Users experienced numerous problems ranging from unable to log in to finding the SR numbers revamped and some going dizzy when the Flash player hung up at 90% mark. The outage lasted more than a day, for sure and my mailbox was flooded with die-hard fans of the Old Metalink spitting their venom over the new makeover.
Yes, its true, it appears that the folks at Oracle didnt get their act correct - not that they have got it right on all earlier occasions, but something drastic seems to have been missed out when the cutover happened. Did they not anticipate the load? This may probably not be the case because afterall, Oracle wasnt releasing anything spectacular - like a MJ-This is it. There are pretty slim chances that their severs crashed due to an upsurge in hits. It appears that the folks at Redwood Shores' blue towers missed some links in the Performance testing which led the users to vent their venom.
Anyway, that brings up an important parallel. Imagine such a situation when a fragile module like Oracle Self-Service goes live. It is very essential to have a solid back-up plan to handle any eventuality including ability to extend the life-support of the legacy systems, having a trained front-end helpdesk staff to handle questions from the users and guide them in navigation etc.
A poorly planned Helpdesk is bound to create a high degree of frustration for the new users leaving a bad patch that takes a long time to clear up. It is very important to incorporate these factors into the Project plan and prepare for the contingency and mitigation.
Talking about poor support, I remember seeing a mail from a frustrated end-user to his manager "This is the worst application I have every seen. If you have paid 1 cent for this product, let me tell you - you have paid too much" !!
I will leave you to guess what the product / application was.