Someone once produced a list showing that every aphorism could be matched with an equally venerable saying that contradicted it. The early bird gets the worm—but haste makes waste.

A modern game of point/counterpoint is going on among those who think about how information technology is changing business. Some argue that companies need to be "first movers" to take advantage of the opportunities being created. Others counter that it's better to let the pioneers make expensive mistakes, then be a "fast follower" that gets the idea right and captures the market.

To sort out the issues, Context turned to two experts who have long debated the issue at Harvard Business School. Warren McFarlan, a professor there, argues that "fast following" is generally better. John Sviokla, a consultant and speaker who was a professor at Harvard until recently, says moving first is usually right. Together, they hash out some illuminating principles for how to figure out which approach works better in which situations.

JOHN SVIOKLA: I'm basically a "first mover" advocate. Digital technologies and the dynamics of electronic commerce favor companies that are halfway around the track before everyone else knows there's a race. First movers can remake the value chain of just about any type of business you can think of. Much lower price, much higher value provided. Seemingly overnight.

In my view, a first mover is not simply a company that makes a big play early. A genuine first-mover play manipulates computerized data flows in ways that traditional competitors haven't learned yet. In this "virtual" realm, some funny, anti-gravity sorts of things can happen. You start seeing enormous economies of scale for companies with strong brands. And the marginal costs of distribution can fall to zero, making it advantageous to operate in more markets from the beginning, rather than fewer.

I know that, traditionally, start-ups haven't been as well-financed as established companies. But look at the market capitalizations of Yahoo or some of these other Internet companies. The capital markets are giving machine guns to kids. A first mover can actually use those valuations to raise money and acquire customers and traffic faster than hulking competitors that have only ordinary price/earnings multiples.

WARREN MCFARLAN: The Goliaths are still awfully strong, though. I'm probably more of a believer in hanging back than you are, John. That's largely for three reasons. One, while the "digital marketspace" possibilities are exciting, the earthbound side of business will continue to predominate in most industries for a while yet. Second, in contrast to you, my focus is on the size of the upfront investment rather than on the marginal costs and marginal revenues of the "next unit" sold. After all, that last transaction takes place months or eventually years after the initial outlays have piled on debt. Third, I worry a lot about the market and the financial power of entrenched competitors. If a start-up goes after a really large market, the big players are going to work like hell to pound it down.

SVIOKLA: Right. An incumbent can't afford to let an entrant's second bite of the apple taste as sweet as the first.

MCFARLAN: So how do you know when to make the first move? I would say, when four conditions are present simultaneously. No. 1, when your bet requires a big investment, in terms of both dollars and time to market. No. 2, when it will provide customers with new and substantial value. No. 3, when the cost for customers to switch away from you will be high. And, No. 4, when you can keep what you're doing confidential until roll-out.

SVIOKLA: I would add two more conditions—when you've used digital technology to establish a distinctly lower price for your category, and when your system configuration is modular. By "modular," I mean that your information-systems architecture is open enough that most of your buyers or partners can plug right in with their systems and work with you. That's how you'll reap the "Metcalfe effects" [creating a network that, like the Internet, seems to increase in value exponentially as new users join] and attain economies of scale. With modularity, cost per additional customer should drop to near zero.

By the way, modularity is more than interactivity. An airline-reservation system, for instance, is interactive, not modular. Airlines can't let outsiders mess with their planes or routing systems. But something like vendor-managed inventory is modular. A manufacturer lets a retailer make entries in its warehouse records, often by as mindless a process as bar code swipes at a checkout counter—which is kind of frightening, but essential to getting the huge gains.

Even stellar first movers probably don't exhibit all the characteristics we've laid out. Some haven't called for a big investment in the usual sense. Microsoft, at its inception, didn't require much money. And others haven't been digital. But, in general, the more ways you can cement a first-mover position, the better.

MCFARLAN: Microsoft is the classic example. It shows how a first mover can literally become so dominant that it eventually embodies its industry. Cisco is another, by virtue of its superior ability to create new operating economies well before competitors could react. Dell has just reached a similar plane of industry domination. Dell carved the heart out of IBM's market share in the PC business using a new, very different distribution concept: over-the-phone direct order.

