(FORTUNE Magazine) – One day last Fall 400 investors and analysts gathered on Wall Street to hear a full day of presentations given by Barry Diller and the cadre of top executives at his Internet company, InterActiveCorp. They came in part because of Diller's unquestioned star power. But they also came because IAC has been anointed one of the "four horsemen" of the Internet--right up there with Amazon, Yahoo, and eBay--although IAC got there in a very different way. Diller, through a staggering number of deals, has assembled a conglomerate that includes the Home Shopping Network, Ticketmaster, Citysearch, Expedia, LendingTree, Match.com, and other well-known brands. Today, although IAC's market capitalization of $25 billion places it behind Yahoo ($36 billion) and eBay ($49 billion), it is slightly ahead of Amazon's $18 billion.

Diller comes from Hollywood, and he was the impresario of a superb show that fall day. The huge ballroom was decorated with multicolored banners bearing the logo of every business. Each one's CEO explained why his particular company could double, triple--maybe even quadruple!--over the next five years. All of this added up to grand goals. IAC hopes to almost triple its revenues in the next five years, from $6.3 billion today to $16.5 billion, while increasing its profits at a 25% to 30% annual clip to $3 billion. IAC also plans to kill its competitors by, as Diller put it, spending "more than anyone else can afford"--some $1 billion this year alone--on marketing. Ultimately, Diller has said, his goal is to create "the largest, most profitable e-commerce company in the world."

To the true believers--and there are many--IAC is in the Internet's sweet spot: As consumers increasingly gravitate toward online commerce, IAC plans to take a cut of just about every kind of transaction, from travel to dating to lending to things you can't yet imagine. But unlike those of Internet companies of yore, most of its businesses are not smoke and mirrors, and IAC has real money in the bank: at least $3.3 billion. And the fact that Diller is the one putting it together just makes it that much more enticing to his legions of true believers. "I'd take Barry Diller over [Cisco chief] John Chambers any day," says James Cramer, the ubiquitous CNBC host. "Put me in the religious camp," adds State Street managing director Larry Haverty, who's owned IAC stock as long as IAC has existed. Diller and IAC, in fact, are a reminder of how much fun it used to be, back in those late-1990s days when anything seemed possible, and no idea--and no growth rate--seemed too outlandish.

Which is precisely why IAC has attracted skeptics as well as ardent fans. It isn't the late 1990s anymore, and we know all too well that grandiose promises can end in ugly disaster. After almost doubling in the first part of 2003, to a high of $42.74, IAC stock has fallen 25% and trails the other three Net titans; there is a healthy chunk of stock--some 50 million shares--currently sold short. When the skeptics look at IAC they see not so much visions of future glory but warning signs in the here and now. IAC indulges in iffy practices, like inventing financial metrics: Its contribution to generally unaccepted accounting principles is the absurd-sounding OIBA (operating income before amortization), which excludes all kinds of costs. In 2003, IAC's OIBA was an impressive $860 million; its GAAP net income was just $154 million. Diller's nonstop dealmaking--more than $8 billion in the past 18 months alone--is not just tortured in its complexity but has resulted in numerous strategy shifts and five name changes. Critics point to the enormous sums Diller has paid himself and his top executives. Diller's complaints about IAC's "undervalued" stock also strike people as unseemly and promotional: IAC has incessantly whined that the other big three Internet stocks are valued much more highly.

But if you think, after reading the previous paragraph, that you know where this story is headed, think again. IAC is a company that can scare you and seduce you at the same time because it contains such a potent mix of reality and fantasy. There is enough reality to give credence to the fantasy, but enough fantasy to cast doubt on the reality. There's never been a company quite like IAC. And before you can understand this company, you have to understand the man who put it together.

His reputation precedes him, of course. Over the years Barry Diller has been described with words like "arrogant," "intimidating," and "bulldog." In public forums such as the one held last fall for investors, he seems even less approachable than your average CEO as he tosses off facts and predictions as if they were holy writ. He has been known to react with vitriol to public criticisms of his company. And IAC's offices--in Manhattan's Carnegie Tower--do nothing to temper that image. The 42nd-and 43rd-floor executive suite has cool, blond wood, a silver spiraling staircase, and fresh flowers everywhere. (The company is about to break ground on a new Frank Gehry--designed headquarters in the hip Chelsea area of New York City.) Several of Diller's assistants--there's a whole pack of them--guard the door to his office, which has a private terrace offering sweeping views of Central Park.

