In every nascent product category, there is a debate about the merits of building applications in-house compared to customising a set of packaged applications licensed from a vendor. That debate is now starting to be taken up in the Internet-commerce arena as IT organisations scramble to gain control of this rapidly growing area. BroadVision CEO Pehong Chen expects the debate to be settled firmly in favour of packaged applications during the next phase of the emerging digital economy. In a recent conversation with IDG's Michael Vizard, Chen explained why.
IDG: The vast majority of Web sites today have opted to build their own applications for I-commerce. Why do you think this will change?
Chen: There are three stages in building an enterprise application. You can start with a tool like Java. You can start with point solutions. And you can start with a more complete, end-to-end package solution.
There are tools that would probably give you some 15 to 20 per cent of the capability you need out of the box, and you write all the other stuff. BroadVision basically takes a more complete end-to-end approach in that the customisation work is just about 10 to 20 per cent.
Typically, a very complex and comprehensive site can be launched, including managing and repurposing all the content, within three months using BroadVision.
If you want to start with a tool, you're welcome to do that, but the industry has been following a trend which tends to favour the benefits of cost of ownership and time to market.
What impact are companies like Amazon.com having on you and your customers?
I think one of the things that's really good is that although Amazon.com is not our customer, they did build everything from scratch.
The best thing that ever happened to us is their level of success. It is getting everybody to take notice and everybody is saying: "I need something like that, but I don't have the time to build that kind of thing."
There's been a fair amount of controversy about the merits of one-to-one marketing on the Web using products such as yours. What's the value proposition for this approach?
The right way to address it is to put it in the context of a business framework. Let's face it, the Web is a channel for doing business.
The nature of this channel is something that you obviously need to understand before you can really do anything with it.
This channel is unique compared to other prior channels in that it's interactive, it is multi-dimensional.
If that is the environment, you're heavily relying on the user to do the task rather than somebody on your staff. Then it becomes a challenge in that it has to be easy, compelling, and palatable for these people to be able to do it. Most people can't deal with stuff that requires too much time to learn. It's not a question of intelligence, it's a question of time.
What we say is that if I know who you are, I'll give you what you need and you can get something done with me in two or three clicks. It shouldn't take 20 or 30 clicks and 20 or 30 pages of browsing. It is very important for me to help you by understanding your interests. So it's absolutely essential that you do one-to-one marketing and personalisation of content.
Why is this important?
The Web is the biggest price-leveller. Unless you do really successful one-to-one marketing, it's very hard to compete unless you can drive the price to the lowest possible point.
How many of the Web sites out there today would you say are passive, as opposed to active, Web sites - and when will this state of affairs change?
I would say almost 99.9 percent are very passive, first-generation sites. There are probably a couple of possibilities for a change. One is, of course, cost of ownership. The other is that the traditional way of laying out Web sites is as a string of pages. If your site is successful, you're going to find yourself having way too many pages to manage.
How does BroadVision differ from what Netscape and Microsoft provide in the personalisation area?
Microsoft and platform vendors like Microsoft and Netscape offer generic Web servers. These are more or less application servers that keep track of state and provide integration into databases. They really don't incorporate specific application logic for e-commerce, financial services, or knowledge management.
Certainly they don't do any personalisation or all the content management capabilities we do.
Given the number of application servers on the market, does it make sense for you to have your own application server? Or will BroadVision ultimately just layer its software on top of application servers from other vendors?
That's a very likely scenario. However, in an enterprise environment, the application server is very critical in terms of its scalability and caching. A generic application server doesn't really have the context about what the application is [that is running on that server]. So it's very difficult for it to do effective caching. We're talking about serving millions and millions of pages a day, so the BroadVision server is highly specialised.
Why is it so difficult today to integrate I-commerce applications with the overall enterprise?
Because the enterprise environment is a mess - it's very complex. Even BroadVision, a brand-new company, has six or seven different back-end systems already. You can imagine if a company has been there for 100 years how complicated it would be.
Most people equate the Internet-commerce trend with the invention of the light bulb or the discovery of fire. This is a very, very big thing. The complexity is humongous. You're talking about trying to get customers to do self-service, and there are just lots of issues there, especially in terms of supply chains.
So at the end of the day, how much effort should it take for someone to find something on a Web site?
Three clicks or 30 seconds. Searches should always be in the background, so it's always there. But that's the problem today. There's no contextual information there, and it's just making searches very, very hard and very unfriendly.