Build and launch get the applause. Staying up gets the business. The half of the work nobody photographs is the platform that has to be live every hour hundreds of stores are trading, and the governance that keeps it that way.
Keeping a mission-critical platform live across 450 stores, and migrating 400 of them in phases without stopping trade, is a discipline, not a heroic. It runs on formal change governance, problem management, and an enterprise capability model that maps what the business actually does, around 500 capabilities across ten domains.
Build and launch get the applause. Staying up gets the business.
The day a store lost its server room, trading was restored by moving it to a virtual server in real time. That is the test reliability is actually for: not the demo, the disaster. It is why we treat run-and-reliability as part of the build, not a handover.