Rackspace is Coming to Dinner!

TH: In your latest press release, Rackspace is referred to as being a “Pioneer of Profitability.” Can you perhaps detail the innovations that led up to this label?

GW: First of all, we focused very narrowly on managed hosting, which means we’ve gotten very good at providing managed hosting: we don’t offer shared, we don’t do design, we don’t do programming or integration. This focus on managed and dedicated hosting is where it all began. Secondly, Rackspace is focused on providing fanatical customer support. These days, when data center or colo space is virtually free, servers are virtually free, and bandwidth is virtually free (by which I mean that all of these elements have dropped in price dramatically over the last couple of years), the thing that people want most is the support for these elements - the part that is still not free, and continues to make deploying on the Web extremely complicated. While the price of servers, colo space, and bandwidth has dropped by more than 75 percent over the last couple of years, maintaining the Web server and the OS is still just as complicated as ever, which is why we decided to focus on support first and foremost, to better help customers pull all of these elements together.

TH: What makes the issue of support such a complicated one?

GW: Let’s just say that a lot of companies operate with the focus that “It’s all about the network.” For example, WorldCom’s current campaign emphasizes how they are “one network,” Dell focuses on their servers when marketing their services, and Exodus goes on about their colo center. I don’t think that it’s about any of these things, it’s about integrating these elements and having someone maintain them for you reliably.

TH: So basically, Rackspace’s definition of “managed hosting” has nothing to do with a client’s hardware or software, but is about the support behind all aspects of the solution?

GW: Exactly. It’s the support for those things that makes a managed host, it’s about having someone to call on to help you complete your deployment, rather than leaving it to you to piece together yourself. Managed hosting should support the total solution, not just isolated pieces of hardware and software.

TH: I notice that it’s stated very succinctly on your homepage that Rackspace does not offer “virtual hosting, co-location, Web development, systems integration, or Internet dial-up” - you’re very clear about that. Does this commitment to managed hosting at the exclusion of all else cause any problems - especially in light of the fact that over 46 percent of potential customers in the US are looking for these other solutions? Is it a different market you’re targeting?

GW: Absolutely. In order to get good at anything requires sacrifices to be made, which is what I mean when I talk about focus, which can only be achieved by eliminating everything else that is not related to your goal. You can do anything you want, but you can’t do everything that you want, which many hosting providers seem not to realize. I think that by limiting our focus to a relatively narrow range of activities or specialties, we’ve gotten really good at it, while the people who didn’t focus are still struggling along.

TH: I can’t argue with that! Is there any room in this focus for the virtual private server? The VPS seems to be sort of the “new kid on the block” for lower price point, high functionality solutions, and there’s been a lot of press surrounding companies like Ensim and Sphera. Does Rackspace have any plans to gain traction in this marketplace - perhaps to target smaller organizations who can’t afford a full-blown managed or dedicated solution?

GW: I think that the answer is definitely “yes.” We think that it’s still only a “niche” of the managed hosting landscape, and may only make up 20 percent of the spectrum. However, while the VPS is a small niche, it’s an important one, because once again it de-emphasizes the technology aspects - making support more valuable and hardware less of a critical issue. It’s not the “shared” aspects of virtual hosting that are becoming unpopular. Many people are willing to share a server with 20 other clients, but only if they have full access.

TH: So basically, hosting providers are finding it increasingly difficult to use technology as a selling point for their solutions.

GW: Yes. When Rackspace first started, there were many people who did not think that an emphasis on support was the right one for us, because it’s too easily copied. To be blunt: while special software can be copyrighted, etc., there’s nothing proprietary about people, which is the very essence of effective support. However, even this was eventually proven wrong, as almost every software innovation is copied - look at Verio’s iServer, which is essentially Ensim, and the iServer product has been around since 1996.

TH: Really?

GW: Absolutely. Virtual Private Servers in concept is not that new, it’s just becoming more popular. Ensim is a commercial version of it, but it’s nothing new. What was proprietary for the iServer in 1996 is now essentially available off-the-shelf for all hosting companies - so you can see how difficult it is to stay ahead with a total focus on technology.

TH: In regards to this commitment, as a Rackspace customer, what would I enjoy in terms of support that would set the company apart from the competition?

GW: To begin, I’d like to also point out that in the beginning, what distinguished us early on from the competition was that when we first entered the dedicated server industry, it was really about leasing a server on a network, and we thought that while that had merit (and there’s clearly a portion of the market that focuses on that today), we also came to the conclusion that the true client-value comes from managed hosting - it’s about people, not about leasing the hardware. Any company that charges for support is not in the managed hosting business. For instance, if you went to get your car washed, but had to pay for the water, it wouldn’t really be a car wash would it?

TH: That’s a good analogy, especially since the majority of consumer complaints surrounding hosting have to do with issues of support, and the cost of that support: “Why should I pay $75 per hour when the host is responsible for the uptime of my server?” - complaints like that.

GW: If that’s the case, where you’re renting the server and support is extra, than that’s a different business altogether - support is what you’re buying. We’ve never charged for support, and we decided early on that support should be unlimited. To return to the car wash analogy, while we won’t repair your car, we will wash it as many times as is necessary to get it clean - we’ll provide you with an unlimited amount of support until your problem is resolved. Obviously, we will only provide this within certain boundaries - there’s still a lot of confusion between the nature of dedicated servers and managed hosting, because there are a lot of people who just want to rent a server. Still, we think the bigger market are the customers who want Rackspace to take accountability for their server.