SVIOKLA: Great examples, but all from young, technology-based businesses.

MCFARLAN: OK. I'll give you personal examples of spectacular first-mover success from two of the oldest businesses around, agriculture and lending. I'm on the boards of both companies, Pioneer Hybrid International and Providian Financial.

A new seed takes about eight years of research and testing to develop. But if you can provide farmers with a significant per-acre advantage on their crop yields, you win their business. Back in 1973, Pioneer came out with a hybrid that gave farmers a 14-bushel-per-acre advantage. At the time, DeKalb had 26% of the market, and Pioneer had 23%. When the dust cleared, Pioneer had more than 40% and DeKalb had about 10%. And that is about where DeKalb has stayed ever since. Ultimately, DeKalb was forced to sell out to Monsanto.

At Providian, our dream was to understand the potential of the unbanked market, the 30 million people who got paychecks but didn't have so much as a checking account. Providian had the courage to offer these people credit. We drove right through that territory where bankers feared to tread.

SVIOKLA: Was extending credit to the financially disenfranchised a hard sell to the board?

MCFARLAN: Not really. After all, it was a large and competition-free niche. Once we spotted it, we did some modeling and built in an appropriate loss ratio. Then we ran a pilot, and then another pilot. By the time we had a working system, we were making money. That made the profit model clear to others, but it was too late for them to knock us out. We had stolen almost a year and a half lead.

SVIOKLA: You've been doing my work for me, giving examples of first movers that made it to profitability. Tell me about somebody who moved early and got hammered for it.

MCFARLAN: Chemical Bank's venture in the early 1980s, Pronto Home Banking, is notorious as a great concept undercut by wretched timing.

The vision was that people with their home PCs were going to, in their enthusiasm, call into their Chemical bank account after work every day and move money back and forth. The trouble was, Pronto did not provide customers enough value to get them to switch from their usual ways of banking. People really didn't, and still don't, want to know what's in their bank account that often. Customers resisted, the red ink just flowed, and Chemical took a $100 million loss on the project.

SVIOKLA: That will get your attention. As I remember from using a similar system at BayBanks, the user interface was brutal. You were essentially a dumb terminal attached to a mainframe system. Response times were terrible, and the statements fed you a lot of jargon. Then the system would boot you out if you had a problem. And you were paying extra money for this!

MCFARLAN: Yes. Chemical's mistake was getting to market before the technology was friendly enough. That mistake is well worth remembering because the standard of user friendliness keeps going up and up.

SVIOKLA: I would also stress that they made a fundamental mistake in pricing. Chemical had an old mindset and viewed Pronto as a service that provided increased value. So it charged something like $50 a month. Really, Pronto was a new channel of distribution...

MCFARLAN: ...and a new way for banks to manage a relationship with the customer. Chemical's approach was clumsy.

Of course, this is where management decision-making gets tricky and interesting. Chemical had to guess whether the Pronto value proposition was powerful enough to stampede customers in its direction and whether it had built high enough switching costs to keep people from stampeding back the other way. In industrial settings, a purchasing agent is never really stampeded. But consumers juggle lots of complexities in their lives and look to scrutinize only what they can manage.

I'm a perfect example. I moved to MCI to get in on its frequent-flier mileage partnership. AT&T's sales program must have come back at me 50 times, but by then MCI had me. Switching costs are what protect you from a deep-pocketed fast follower buying its way back into the game.

SVIOKLA: Who's a great fast follower then?

MCFARLAN: Microsoft, in its mature phase. It used the cash it gained early against Lotus to do exactly what AT&T refused to do against MCI—slash price drastically. Lotus depended on its 1-2-3 earnings, but Microsoft didn't need any cash flow from Excel. It used its superior financial strength to offer Excel as a loss leader, smashing Lotus's marginal costs from underneath.

SVIOKLA: I would argue that, increasingly, big players can't afford to assume a fast-follower strategy will save them. W.W. Grainger is a huge distributor of motors and equipment. It has a tremendous physical business at risk. Right now, its marketing front end is still a product catalog with 120,000 listings. Imagine the frustrations of a purchasing agent who has to wade through all that. Now imagine a digital first mover with a more efficient search alternative. Maybe it's a Netscape teamed even with a marginal distributor. Together, they could inflict not just a flesh wound on Grainger but strike right at its heart.

