development time improvement, but also analyze the entire workflow. For instance, if a team reduces development time from 2 days to 1 day, the total delivery time would decrease from 10 days to 9 days, which is only a 10% improvement. The real question to ask is what is limiting the other stages of the delivery path.

To identify the bottlenecks, the article recommends breaking down the delivery time by workflow stage. For example, if the average development time is 1.8 days, code review takes 1.2 days, awaiting QA consumes 3.4 days, and testing takes 1.5 days, it's clear that the delays are primarily in the QA and release stages. The article emphasizes that improving developer productivity alone will not significantly impact overall delivery performance.

To analyze workflow stage times, the article suggests using Jira's changelog history or a Time in Status report. By breaking down the total delivery time by status, teams can identify where the delays are happening and focus on addressing those specific areas to improve overall delivery efficiency.