In April 2026, the Welsh Government escalated Digital Health and Care Wales — the national body running NHS Wales' digital programmes — to Level 4 "targeted intervention." The number that should stop every CIO cold: roughly £600M in cumulative funding (about £191M a year), headcount up from 675 to 1,263, none of nine major programmes delivered on time — and, by the Finance Director's own account, £0.5M of quantified delivered value.
Inputs are not progress
Headcount grew 80% while delivery declined. That should be impossible under reasonable management logic, yet it's the most common failure pattern I've seen in enterprise transformations. Budget and headcount are the comfortable levers: they look like action and feel like momentum. But they're throughput, not outcomes. Add people to a broken delivery model and you get more meetings, more governance decks, and the same gap — just at higher cost.
The confidence cascade
Wales' own oversight framework named the real killer: "optimism bias and late challenge." Programmes sit green until they fall off a cliff, because no one wants to be the person who escalates early. So DHCW sat in Level 3 "enhanced monitoring" from March 2025 to April 2026 — a year of watching a programme fail without intervening. Monitoring is not management.
Value asserted, not measured
£0.5M against £600M isn't an accounting quirk. It's what happens when every business case promises value and nobody is accountable for proving it. Benefits were asserted at approval and never audited after.
The takeaway: post-mortems rarely blame the technology. They blame a governance culture that rewarded announcements over evidence. Audit value quarterly the way you audit spend monthly — before a regulator does it for you.