But Diller, like his company, is difficult to characterize. One-on-one he appears anything but arrogant and slick. In fact, for a chief executive, he seems unusually willing to question himself. He rambles. He expresses self-doubt. He admits to mistakes before you can play "gotcha." For instance, during IAC's fourth-quarter conference call, Diller told analysts that it was absolutely critical for the company to be "consistent." But "consistency" is hardly the word IAC's history brings to mind--and it's refreshing when Diller admits in an interview that it won't be easy. "It's hard to be consistent when you're making it up every day," he says. "That's the bitch of it."

Nor does Diller seem to believe he's infallible. He says that "the most solid thing about this enterprise is that we enjoy being told we're stupid," by which he means he likes to be challenged. Though the 62-year-old CEO has been in business longer than many of his executives have been alive, he doesn't impose top-down rule; rather, he has fostered a culture of debate. "You must have people who will say, 'That's nice, you fool, you're wrong,'" Diller says. Meetings at IAC are not decorous affairs, but loud and even combative. "People are interested in finding the right answer, not in affirming their own answers," says Karl Peterson, CEO of Hotwire.com, which IAC bought in 2003. "Diller," adds Doug Ledba, CEO of LendingTree, another 2003 purchase, "has a great ability to hold two simultaneous, seemingly opposite thoughts in his mind." His executives find it impossible to sum him up easily. He's arrogant and self-deprecating; certain of his views and questioning of them; he's created IAC both to satisfy his intellectual curiosity and to make money. "Diller is intense, a builder, an opportunist, a capitalist ... a personality in your face, always pushing," says John Pleasants, CEO of Ticketmaster, one of Diller's first purchases.

IAC's convoluted history began in August 1995, when Diller invested $10 million (half of which was a loan from the company) to buy a 20% controlling stake in a tiny TV network called Silver King Communications, which broadcast the Home Shopping Network. Silver King was controlled by John Malone's Liberty Media, as was HSN, and Malone had recruited Diller. (Liberty still owns a 20% stake in IAC.) By the end of 1996, Diller had merged the Home Shopping Network, Silver King, and Savoy, a small film production company, into an entity named HSN Inc.

Diller, of course, was well past the point where he needed to work for the reason most people do. After starting in the mailroom of William Morris, he had rocketed through the entertainment world to head ABC, run Paramount, and turn Fox into a viable fourth network. But he abruptly departed Fox in 1991--with a reported $140 million--after Rupert Murdoch refused to make him an equity partner.

Diller became fascinated with interactive shopping in 1993, when he visited QVC with his now wife, designer Diane von Furstenberg, who was selling her dresses on the air. Diller called home shopping the "fastest link between action and reaction I've ever seen." Within a few months he became chairman of QVC, a platform from which he tried to buy Paramount, losing a bitter battle to Viacom's Sumner Redstone. When he then tried to buy CBS, Comcast, which had a majority stake in QVC, nixed the idea. By 1995, Diller was gone.

Back to HSN, which Diller first envisioned as the foundation for a new network he would build, a la Fox. But it didn't work out that way. On the one hand, Diller continued to do media deals, purchasing Universal Studios from Edgar Bronfman's Seagram for just over $4 billion in 1998. Diller paid $1.6 billion in cash, gave Bronfman a 46% stake in his company, and changed its name to USA Networks. Four years later Diller sold Universal to Vivendi for $11.7 billion--although only $1.6 billion of that was cash; the rest consisted of the return of USA stock Vivendi owned and various other convoluted securities in a new entity called Vivendi Universal Entertainment. (The ultimate value of those securities is the subject of a lawsuit between Vivendi and IAC.) Diller also sold all of USA's TV stations in early 2001, for $1.1 billion in cash.

But Diller also began making deals that had nothing to do with traditional media. He acquired Ticketmaster, the nation's dominant ticket seller, for about $680 million in stock. He also began investing in an Internet company called Citysearch, which planned to make its fortune from local advertising and commerce. The stated rationale for the Ticketmaster acquisition was that Diller could combine HSN's delivery infrastructure with Ticketmaster's phone capability and outsource its "integrated electronic commerce solutions" to other companies. Though Diller invested roughly another $1 billion to further this strategy, that never worked out either.