TH: I can’t deny that there’s been a large shift in favor towards managed hosting solutions over the last year - do you think that it’s support that is driving this?

GW: Unequivocally. The hardware space is fairly level right now. Every server today is at least adequate, most bandwidth providers are adequate, colo space is fairly similar across the board. What I’m getting at is any hardware solution you choose will get the job done, so relying on what’s in your rack as a selling point simply won’t work. Two years ago, people would brag about having back-up power, fire suppression systems, high security in their data center etc. These days, try to find me a host that doesn’t - it’s a given, anyone who doesn’t have these things isn’t even in the game. You won’t get far trying to sell cars just because your model has wheels. Who cares? They all do!

TH: Still, the homepages of many Web hosts still outline their collection of “gear” pretty prominently, so some still consider it an important selling point.

GW: And more power to them! WorldCom’s current campaign is focused on the message that it’s better to host with the company that “owns” the network. And if you’re WorldCom, it makes sense to talk about that, they do own the network. However, in the end, while it’s an interesting discussion, it’s the wrong one. The first generation of the Web was driven by technical people, and the prevailing attitude was one of “Isn’t this cool? Look at everything we can do!” I think about a year ago, the businesspeople started walking down the hall and saying: “Look. We’ve been spending a lot of money on this stuff. It’s not longer an issue of ‘cool,’ we have to make this work for us.” Eventually, the tech side of the house had to jettison all those operational functions that are not necessary for the internal IT person to control. Internal IT’s job is to make the site run better, not to perform the hosting, which is where our industry comes in.

TH: Are there any cost advantages to this non-technical focus? Does not having to “keep up with the Jones’s,” by constantly upgrading your technology and capacity translate into savings when running your own data center?

GW: What do you mean exactly?

TH: I mean the massive expansion race of a few years ago, when the perception of what makes a successful host was totally based on the philosophy of upgrade, build, expand, and repeat.

GW: The technological fundamentals are no different today than they were years ago. Take the IP network for example - the Cisco GSR routers that sit in our network are essentially the same ones you would buy today, so I think that such innovations (except for the long haul network where DWDM has come on the scene), have had little impact - routing at the data center level has not made existing technology obsolete. I think that a lot of the dot com boom had to do with getting every data center stocked with these routers, back-up power, and the like, the first time through.

Nowadays we’re not continuing to add and add and add to these basic facilities - because they already exist. I think the demand for Cisco etc. has gone down. There’s a huge amount of infrastructure - where there was tens of thousands of miles of fiber laid last year, there will be much less this year, because the world has been criss-crossed enough times. While everyone is still upgrading their network in terms of its capacity, it’s not because new technology is forcing them to.

TH: So the trend now is how to use the technology you have rather than expanding needlessly.

GW: Exodus is a good example of this (even though their problems reached far beyond this) , as they built far more capacity than they needed, and could sell.

TH: To be fair, I think Exodus simply counted on a growth that never happened.

GW: True. In Austin Texas for example, until Exodus came on the scene in 1999 and opened their first data center, there was no true co-location center in this region. So when Exodus opened its massive 75,000 square feet data center, it filled up instantly. So then they reasoned that they needed a second one. It’s always dangerous to try to extrapolate the utilization of these hi tech assets, especially when this extrapolation is based on years of pent up demand, rather than continuing demand. So when they opened their second data center, it didn’t fill up as fast as the first one did - because the pent up demand had already been satisfied. But Exodus continued to build data centers long after they saw this pent up demand satisfied. Obviously, the dot com bust contributed to this as well - but it’s still puzzling.

TH: Back to managed hosting, which is your focus. Do you think that managed hosting has the potential to in essence, obliterate all the other genres of hosting - like shared, dedicated, colo etc.? Quite simply, will managed hosting eventually be “the way it’s done” for the entire hosting landscape?

GW: Colo will always be around, but it will become a niche product. It will be used only for those who have the deep pockets and deep in house expertise. Think about it like payroll. It seems so logical to outsource payroll - it’s not strategic to your business, but it’s complicated. Likewise, managed hosting makes sense for all businesses where hosting is not strategic, but it’s complicated - which applies to pretty much everyone. It all depends on the type of business in question.To use another analogy. Compare a newspaper with a magazine. For a newspaper, printing the paper on a daily basis is strategic for them - getting the paper printed and delivered that day is very important to them. A magazine on the other hand can outsource their printing much more easily, as the timeframe they need it in is much more tolerant. So I think that co-location will still exist for those businesses where hosting is strategic to them - such as eBay, eTrade, and others - whereas almost everyone else will go to managed hosting. At the same time, shared hosting will also continue, as it’s still the best solution for 70 percent of the small Web sites out there, and at the high end you’ll still have major clusters. Still, even these people will be comparing, and asking themselves if they should be switching to co-location or managed hosting. Rackspace fits into the space below that, where most of our clients do not have the expertise sitting around in house waiting to be deployed.

TH: It seems that everyone’s trying to get in on this trend, and everyone seems to be scrambling to put together some sort of managed hosting offering. Even the Telcos seems to be edging into this territory. If support is the primary function and benefit of managed hosting, is it even possible for a large Telco to provide managed hosting effectively? Obviously Rackspace is competing well, but is this a concern for the future?

GW: It has never been done in the history of man. There has never been a Telco who has ever been able to run a high-support business well.

TH: Well, their revenue is based on volume-sales.