GAC Berlin 2000: IT Links for Boundaryless Companies

By Kenneth Liss, Editor, HBS Working Knowledge

David Upton's Alumni Conference report on the state of information technology for business-to-business began with a lesson in technology history because, said the HBS professor, "understanding doesn't come easy without an understanding of the Internet and where the Internet came from."

That includes, especially, an understanding of the open standards process through which the Internet was built, with engineers stacking one set of open protocols on top of another from the physical—how you plug the wires together—through successively higher levels of data, network, transport, session, presentation and application activity.

"The levels at which computers communicate have been formalized," said Upton. "This is important as we begin to understand what's happening on the 'Net as we go forward."

As this growth of standards up the protocol stacks continues, he said, it's time for a new kind of standards to emerge: standards not just for how computers communicate, but for how business processes work.

Even an advanced technical protocol like XML, said Upton, "with all its cleverness, doesn't know anything about business semantics."

"This is where the engineers leave off," he told the audience of business leaders. "This is our territory.

"Think about this for your industry," he continued. "Who defines these standards for your industry? Should it be you? Is it your firm? Another existing firm? A dot.com? What happens in the vacuum?"

These are important question to ask, said Upton, amid the emergence of extranets, the fastest growing segment on the 'Net and potentially the most important in terms of transforming business.

"We're developing new business models as we go along," he said. "We're seeing real innovation in business, not in products but in what business does."

Economic Facts of Life

There are three key characteristics of information technology markets—three economic facts of life, said Upton, that can help show how that growth will play out.

The first is switching costs: "There are huge switching costs in information technology. If you're selling stuff, you can lock in the buyer around your proprietary standards. But the buyer has to beware they're not locking themselves into a closed standard.

"When you invest in information technology, it's not so much the investment in IT as the investment in standards."

The second characteristic, said Upton, is network effects. "Think about the value of owning the first telephone. The value of owning the first telephone is zero. It's only valuable when someone else gets one. As more and more people get phones that can speak with your phone, your phone becomes more valuable.

"We see that very strongly in information technology. That effect is very important not only in terms of the hardware protocols, but also in terms of virtual networks, networks of users.

"We don't like to think of information technology as being subject to fashion, but the same thing happens, the network effect. This is important when we start talking about B2B hubs."

The third characteristic is differentiated products and prices. Most of the costs of producing pieces of information are fixed costs, paid up front. "The price of a piece of information is vanishingly small, so we have to play tricks to make sure it doesn't disappear to zero," said Upton.

"That doesn't work unless I make some artificial distinction between things. That's why, for example, we have Adobe PhotoShop, for which you pay full price, and Adobe Limited Edition. It actually costs more to make that, because you've got to change the code to strip things out. But at least you've segmented the markets."

Market Makers and Inter-Business Processes

The two primary sources of growth in B2B uses of the Internet, said Upton, are market making and inter-business processes.

Market makers, he said, go by many names: hubs, portals, infomediaries. "There are just so many words for it because everyone wants to say ours is different.

"But the key thing is if you're market making you're trying to aggregate buyers and sellers. Often, they're people who would not have met if not for the Internet. The advantages of this are you can reduce transaction costs if you're buying and selling, particularly if you're buying."

This is where the network effect comes into play, said Upton. "The value of the markets grow as the number of people who are on the network grow, same as we saw on the phones. As they get bigger, many more people will come to them because that's where you go.

"Now we see that the Internet land grab is really important when it comes to B2B e-commerce."

Vertical Hubs and Functional Hubs

The two general types of B2B hubs, said Upton, are vertical hubs, focused on a single industry segment, and functional hubs, focused on a common process or requirement, such as maintenance or human resources, across industries.

There are several key things that vertical hubs need to succeed, he said: fragmented buyers and sellers ("If people have other ways to meet, it won't work."); an inefficient existing supply chain; and, most important, knowledge of the industry domain. Functional hubs, on the other hand, can depend to a great degree on the standardization of processes and procedures and on complementing process automation with deep content knowledge.

Upton also outlined three key challenges for B2B hub builders. First is the domain expertise. ("Important in making the thing work, but also in building a barrier to others.") Second is acquiring participants to get big fast. ("What we're seeing in this arena is very traditional industrial marketing.") The third is retention. ("The competition is only a click or two away. What you want to do is increase the participant switching costs.")

On the inter-business process side, said Upton, the key is the detail of what happens in a business transaction and the task of defining standards for B2B e-commerce. "If we accept the argument that we've got strong network effects in this kind of B2B commerce, that standards are starting to be set, what do we do with this?"

He noted three different models for developing these standards: an industry leader setting the standard, to their own benefit but for the good of the industry, as well; a more collaborative, but very slow, industry association model; and a middle ground of a business consortium.

Another Way of Doing Things

Upton concluded his presentation with a look at two contrasting approaches to information technology development.

One approach, what he referred to as the periodic project, involves a lot of time and a lot of planning, with value delivery going underground. "By the time the project happens, it's late, it's over budget, and it just about fails to automate your business the way it was three years ago."

The vast majority of projects happen this way, said Upton, but there is another way of doing things, what he called the path-based approach.

"The other way is to do a lot of projects, to try things out, to see what works. Delivery of value is ongoing. The default in this model is action. The idea is you do custom assembly, but you put a premium on the ability of those things to fit together like Lego bricks.

"The primary concern in the periodical project is control. With the path-based approach you've got to have your eye out for how easily is this thing going to integrate with things I haven't even thought of yet.

"A key question to ask yourself as you are building systems for this new, interconnected world is not 'Will it do the job?," but 'How will it change? How will it improve in the future?'"

The role of IT in this model is essential, said Upton, but the primary managerial effort, which starts at the top of the company, is in managing the standards that you allow to be in play in your firm.

"Bear this in mind as you encourage your IT people to grasp this model of experimenting with both hands as we move forward."