By then, though, Internet mania was in full bloom. So Diller spun off Ticketmaster's nascent online division, grafted it onto Citysearch, and sold stock in a new entity called Ticketmaster Online--Citysearch, or, mercifully, TMCS. (The rationale? Citysearch would help steer people to Ticketmaster events.) TMCS's stock shot from $14 to $80, allowing it to acquire a slew of companies, including Match.com (a dating site) and Evite (an invitation site), before plummeting back to earth. Losses ballooned, and in early 2001, Diller reunited TMCS with its brick-and-mortar half by having TMCS issue $670 million of stock to his company in exchange for the old-fashioned piece of Ticketmaster.

Then came the travel business. In a 1999 deal Diller paid $245 million for a tiny Texas-based company called Hotel Reservations Network that sold blocks of hotel rooms at steeply discounted rates over the Internet. (In 2000, Diller sold a stake to the public and later on renamed the business Hotels.com; in 2003, IAC reacquired the shares for $1.2 billion.) Diller also bought a $1.5 billion stake in Expedia, which had been spun out from Microsoft a few years earlier. Erik Blachford, an Expedia executive who would later become its CEO, recalls Diller's saying, "You guys think this is the future of travel. I think it's the future of everything."

The deal that garnered the most headlines in the 1990s was the one that failed: Diller's attempt to buy just over 60% of search engine Lycos for $3.8 billion. Though his critics today say he dodged a bullet, Diller vehemently disagrees. "Those dopes!" he exclaims. "I did not dodge a bullet! We were going to wire all of our commerce to the No. 3 search engine at the time when habits were just changing. Our company would have been so far advanced! I would have loved it!"

With all the bewildering wheeling and dealing, Diller achieved four undeniable results. First, he created a cash-rich company with a healthy $3.3 billion on its balance sheet (not including the value of the Vivendi securities). Part of that came from Diller's dealmaking, but he has also owned businesses that have produced lots of cash. Second, Diller did quite well for those shareholders who stuck with him. IAC dates its founding to Dec. 16, 1996; from that starting point, IAC's stock has returned around 600%, or more than triple the S&P 500's return. Diller views this as the answer to just about any form of criticism. "It's amazing that anyone would look at our company any other way than its only reality," he says. He has done particularly well for Malone, who is entitled to keep his IAC stake at 20%. This past spring, for instance, after IAC planned to issue a ton of shares to make acquisitions, Malone's deal allowed Liberty to buy over $1.6 billion of stock--for $1.16 billion.

Diller also made himself a fortune. He's guaranteed--yes, guaranteed!--$255 million from the deal with Vivendi. Add in Diller's two million shares of IAC, 42 million stock options, gains from stock sales over the years, salary, and bonuses, and he's reaped some $1.6 billion. That's not much less than the $2.4 billion the company has reported in net income over that same time, which helps explain why Grant's Interest Rate Observer opined last year that "no public investor in Diller's company will ever do as well as Diller himself." (To be fair, Diller has not taken any stock options in six years.)

And finally Diller's wheeling and dealing has given rise to a company--the name was ultimately changed to InterActiveCorp. in mid-2003--that, to hear Diller and his executives tell it, is no longer a hodgepodge of disparate assets but makes coherent business sense. It's all about selling to consumers, whether your interest is travel (Expedia) or treadmills (HSN), finding a mate (Match.com) or a mortgage (LendingTree). Assets that didn't mesh with that focus have been sold. Companies that extended that focus, such as LendingTree and Priceline competitor Hotwire.com, have been acquired. Diller took almost $1 billion in pretax charges to clean up the company's balance sheet and spent $5.6 billion buying the publicly traded shares of Ticketmaster, Expedia, and Hotels.com in an effort to simplify the company's structure for investors. "We finally have a poker hand we like," CFO Dara Khosrowshahi has been known to say.

