Today was a fascinating day to watch live-blogging services
fail in a rather dramatic way.
Engadget... Gizmodo... MacRumors... ArsTechnica... and more...
Unless you were under a rock or otherwise hiding offline, you know that today was the big, huge, ginormous "iPhone 5" announcement event from Apple. (It turned out, of course, that there was no iPhone 5, but that didn't stop the media frenzy.)
As you probably also know, Apple does NOT live-stream these events. I think Robert Scoble nails their reasons - it's all about control. The PR folks at Apple are also masters at "creating spectacle". These "events" become the huge media events they are precisely because there is scarcity... you can't get the info unless you are in the room.
And so, the "media" get to be gatekeepers to the knowledge again.
Lacking a live video or audio stream, all the interested techies, media and fanboys must turn to live blogs and to Twitter (and Facebook and Google+) to get their updates.
But boy did those live blogs fail today!
Now, don't get me wrong..
I DO understand that providing live updates to an unknown - but very HIGH - number of visitors is hard to do.
I get that... but still it was interesting to see who survived and who didn't (and I mention both below).
Live Blogs That Struggled
One of the first I saw go was MacRumors, who was originally using a service that embedded "live blogging" directly into their web page. That seemed to fail under the load and they dropped back to simply providing bullet updates on their live page.
I was watching Engadget's coverage for a while and it was great ... until it wasn't:
Even worse, the entire Engadget site seemed to be down at times:
The site went in and out during the course of the coverage but was mostly out for the latter half of the coverage.
Gizmodo's live blog didn't give the same kind of errors, but simply stopped updating for long periods ... and then had problems loading display elements (which I missed capturing):
Ars Technica did better with their coverage up until about 40 minutes into the event when they stopped updating the site and pointed people over to Twitter:
Their coverage came back... and then froze again several more times.
I would have loved to be watching the stats on the traffic these sites were getting as it had to have been a TON of traffic.
Live Blogs That Worked
Still, some sites seemed to work well through it all. And while I have no insight into how much traffic these sites had versus the ones above, it could also be the architecture they chose to use as well as their choice of content.
Ryan Block's gdgt live gave the best experience I found, integrating both text and pictures to provide a great way to know what was going on:
They had a couple of momentary hiccups, but overall they seemed to consistently be publishing more and more content.
Mashable's live coverage was also consistently available, although they went with a more Twitter-esque series of mostly text updates. They added in polls for some more interactivity and also had links to other posts and info. They had a few photos, but not all that many compared to others. However, the coverage was consistent and always there:
Finally, GigaOm's coverage was noteworthy in that they started out from the beginning to just provide simple text updates to a blog post that you had to manually refresh. No auto-updates... no embedded widgets... just a straight-up blog post with a mixture of text and pictures. Not as sexy as other sites, but every time I refreshed it the content was there with updates.
Sometimes, simplicity can win.
(Now, in fairness, because there was no auto-updating and because I wasn't sitting there hitting the refresh (this was all running in the background on my computer while I ate lunch and was doing some other work), the GigaOm site could have gone down several times between my refreshes.)
Again, I do realize that providing this kind of large-scale coverage is hard, but in the era of "web-scale" and with the availability of content delivery networks, caching services, etc., there are certainly options available to companies providing "live blogs" of events.
Now maybe some of the sites that "struggled" had all that kind of stuff in place and still succumbed to the overwhelming traffic.
Regardless, today was an interesting experiment in seeing what worked and what didn't work. Personally, I'd love it if some of the services that had technical difficulties would write up a bit about what happened and how they were hit.
It would help all of us learn how to scale our sites.
And help people get ready for the NEXT Apple event ;-)
P.S. And yes, there were undoubtedly other sites that were offering live blogs of the event... these were just the ones that I happened to know of or find.
If you found this post interesting or useful, please consider either: