IT Revolution: only 8% time building
- IT Revolution said April 27 that data from its Project to Product reports shows software teams spend just 8% of delivery time building code. - The same dataset, covering more than 600 organizations and 8,000 value streams, found 48% of time goes to analysis, approvals, and coordination. - The claim lands as firms rush into coding copilots while DORA says AI mainly amplifies existing systems, good or bad. (itrevolution.com)
IT Revolution said the bottleneck in software delivery is no longer writing code. Its April 27 article says teams spend only 8% of end-to-end delivery time actually building software. (itrevolution.com) The figure comes from the 2023 and 2024 Project to Product State of the Industry Reports, which IT Revolution said analyzed more than 600 organizations and 8,000 value streams. In the same dataset, 48% of delivery time went to analysis, approvals, and stakeholder coordination. (itrevolution.com) Another 44% went to post-development testing, release management, and infrastructure overhead. That leaves coding as the smallest slice of the delivery timeline in IT Revolution’s breakdown. (itrevolution.com) The argument is aimed at companies buying artificial intelligence coding tools to speed up developers. If most elapsed time sits outside coding, faster code generation does not remove the biggest delays. (itrevolution.com) Google Cloud’s 2025 DORA report makes a similar point in different language. It says artificial intelligence acts as an amplifier, and the biggest returns come from workflow clarity, internal platforms, and team alignment rather than tools alone. (services.google.com) DORA said its 2025 findings were based on survey responses from nearly 5,000 technology professionals and more than 100 hours of qualitative research. The report says AI adoption now improves throughput, but still increases delivery instability. (services.google.com) IT Revolution ties that gap to operating model design. Its article says many enterprises improved Agile and DevOps practices at the team level, but left funding, approvals, decision-making, and measurement systems largely unchanged. (itrevolution.com) The proposed fix is a product operating model, where funding stays attached to long-lived value streams and teams remain stable over time. In IT Revolution’s description, success is then measured through adoption, revenue impact, and business outcomes instead of project completion against scope. (itrevolution.com) That reframes approvals, policy checks, release steps, and infrastructure setup as delivery work, not side issues. If those steps still require handoffs across separate functions, AI can make code arrive faster than the organization can safely review, test, and ship it. (itrevolution.com)
