What Web Hosting Customers Want

By Doug Kaye

March 15, 2002 -- (WEB HOST INDUSTRY REVIEW) -- The Web hosting industry is less than a decade old, but already we've seen dramatic changes in the attitudes and priorities of customers. In these few years, the concerns of my clients have switched from, "Who's the best?" to "Who's the cheapest?" to "Who's going to be around for a few years?"

At the beginning of the Internet boom (circa 1995), Web hosting was driven by dot-com customers who had no choice but to outsource. These customers were pure startups, and although they had some of the industry's best young programmers, very few companies had the resources necessary to construct and operate even a small data center. The first Web hosting vendors were the same ISPs that provided Internet access, and they found low-hanging fruit when it came to harvesting customers from among the dot-com startups.

Shortage of qualified staff was another reason companies outsourced their hosting. If you were a system administrator in 1997, it was a lot more fun (and you got a lot more respect) running hundreds of Web sites for a hosting vendor, than having to cover 24x7 for one site all by yourself.

For five years, the dominant customer requirement was for performance, and everyone sought the Web hosting vendor that could deliver content the fastest. The Internet's infrastructure wasn't keeping pace with demand, and most customers looked to their Web hosting services to provide a competitive advantage through faster content delivery.

Driven by this passion for performance, we saw the emergence of content delivery networks (CDNs) such as Akamai and Sandpiper. They bypassed the slow and unreliable core of the Internet, and delivered content from as close to visitors as possible. It made sense back then - performance meant everything, and at nearly any cost.

But by 2000 things were changing, and changing faster than anyone realized. The millions of dollars that had been poured into building infrastructure were taking hold. Before we knew it, we found ourselves with an excess of long-haul fiber capacity and (except in a few locations) more data center space than could be filled. Then, on April 14, 2000, the dot-com stock-price bubble burst, and the pyramid collapsed. Suddenly, those customers weren't looking for performance; they couldn't even pay their bills. Contraction, rather than expansion, was the word from the boardrooms.

Hardest hit were the CDNs, for they had bet their farms on high-cost high-performance delivery. Weakened demand, combined with an artificially low cost of traditional delivery (due to the long-haul fiber glut) meant not only that CDNs could no longer charge a premium for their services, but that they actually had to sell at below their costs. And all at a time when they, just like their customers, were cutoff from sources of capital.

Eventually, those vendors that had over-built and over-committed their financing began to fall on hard times as well. This was felt hardest by the vendors that primarily offered debt-intensive colocation, such as Exodus. Once the word got out that Exodus was in trouble (last summer), customers became nervous. And when the events of September 11 occurred, customers' concerns changed suddenly once again.

It was a triple whammy: an already weakened hosting industry, terrorist attacks, and the widespread downturn in the economy. Customers' attentions quickly turned to the economic stability of their Web hosting vendors. When Exodus filed for bankruptcy protection in late September 2001, the change in my clients' concerns was dramatic and nearly universal.

While Exodus was tumbling, customers were having problems of their own. Most noticeably, IT budgets for calendar year 2002 were slashed. Fears of recession were so great that CIOs and CTOs were not only being asked to cancel new projects, but substantially curtail existing systems and costs. Some CIOs were even charged with converting cost centers into revenue-generating opportunities.

So what are clients asking for today? It's fairly consistent. In order of priority: vendor stability, cost, performance (service) and security.

Why isn't security higher on the list? Unlike the other criteria, most clients are willing to take moderate risks when it comes to the security of their sites. They believe they can mitigate risk through good practices, vendor assurances and insurance. It's somewhat easier to measure stability, cost and performance than it is to measure security, so security is less often a criterion during vendor evaluation. (One vendor, Pilot Networks, emphasized security, and subsequently went out of business.)

In the future, the swing of the pendulum will dampen substantially. If the economy continues to improve during the second and third quarters of 2002, IT budgets for calendar 2003 will include at least moderate increases. 2004 looks great. Perhaps in 2005, enough of the long-haul fiber will be lit to end the deep discounting of bandwidth. Once the unstable vendors are weeded out and the supply/demand ratio stabilizes, I expect the top customer criteria will become a healthy balance of price and performance.