If you were looking to cast someone to play the CEO in a movie about the dot-com era, you could pick the men running the various IAC businesses. They're mostly young, white, thirtysomething MBAs--Harvard appears to be their business school of choice--many of whom once worked at TMCS. They're trim and handsome, and favor the open-collar-shirt look. They say the kind of optimistic, enthusiastic, hyperaggressive things that dot-com executives used to utter. Listen, for instance, to Tim Sullivan, CEO of Match.com: "Our mission is truly bringing happiness to the world. I'm not afraid to say it!" Or CFO Khosrowshahi: "We're here to conquer." You'll also hear more than one say some variation of "Barry pushes me, but not as hard as I push myself." You'll discover a little bit of arrogance and a lot of ambition, but as with Diller, you don't get the sense that they're selling a story they don't believe. Indeed, if there's Kool-Aid drinking going on, they're drinking it.

Most of them are also dot-com rich; Khosrowshahi, for instance, made $8.3 million from selling Hotels.com stock. (He received that stock because he sat on the Hotels.com board.) John Pleasants has netted $3.8 million from stock sales; there was also very heavy insider selling at Expedia and Hotels.com before IAC bought back the publicly traded shares. Erik Blachford got a slew of Expedia options right before the buyout was announced, resulting in a windfall of $3 million. Most of the top executives have millions of dollars' worth of stock options. IAC has 98 million options outstanding, which is why some eyebrows were raised when Diller announced in 2002 that IAC would no longer hand out options--even characterizing them as a "get rich quick" scheme. Says Albert Meyer, an IAC skeptic who founded 2nd Opinion Research: "The horse has bolted, and now the stable door is closed."

But if the CEOs fit the dot-com stereotype, the companies themselves are all over the lot. Yes, they are all in the business of selling things to consumers, but not all of them do so online--in fact, about 50% of IAC's sales are offline--and the closer you look, the more you're struck not by how similar IAC's businesses are, but how different. Everything about Ticketmaster, for instance, from its Sunset Boulevard address in Los Angeles to its near chokehold on the ticketing business, exudes power and success. Its old partner Citysearch has never made a penny despite numerous strategy shifts. HSN, with its $2.2 billion in revenues, 53-acre Florida campus, and omnipresent screens that track every single sale every second of the day, represents a certain solidity. Yet while HSN does 15% of its sales online, it sells most of its wares on TV. Today it has virtually nothing in common with the rest of IAC.

And on it goes. LendingTree is a classic dot-com, with all that implies: It could represent a revolutionary way for consumers to fulfill all their financial needs, or it could turn out to be a flash in the pan. IAC's crucial travel business--which accounts for 41% of revenues, and which Diller predicts will become the largest travel company in the world--has been terrifically successful so far. But it is engaged in a titanic struggle with its most profitable suppliers, the hotel industry. Starwood Hotels chief marketing officer Steve Henkin explains the state of the travel business this way: "This industry is changing at a phenomenal rate, and for anyone to say he knows where it's going is crazy." In truth, in virtually every business IAC owns, the size of the opportunity is matched only by the size of the risk.

Let's look again at Ticketmaster. With its $743 million in revenues last year--and its 19% profit margin (using, alas, OIBA)--it is by far the dominant player in its industry. And as ticket sales have gravitated to the Internet--Ticketmaster now does half its business online--its profitability has risen. To boost ticket sales, CEO John Pleasants has pioneered all sorts of new Internet-based strategies, such as last-minute e-mail alerts. (It sent 300 million last year, which resulted in an additional $43 million in ticket sales.)

But the Internet also has the potential to diminish Ticketmaster's importance. What's to stop its clients, such as event promoters, from doing their own online ticketing? Pleasants's challenge is to ensure that doesn't happen. He is trying to convince clients that by using Ticketmaster, they will sell more tickets--and make more money--than they would otherwise.

Now look at Citysearch, which has been subsumed into a new division called Local Services, where it exists alongside Evite and a recently acquired business called Entertainment Publications (EPI), which sells coupon books offering restaurant discounts. Diller has big plans for EPI, which currently makes up most of Local Services' revenues and all of its OIBA. But those plans revolve around turning the business into an online service--and currently EPI has a minuscule Internet presence. ("Once you're able to throw targeted discounts to people online instantly, we think the business frankly explodes," Diller has told the Street.) It also requires some sort of synergy with Citysearch and Evite. Given Citysearch's previous failed attempt to mesh with Ticketmaster, skepticism is warranted.