MCFARLAN: But the issues are complex, because there are so many factors that can influence the outcome. For instance, some years ago, Wal-Mart set up bar code scanning in a way that would greatly expedite its ordering decisions. Its competitors were unprepared, and, based on my criteria, Wal-Mart should have enjoyed a large first-mover advantage. But H-E-B, a competitor, was scared to death. Wal-Mart's technology could have disrupted industry economics, driving it out of business. If Wal-Mart became so dominant, it would also end up with tremendous leverage over major suppliers, such as Procter & Gamble.

Within two weeks, H-E-B was fighting back. It became a P&G customer and, with help from P&G, began using similar technology. Over five years, H-E-B added 61 other suppliers and drove its inventory turns from 11.4 per year to 38.6. Wal-Mart's hopes for crushing economies of scale were destroyed. The H-E-Bs of retail were saved from annihilation.

SVIOKLA: What the H-E-B story dramatizes is that survivors don't sit back and wait. They put the accent on "fast," not "follower."

MCFARLAN: Yes. And, collectively, these stories illustrate why it is hard to generalize. A first-mover approach fits some situations, a fast-follower approach others. Management has to analyze the buying proposition, the switching costs, and the upfront investment to discover which way the decision breaks.

SVIOKLA: I agree that the decisions are complex. Look at Amazon.com, everybody's favorite example of a successful first mover. I am not bullish on Amazon. It is not clear to me that they are going to get to an economy of scale. At the margin, their economies are actually diseconomies—the more they sell, the more they lose. I say that because they are diversifying into non-book businesses like CDs that are money-losing to begin with. The opportunity is elsewhere, in businesses that have some margin like advertising, like brokerage (which, incidentally, is why E*Trade has a good shot as a first mover).

MCFARLAN: Even worse for Amazon, they have gone after a really powerful competitor. Barnes & Noble has a lot of money. They have a huge reputation at stake. And they will literally, I think, give books away on their Web site if necessary to stop the bleeding. When you come in from left field against big, strong, well-financed competitors, you had better have enough power to kill them, or they will come back and do what they have to.

[Shortly after this discussion, Barnes & Noble raised additional financing by selling Bertelsmann a 50% stake in its on-line business for $200 million.]

SVIOKLA: I think Amazon has laid itself open to just that. Here we return to the element of secrecy that you insisted on earlier, Warren. A first mover's imbedded proprietary knowledge and digital links have to attenuate the risk of being copied. That hasn't happened for Amazon. Already, Barnes & Noble's site is nearly equal to theirs.

I do admire Amazon for its guts. It takes great managerial courage to do two hard things at once: price aggressively enough to set an industry standard of operation and expose your inner workings to the competition. Not only do you have to run fast; you have to get naked, too.

You talk to a lot of senior executives, Warren. The thing that troubles me as I look at the kind of investments that companies make is that they do not seem to have an appreciation for how a radical entrant can screw up their economics. While established companies may talk a good game about entrepreneurship and innovation, at the end of the day it is margin preservation that dominates their actions. How do you wake up a management team to be worried about something like this?

MCFARLAN: If an established company has deep enough pockets, it thinks that all it has to do is outlast a start-up. The instant the start-up is gone, it can move its prices right back up.

SVIOKLA: That is the problem. With Amazon, I expect they'd be right. But for many other digital players, I doubt it. Look at Auto-by-Tel [which locates prospective car buyers on the Web and sells the leads to dealers]. It operates essentially without real estate, personnel, or inventory, so its marginal economics have to be lower than those of AutoNation, [a big chain of dealers] and far lower than a conventional car dealership's. When you consider that scalability is easier and faster for Auto-by-Tel, then you have to pause.

Maybe it's most helpful to end with what the adman, David Ogilvy, used to say: A great concept with different execution can get you 20 times the sales.

MCFARLAN: And dumb luck is always better than strategic planning. [Both laugh.]

SVIOKLA: If you execute.

Dr. McFarlan is reachable at fmcfarlan@hbs.edu. Dr. Sviokla is reachable at jsviokla@sviokla.com.