The news of the Facebook outage, and shorter but similar outages for Twitter, got me thinking about the long ago spate of outages in the late 1990s.
eBay led the pack with outage after outage in 1999. The company I was working at back then made a lot of hay over it because it was our contention that eBay didn’t have the right kind of infrastructure.
But, then, eBay survived and lots of companies with iron-clad infrastructures were soon out of business and liquidation companies were hawking previously pricey technologies for pennies on the dollar.
Online grocery retailer Webvan was perhaps the most spectacular dot-com bust. A $375 million IPO and a plan to spend $1 billion on warehouses–all to do something that maybe a company (oh, I don’t know, maybe a grocery store chain) with an existing infrastructure could have done just as well.
Much was made of Webvan’s strategy to build technology for the future, but it was all for naught. It tried to grow too fast, didn’t really consider what consumers want, and flamed out in 2001, like so many other dot-coms.
Retailer Value America is another sad tale. Its novel idea of connecting consumers directly with manufacturers seemed like a good idea. So much so, that when it went public in 1998 it had a valuation of $2.4 billion (that’s a B, folks).
You’d think that a company that connected consumers with manufacturers would have thought out the communication aspects of such an operation a bit. Maybe they did. But it never really worked. All the infrastructure in the world couldn’t have kept it alive. It finally vanished for good in the fall of 2000.
Infrastructure might not be everything.
eBay survived 1999 and 2000 not because it invested in its infrastructure (which, of course, it did), but because it had something people valued–connecting buyers with sellers at the paper-thin end of the tail.
At about the same time, Google was springing to life. Did Google invest in infrastructure? Yes, but they didn’t rush out to the big server companies (although I’m sure they bought their gear from someone). They built it themselves using, literally, Velcro. Their philosophy probably was this: why spend money lining the pockets of a server vendor when we can build it ourselves to our exact specifications. Plus, it was probably fun.
So, what’s my point? Facebook and Twitter will survive.
And, during the occasional outage, you might have to emerge from the darkness of your cubicle and actually talk to a colleague or pick up a phone and call an old friend. It’ll be good for you. Take it from me–I’m still waiting for that grocery order I placed 11 years ago. Might have to head down to Stop & Shop myself.