Other companies learn from their mistakes and implement tooling, automation, and redundancy to help reduce the pain of outages. Netflix created their own CDN, when they had the big Amazon outage it was not a total outage for Netflix because the traffic and content is distributed.
Many companies anycast their public DNS terminating traffic at multiple points of presence, so a consumer never actually sees when a site goes down.
Disney's public / consumer facing systems quite often do not meet even basic capacity requirements. How often you see stitch eat the page or MDE not work properly? Often.
Downtime does mean a bad system, every time. It may be acceptable downtime, but it's still bad even when it's planned because it impacts customers and that is never good.
MDE and their reservation/ticketing/dining websites are all served by the same service bus, the same databases, and more than likely the same front end load balancers. The only difference is more than likely that MDE traffic passes to their API servers, and their website traffic passes to their reverse proxies and content servers.