As for Citysearch itself, now on its third or fourth business model, it remains IAC's black hole. The latest plan is a "pay for performance" model, in which local merchants pay Citysearch whenever users click through to their establishments. IAC has told the Street that Citysearch hopes to turn a profit in the fourth quarter of this year. But Diller is willing to entertain the notion that it might not work out that way. "We will either pull it through, and people will say, 'Well, it took them a long time,' or we'll apologize and go on," he says. "I think it'll be the former, but I don't know."

HSN creates a different sort of question: Diller's dealmaking skills notwithstanding, how good an operator is he? Analysts constantly speculate that HSN will be sold, not only because it's not an Internet business but because it has struggled. In the nearly ten years Diller has controlled it, HSN has always been the also-ran to QVC; currently it has half the sales and just a third of the profits. It's also on its third management team. New CEO Tom McInerney says that when he first joined HSN, he thought there was some secret to the performance gap, so he spent three months digging into it only to conclude that QVC "just had better execution over the past decade," meaning that everything from its merchandise to its service was better. He adds, "If we sold the exact same thing head-to-head they would beat us."

Without question, part of the problem was Diller himself--as Diller forthrightly admits. "We were fed up with the fact that they hadn't grown, and we forced them to push sales beyond their capacity," he says. "I blame first us--without question me--then I absolutely blame management for not saying to us, 'You can't do that.' " After he brought in McInerney last year, Diller says, he tried to push the new guy into doing acquisitions to juice growth. To his credit, McInerney refused. ("Thank God," Diller says now.) Today McInerney's focus on better execution seems to be working. Sales are growing at double-digit rates, and last year HSN took market share from QVC.

LendingTree? Diller called it "probably the most important strategic foot we've put down in the last year" when he announced the purchase in the spring of 2003. IAC has also made the astonishing prediction that LendingTree's revenues, which currently stand at $160 million, will reach over $400 million in 2008, while its OIBA will increase a stunning 48% annually to reach over $125 million.

LendingTree's core business--helping consumers find mortgages--has climbed like a rocket, and the company is now trying to break into the business of matching consumers with real estate brokers and take a piece of the $1.2 trillion real estate market. But Diller bought the company--at a 48% market premium--at the height of the refinancing boom, and much of its business has been refinancings. In the fourth quarter of 2003, with interest rates climbing, LendingTree lost money. Which raises the obvious question: Was its early success a result of the lowest interest rates in decades--or can its profits still rise meteorically when rates increase?

And then there's the travel business.

Travel is the main reason that IAC's stock shot up in the first half of 2003 and the main reason that it has since lagged. Expedia, Hotwire.com, Hotels.com, and the other businesses that make up IAC Travel are at the heart of the debate over IAC. This debate is partly about the prospects of the business and partly about the future of the Internet. But it is also about the level of faith investors are willing to put in Diller and his executives. Following the buy-in of the publicly traded shares of Expedia and Hotels.com in 2003, IAC decided to provide investors with just three data points on the amalgamated travel business: total revenues, OIBA, and operating profits. Khosrowshahi says that since this is how IAC plans to measure the business, Wall Street should look at it that way too. Skeptics say that IAC is trying to hide deteriorating results in the individual businesses, particularly in the profits that Expedia and Hotels.com earn from hotels.

There is no question that IAC Travel has been phenomenally successful: Last year it sold $10 billion worth of travel--that's hotel rooms, plane tickets, car rentals, time-share stays, and so on--making it one of the world's five largest travel agents. Blachford has said that IAC Travel could sell as much as $86 billion in travel by 2010. IAC is also gunning for a piece of two huge additional markets: corporate travel and the $350 billion European market. "Our goal," Diller has said, "is to be the No. 1 seller of worldwide travel on-or off-line in three years."

The core issue right now is the hotel portion of the business. After 9/11, when the travel industry was in the midst of a terrible recession, hotels handed blocks of deeply discounted rooms to Expedia and Hotels.com. The travel sites, in turn, would take the rooms, mark them up by 25%--or more--and sell them online at rates that were still lower than you could find elsewhere.

The steep discounts drew a flood of consumers, giving the online agencies power over hotels: Expedia can shift market share between Manhattan hotels simply by placing one higher in the search results than another. But more recently the major hotel chains, as the economy improved, realized that allowing IAC to control their pricing wasn't such a good idea after all. Today Expedia and Hotels.com have deals with five major hotel chains in which they get access to more rooms--but promise not to undercut the hotels' pricing and accept a lower margin. Most of the chains now guarantee that you'll find the lowest price--usually equal to the Expedia or Hotels.com rate--by booking directly with them.

