Rebuilding the core: the MINORI system (2012)

Completion did not promise stability

The heart of this decision was the resolve to abandon the path of patching and getting by, and to rebuild the core banking system itself. The two failures had exposed the limits of a design that interconnected the systems while leaving each of the three predecessor banks’ logic intact. What Yasuhiro Sato chose was not a near-term fix but a full migration onto a platform recomposed into functional units. Its scale — some 350,000 person-months and roughly $4.1B (¥450bn) — would tie up domestic IT engineers for years and bind Mizuho’s financial stamina over the long run. Even so, set against the danger of forever deferring an unfinished task, it must have looked like a price worth paying.

The verdict is still unsettled. The full cut-over, nineteen years in the making, resolved on the technical plane the integration that had been deferred at the group’s birth. But immediately afterward MINORI became the stage for eight failures, proving that building a thing does not promise it will run stably. Where the 2011 failure pulled the trigger on the overhaul, the 2021 failures happened on the brand-new platform itself. A vast IT investment easily makes completion its own goal, and the difficulty of keeping the thing running remains after it goes live. Finishing an integration and operating it stably are separate tasks — Mizuho’s MINORI will be remembered as the case that left that lesson behind, together with a price tag of roughly $4.1B (¥450bn).