The Libra magistrates' courts IT project is one of the most extensively documented government IT failures in UK history, a £146M bid that became a £390M liability. The Committee of Public Accounts called it "one of the worst PFI deals that we have seen" (New IT systems for Magistrates' Courts: the Libra project, 44th Report of Session 2002–03, HC 434). The root cause wasn't technology. It was requirements.
Analysed primary sources including the National Audit Office and Public Accounts Committee reports to establish the factual timeline and decision chain. Applied requirements engineering frameworks to diagnose failure at each phase: procurement commitment before requirements validation, single-bidder lock-in, and scope instability under fixed-price contract. Wrote a complete alternative requirements specification from scratch. MoSCoW-prioritised functional requirements with per-requirement rationale drawn from real stakeholder perspectives (Legal Adviser, Finance Officer, Court Manager, Policy Official), and scalability-focused non-functional requirements. JAD workshops and DSDM iterative delivery evaluated as the structural remedies that would have prevented the overrun.
The root cause was committing to a fixed-price, single-bidder procurement contract before redesigning business processes or validating user needs structurally. Scope instability and contractual lock-in were the inevitable result. JAD workshops across representative court types (bringing business users, IT professionals, and decision-makers together before procurement) would have surfaced the operational diversity that made a one-size-fits-all system impossible. DSDM iterative delivery governance would have provided structured checkpoints to catch scope drift before it became irreversible. Staff distrust of formal systems was a known social risk that was never addressed in the requirements process.
This analysis sharpened what requirements engineering actually protects against, not just technical ambiguity, but organisational, social, and procurement risk. A £244M overrun is the cost of skipping stakeholder alignment.