There is little doubt that the chains see themselves in a struggle with IAC. Some are doing everything in their power to get customers to bypass IAC, such as offering loyalty points only on rooms that are booked directly. And they're having a surprising amount of success. John Davis, CEO of Pegasus Solutions, which provides services like reservation technology to the industry, says that at a recent conference he attended, all the bankers were talking about how they now checked prices on Expedia--and then went directly to the hotel's own websites to book the rooms.

For their part, IAC Travel executives--Diller very much included--insist that those are nonissues. CEO Erik Blachford says, for instance, that IAC does only 20% of its business with the big chains. (But he's not including franchisees in that number.) The rest comes from independent hotels--which still offer a rich margin. Indeed, on the last conference call, Diller said that "visions of serious margin deterioration are not realistic." But there will be no way to verify this claim, because IAC no longer provides information on gross margins to investors.

Here's the larger question: For all the money IAC Travel is spending on advertising, are customers drawn to Expedia because it's a great site with a great brand--or because they're looking for the kind of cut-rate deals that were more prevalent a year ago? PhoCusWright, a travel industry research firm, has data that would suggest that consumers are less than brand loyal. It says that the average customer searches three travel websites before making a purchase. But if customers continue to flock to--and buy through--IAC Travel, then over time IAC will have the leverage to keep its margins high. "You can't forget, these guys are from Microsoft," says analyst Paul Keung of CIBC. "It's all about getting volume, and then the margins are here to stay."

The continued growth of travel is important for another, less obvious reason. IAC's operating cash flow--$1.3 billion in 2003--is one of the most appealing things about its business. But around $330 million of that cash flow is "float" from the travel business: that is, money customers paid IAC upon booking their trip, but that IAC has not handed over to the hotel or the airline because the trip hasn't occurred yet. IAC says that this is "permanent cash that we can put to work."

Not everyone agrees. Although Meyer thinks the float may be sustainable, he also argues that "IAC does not gain title to this cash." He thinks a more accurate figure for IAC's operating cash flow is just over $850 million, which also excludes the tax benefit IAC receives from employee stock option exercises.

You can't help but be struck by Diller's willingness to entertain questions about IAC's ambitions. He is not like a typical CEO, who stays "on message" at all times. Listen to Diller talk about eventual synergies among the various business--synergies that would seem critical, because otherwise what's the point of collecting these disparate entities under one corporate umbrella? "It's the most important potential benefit we have," says Diller. "If we're right about just that, there will be enormous competitive advantages and very high barriers to entry." If IAC is able to pull this off, which Diller admits "is a bigger 'if' because it's an executional nightmare," then, he says, "this company has the kind of power that is only dreamed about." But he also concedes, "This idea should be treated skeptically. We haven't proven it yet."

Diller is also willing to hold another possibility in his mind: that the Internet ultimately destroys IAC's business model. You only have to visit Froogle, a recent offering from Google that scours the web for the lowest prices on any item, to see the risk. Five years from now, will most of us go to Expedia to book a trip to Hawaii, or will we use a version of Froogle to instantly search the Net for the best deals?

"I think about that a lot," says Diller. "I believe that intermediaries who deliver services to the consumer that are of value are going to be able to sustain their margin. All of my experience tells me there's a chance to do that." Then he adds, "I could be completely wrong, but I like my bet."

Nor, for all his talk of consistency, is Diller promising that IAC won't change course again. "We reserve the right to zig and zag on various dimes in the road. That's one of the things that created this thing." He adds, "I'm never absolutely sure of anything, and I don't want to be. You're either right and you'll pull through, or you're not. We're never going to be right about everything, and we've certainly been wrong."

Ultimately, Diller's willingness to be both a believer and a skeptic is reassuring. He is fully aware of the gritty realities he faces in making the fantasy of IAC come true. He'll adjust to those realities if he has to. But just don't trash his fantasy. "The only thing that makes me crazy is cynicism," he says, "because I've been fighting it all my life."

FEEDBACK bmclean@fortunemail.com