Report: Small dev teams with highly productive senior developers in the future

A new report from the developer analysts at GitClear takes a look at the productivity of software developers. It suggests that companies could demand more from their developers in the future and - thanks in part to the support of artificial intelligence (AI) - expect more from smaller team sizes.

For his analysis of developer productivity, the Git Commit Count Percentile Stats now presented by GitClear take a look at the number of commits per year as a metric. This shows that full-time developers deliver a median of 673 commits per year, which corresponds to around 2.8 commits per day. There is also a trend towards a decline in the number of commits in connection with the use of artificial intelligence (AI).

Relevance for developers in top organizations

Particularly experienced developers achieve at least 1000 commits per year (median: 1,563). This is relevant insofar as this group - according to the study - coincides with the top positions such as Senior Developer or Staff/Senior Staff Engineer ("L6/L7") in large technology companies such as Google, Amazon or Facebook ("Tier-1"). The figures are therefore intended to provide guidance for those seeking a career in a top company.

The analysis also shows that this group of developers are only active on working days on average, rather than at weekends or on vacation. This indicates that it is currently possible to find a healthy balance between work and private life, even in demanding positions.

Artificial intelligence reduces the number of commits

With regard to the impact of artificial intelligence, the report concludes that the number of commits fell by around 9.5 percent between 2020 and 2023. The study explains this trend, which at first glance seems contradictory, with an increase in commit change density; developers bundle larger changes into fewer commits.

Based on its observations, the report propagates three key trends that will have an impact on expectations of software developers.

1. Organizations will encourage their developers to commit code changes more frequently. They assume that modern tools will allow them to do this and want to ensure transparency for management and facilitate review thanks to smaller work units.
2. Commit messages and pull request descriptions are becoming more extensive, especially as a result of available AI tools that generate corresponding descriptions.
3. Start-ups and teams will shrink in size and favor senior developers. Here, too, the influence of AI can be felt: with modern language models, developers can also implement large ideas with relatively fewer commits. And increases in the productivity of highly qualified developers in particular will contribute to small teams of five or fewer people being sufficient for product development. The report makes particular reference to the 99th percentile, which implements up to ten times as many commits as an average developer.

Public GitHub profiles that recorded between 100 and 20,000 code commits in at least one year between 2020 and 2024 serve as the data basis for the Git Commit Count Percentile Stats report. Cumulatively, the report thus records 878,592 developer years of information. Alloy, a provider of the "engineering intelligence platform" GitClear, is behind the analysis and has now presented the results on its blog. The company became known to a wider public with theories on the deterioration of code quality due to AI.

Even the authors of the study express doubts about the extent to which the commit count is a viable concept for measuring productivity: "The commit count is anything but a perfect barometer for development progress. Manipulating the number of commits is trivial."

(dahe)