Why most engineering orgs can't justify their teams' cost — and what LLMs change

An eight-engineer team in Western Europe burns roughly €87,000 a month, or €4,000 per working day, yet most engineers and many of their managers can’t quote that number. The author walks through the break-even math: an internal platform team serving 100 engineers needs to save each of them about three hours a week just to cover its own cost, while a customer-facing team needs to protect or generate around 1,740 users worth of revenue at €50 ARPU. Break-even, however, isn’t the right bar — once you account for a 50–70% initiative failure rate plus the long-tail cost of maintaining whatever the team ships, the realistic threshold is three to five times annual cost.

The piece argues this financial blindness is structural, not accidental. Teams measure activity proxies (velocity, story points, deploys) instead of churn impact, activation lift, or hours saved, because the prevailing operating model of the last two decades treated headcount as an asset rather than a capital allocation. Concrete examples — a five-point activation lift worth €25k/month, a one-percent churn reduction worth €50k/month — show how small metric movements dwarf most roadmap items, but only if anyone is doing the arithmetic.

The implicit punchline is that LLMs make this reckoning unavoidable. Organizations that have been carrying large engineering headcount without a viable return calculation are now competing against teams whose per-engineer leverage is materially higher, and the teams that can’t articulate their economic contribution will be the first ones re-priced.