I had a college roommate who after taking business classes would come back to our house, pound a few beers, proclaim, “Dude, when in doubt, get horizontal”—and then proceed to pass out in front of the TV.
It’s probably not what he had in mind, but his maxim has great applicability to business. Horizontal firms like Intel that dominate one element of goods production are doing well. Vertically integrated firms like General Motors are not doing well.
The old IBM was vertical. It made everything in the computer but the squeak: the chips, the software, the box—even the sticker with the number to call for repair service.
The whole thing was copied from the GM model implemented by Alfred Sloan, a model that today is starting to unwind. Digital Equipment Corp. was in the same vertical mold, selling everything from alpha chips to the VMS operating system.
In 1987, when DEC’s stock was close to $200, Kenneth Olsen declared that DEC’s only constraint to growth was the number of sales representatives it had. He was right, but not in the way he meant. Instead of hiring 25,000 more sellers, he should have been firing them all—and getting horizontal.
AT&T was another vertically integrated monster. While working there I learned about “transfer pricing,” the fictitious accounting used inside a vertical company in order to assign profits to divisions. Result: Each division could show a profit, even when the end product was selling at a loss.
The chip division would charge $10 a chip—or whatever price included a profit, a huge R&D budget and massive overhead. The manufacturing division would buy 20 of these chips for $10 each, assemble them on a board with a power supply, put it in a box and sell it to the sales division for $400 as a modem. The markup covered lots of new robotic assembly toys the engineers wanted to play with and, of course, the annual management meeting in Laguna. The sales division would try to sell the modems at $600 but had trouble doing so when competitors were selling the same thing for $299.
In a vertical company each division could show a profit, even when the end product was selling at a loss.
The PC almost toppled these vertical manufacturers. In a hurry to compete with Apple Computer, IBM used Intel microprocessors, Microsoft software, Western Digital disk controllers and so on. Then Compaq came along with hardly more than an assembly line and a sales channel and beat IBM to market with a new 386 PC. Next Dell asked, “Who needs a sales channel? We’ll assemble these things and sell direct.”
What started with a few horizontal specialties has spread to at least a dozen, each with its own leader, producing the various peripherals, parts and software that go into the PC on your desk. With the exception of the somewhat artificial pricing in horizontal segments led by near-monopolies (Intel, Microsoft), pricing in the vertically disintegrated computer business is competitive and efficient. You don’t see the delusional transfer pricing that hides inefficiencies.
This horizontalizing of electronics is spreading. The data networking industry, for example, was dominated by vertically integrated players like Cisco. But lately Cisco has learned the lesson of the PC market, and it now buys components outside—for example, chips from MMC Networks for its Ethernet and asynchronous transfer mode switches.
Internet firms are getting horizontal at a blinding pace. Yahoo! started as a search engine and built a portal business around it. Recently it outsourced its search capabilities to Inktomi. The Inktomi service is run out of Exodus Communications, a specialist in so-called Web hosting. (Please note: My firm has stakes in both Inktomi and Exodus.) Exodus buys bulk data lines from telecom companies like Qwest Communications and Sprint and resells them to publishers of Web pages who locate their Web servers at Exodus’ hosting locations. In short, you have at least three service providers inserting themselves into the communications chain that connects a computer user to the page he wants to look at.
The early part of this century was marked by a powerful movement toward integration. The next century will begin with a movement the other way.