The economics of software teams: Why most engineering orgs are flying blind

Article Summary

Viktor Cessan breaks down the true cost of software teams, showing that an 8-engineer team in Western Europe costs roughly 87,000 euros per month, yet most organizations have no visibility into whether their engineering output justifies that spend. The article argues this financial blindness has persisted for two decades and examines how LLMs are now forcing a reckoning by making it possible to replicate significant portions of software output at a fraction of the cost. It positions this as a structural issue where engineers, managers, and executives rarely connect prioritization decisions to their actual financial impact.

Discussion

• The top-voted rebuttal argues the article misses the fundamental point: figuring out what needs to be programmed is the hard part, not the programming itself, and that insight only comes from actually building something
• Multiple commenters push back on the claim that messy AI-generated codebases can just be thrown at more agents, with one sharing experience of two failed fully-AI-generated projects where agents became completely unable to make progress
• The Slack clone example drew heavy criticism, with commenters pointing out it ignores scale, reliability, observability, and the non-trivial engineering that separates a prototype from production software
• Several note the article is written from a narrow high-tech lens, while industries like utilities spend far more on physical infrastructure than software
• Others argue that reducing engineering to pure financial ROI leads to mediocre products, comparing it to Hollywood-style optimization for minimum acceptable quality

Discuss on HN