This is one of those cautionary tales from the dot-com boom. Do we really need another, you ask? Well, there's always room for a classic story of hubris, terrific marketing, and a total misunderstanding of the importance of infrastructure.
In 1999, I was involved in a startup consumer Web site that decided to launch before the technology was fully baked. The thinking -- remember those days?-- was that we didn't want anyone to get the jump on our brilliant idea. So our little company put its effort into a PR campaign and a glorified signup page to capture users, which at the time were valued at about US$1,000 a pop. Just think, if we got a million of 'em, we'd be a billion-dollar company.
We had a great team of developers, although in our shoestring operation, most were moonlighting (which may indicate the level of faith they had in the CEO). They used the occasion to develop an elegant, JSP-based site, which was fairly cutting edge at the time. A free version of WebSphere was the app server of choice. To keep costs under control, a decision was made at some point in time to use Windows NT instead of UNIX servers. I helped set up the hardware in our little half-cage myself, and everything worked flawlessly.
Nothing was spared in the PR campaign to launch the site. A top-tier company wrote our press release and cut together some "b-roll" -- stock footage bounced to a satellite for download by local TV stations. It was a long shot, but we were hopeful that a couple of stations would pick up the feed and drive a few thousand signups.
On launch day, we flipped the switch and everything worked great. Traffic wasn't quite what we'd hoped, but then, we felt we had a good viral story to tell. Of course, the hard work of actually building out the site lay ahead of us, but first, an expensive launch party at a trendy restaurant was in order.
We'd all worked like dogs and the party started early. In those days, at startup events like these, people fell into two groups: those who couldn't believe their luck and those who felt that impending riches were a natural outgrowth of their own brilliance. Either way, eyes were shining all around the long banquet table.
Until the inevitable happened. Word spread quickly: The servers were down! Victorious smiles stretched unnaturally at their corners. Apparently, not one or two but 200 TV stations had picked up the b-roll. For our little half-cage of servers, it might as well have been a DoS attack, one that came every hour on the hour as station after station picked up the story across the country.
It took almost two days to recover -- and by then, the traffic was gone. Looking at the forensics, there was little doubt that we had lost a million signups, easy. It served us right. We had jumped the gun and launched before we had a site that did much at all, even though its value proposition was intriguing enough to attract a ridiculous number of users. Of course, it's possible that if we'd spent just a few thousand bucks more on server hardware, we could have been millionaires. But when you're so sure success is around the corner, there are many, many opportunities to screw up.