Steve Job’s keynote for Macworld on 15 January 2008 began at nine in the morning, San Francisco time. That is five in the evening here in Dublin; still within my working hours. I was able to keep a weather eye on the proceedings with a Twitter browser window open in the corner of my screen. At least four of the people in my Twitter feed were live-blogging the keynote event.
Obviously they were not alone, as the level of Twitter activity (twittering? tweeting? twoottering?) coming from Macworld quickly began to put a serious load on the Twitter infrastructure. Very quickly the updating of posts began slowing down noticeably and after acting erratically just stopped completely about five minutes before the keynote started.
Reloading brought up this nicely crafted error screen. The little bird reminds me of that ironic smiling puppy with the electrical cord in his mouth that is used on The Simpsons whenever an in-show TV station experiences ‘technical difficulties’ or a presenter’s on-air meltdown. The little Twitter bird may be just as cute, but he does not make me feel any better.
Further reloads and the cutesy illustrated error message was replaced by this far less sugar-coated version. Obviously the problem was escalating. Devoid of all decoration, at least this screen presented me with some potential options. Although, given the scale of the meltdown going on, none of those three links actually worked
There is a definite argument that a disappointing customer experience can be mitigated (somewhat) by considered, thoughtful error screens that inform the user. But all of that is absolutely no substitute for getting the product delivery right in the first place. In this case engineering the server-side to take the expected load.
Update, June 2008
Twitter has learned some lessons and put measures in place to take the strain for Steve' Job’s next keynote at WWDC 08. The Twitter infrastructure made it through that keynote without falling over.
No comments:
Post a Comment