One on One

Xuma's CTO has helped build a number of e-business sites. He's got some sage advice for anyone involved in similar endeavors.

As Xumas chairman, CTO and co-founder, Jamie Lerner is at the leading edge of the volatile e-business integration space.

Xuma, which started as an e-commerce ASP, found itself having to go into reinvention mode last year with the fall of the dot-coms because those types of businesses accounted for 75 percent of its revenue. Since last year, Xuma has recast itself as both a product vendor and an MSP.

On the product side, the company sells software stacks, the very thing that it used to build for end-user customers. These days, its customers are solutions providers that resell a suite or components of the stacks without having to provide the coding glue, something that Xuma has already done. Using Xumas rapid-delivery model, equipment can be sourced, installed, brought up, and have bandwidth allocated and applications installed in a matter of weeks. Xuma uses preintegrated, preconfigured combinations of hardware- and software-based solution stacks to accomplish this rapid delivery of infrastructure.

Lerner started Xuma in 1998 with fellow ex-Andersen consultant Joe Cha. Since that time, the company has launched more than 70 Web-based businesses for various clients, including Maxim Integrated Partners, Riverdeep, ZoZa, and ConvaTec, part of Bristol-Myers Squibb. Hes seen the good times and, more recently, the bad. What has he learned? Plenty. We caught up with him for a core dump.

SP: Whats Xumas claim to fame?

Today, why would anyone go out, look at offerings from IBM and Oracle, and go back to their boss and announce, "We need to write our own database." Can you think of any reason to write your own operating system when you can buy Solaris? You can use the same logic for e-business infrastructure. Seven years ago, you had to build everything yourself. Today, you can go out and buy it, prepackaged. You can buy hosting infrastructure, application server infrastructure. Its a simple matter to find managed service providers and application service providers to deliver the computing environment you need.

SP: Whats happening these days?

People are stepping back now. We get responses that range from, "Well revisit this in six months," to, "We dont know if the Internet is going to be around in five years." The mantra used to be, "Get on the Internet or youre dead." Now all the brick-and-mortar companies are looking around and saying, "Hey, were still here."

SP: How do you convince them that your solution is the right one?

My two most important sales tools are a set of return on investment (ROI) calculations that can be believed and my CFO.

SP: Your CFO?

These days, people are looking at my long-term viability and stability just as carefully as they are looking at the technical elegance of the Xuma solution. Were finding that stability wins out over technology. You may have a solution thats higher performance, with more features. But your prospective customer is looking to make sure that youll be around. State-of-the-art technology doesnt hold a candle to stability.

SP: What are some of your best practices when architecting an e-business solution?

Build a solution with maintenance in mind. The cost of maintaining these systems can be astronomical. Weve seen sites that spend upwards of a million dollars in the first year exclusive of development costs—and these expenses are just for making fixes to get the original functionality implemented.

Think about the future—and upgrades. About 70 percent of your vendor partners arent going to be around in three years. So if youre using their products or services as a component of your solution, youll have to worry about replacing that component.

When youre building, youve got to think about scale. How do you plan for large traffic increases? Sure, you can throw more servers and more bandwidth at the problem, and that will probably work, most of the time. But how does your site scale to a requirement for a large number of feature or functionality improvements? Suddenly, your site may have to start interacting with legacy systems, or a Java process, or Perl, or add much more e-mail capability. Youve got to realize that Internet-based systems are a fraction of the computing infrastructure already deployed in an enterprise. And if you need to access the data in these other systems, youre going to have to go to them, because theyre not going to come to you.

If the system was never designed to deal with these (as well as new technologies that are always being introduced), youll be spending most of your time ripping out, rewriting and patching.

Youve got to give your customer options with regard to operating in-house or outsourcing. Some customers want to write the check and have you do it; others want more hands-on control. Everyone wants option cards. We try to give customers the ability to change their minds as business needs dictate. If they decide they want to move operations in-house, we give them that option.

SP: Do you see more people bringing their operations in-house?

I think that going forward, customers will be less frightened of hosting themselves. Theyll know what it entails and have the staffing that is able to perform the necessary tasks to do so. So yeah, I see that happening.

SP: Dont people worry about reliability?

When considering reliability, you need to be realistic. Does the site really need to have .99999 reliability? Some sites do, but others dont, so the assumption may be flawed. Its tough to say as an MSP, but its reality. Youve got to tailor the solution. Not everyones running Fort Knox. Some CIOs, when they hear about the extra expense involved in getting to that level of reliability, will accept that being up 97 percent of the time is satisfactory. They say, "When the sites down, well just put up a page that says the sites temporarily unavailable." They leave it at that, and save hundreds of thousands of dollars. Its purely a business decision.

SP: What do you see happening in the future?

I dont see the rate of innovation slowing down, but I think the number of companies involved with development of products in the space we play in, will. At present, the rewards for the entrepreneur arent there.

Still, we have a lot of work to do. A lot of attention needs to be paid to application servers, content servers, routing, monitoring and security. If you look at an e-business system today, all of the pieces are there, but theyre all point solutions. For example, eCRM is great for call centers. The applications from eGain (for e-mail), Vantive and Scopus (for voice) are all wonderful in their own right, but they deal with a specific line of business—the call center.

But what about the customer who abandoned the shopping cart halfway through the configuration, called the 800 number and left a message, and then sent an e-mail three days ago? How do you track them? Nobodys integrating the whole ball of wax; at present, nobody can track all of these interactions.

SP: Why is this important?

Its very difficult to ascertain how valuable some features of a customer-facing Web site are. Systems exposed to the customer and the outside world have more of a need to be viewed as "competitive" by many within the organization, but it has to be tempered with common sense. While the customer can apply more pressure to adopt new features and has a relatively strong voice regarding the feature set youre offering, youve got to ask how valuable all of these new features are, and what the return on these features is. Do we need a call center? Do we need Web-based self-service? Will most people be happy with an 800 number?

The technology exists to do this sort of basic correlation and multidimensional analysis. We can get the data from the system, throw it into an OLAP system and start cranking out reports.

SP: What are you doing on the wireless front?

Were doing a little here, but we dont see anywhere near the hype that we saw a year ago for wireless. People just dont want to view a video through their Web-enabled cell phone. Its nice to know you can, but people arent doing it. Dont get me wrong here—wireless has its place, and its just awesome for certain business processes. I think its great for approvals, which makes it ideal for workflow. But it has to be simple, with a very small amount of data—just a snippet—presented to the user.

SP: Is security a big issue for you and your customers?

Absolutely. Our systems are being hacked away at almost continuously.

SP: Do you worry about the competition?

Sure, I worry. But we play in a niche with very few competitors, and its not as if were competing against Microsoft, Sun or IBM. In fact, were partnering with them and gluing together their technologies. In many ways, were agnostic toward their products.

SP: Whats next up for Xuma? Any acquisitions?

Theres nothing on the plate right now. Weve found that most technology that were interested in is more expensive to acquire than it is to go out and build it ourselves. Were continuing to invest in monitoring and middleware reporting tools, and to connecting disparate systems. Its not sexy or fun, but its necessary.