My greatest revelation as a businessman was crystallized in two words: Let go.

It was 1994, and a team of developers at Knowledge Adventure (KA), the educational software company I founded, was building a “virtual museum” for a new CD-ROM title called Science Adventure. In the process, the team developed a powerful technology for three-dimensional visualization and navigation—so powerful, in fact, that I decided to start selling the technology to other software firms as a stand-alone product.

Problem was, KA’s infrastructure wasn’t well suited to the task. The sales staff, for instance, was trained not in consumer but in business-to-business sales, and the new product began burning through an inordinate amount of resources. So KA’s board suggested that we “spin out” the product and its ten-person development team. (I use the term “spin out” because, as I’ll explain, the venture breaks from the main company but, unlike a “spin-off,” never entirely leaves its orbit.)

I rejected the idea out of hand. Why? Mine was the typical reasoning of any owner or manager. First off, I was loath to release my clutch on such a hot property and the potentially sizable revenue stream it promised. And I was even more put off by the KA board’s insistence that we keep only 19.9% of the new company’s equity. The board argued that the new venture would not turn a profit for several years; by taking less than a 20% equity stake, KA wouldn’t have to report the spin-out’s losses on its consolidated statement of income (and thus dim KA’s luster as an acquisition or initial-public-offering candidate). This was too much to bear. I protested that KA was leaving money on the table, that it was handing over too much equity to the new company’s managers, and that taking an 80% stake would be more appropriate.

The board, however, would not relent. After some struggle, I finally capitulated. With Knowledge Adventure taking a 19.9% slice of equity, Worlds, Incorporated—as the new company was christened—was spun out later that year.

What ensued astonished everyone. Within a year, Worlds grew almost as large as KA itself. Its employees seemed to rise to entirely new heights of creativity and passion—putting in Herculean efforts to close deals, to improve the product, and to recruit new star employees. My earlier reluctance suddenly seemed laughable: instead of owning 80% of a $5 million business, KA now owned 19.9% of a $77 million business.

To be sure, that rapid growth was likely the result of a number of forces. The act of spinning out Worlds sharpened its strategic focus, enabling it to communicate a consistent message to its customers. Spinning out Worlds also enabled it to attract outside investors, giving it greater access to cheap capital—far more than would have been available inside KA. But by far the greatest factor, I came to believe, was the magnification of human potential. That’s because we gave employees near total ownership of Worlds and unleashed a new level of performance, building economic value that more than made up for the fact that KA had kept only a fraction of the equity.

This was my revelation, which I’ve since dubbed the “new math of ownership”: that the multiplicative effect of setting employees free and giving them significant equity has a net positive result. It’s a counterintuitive arithmetic, one that I came to embrace despite myself. But I decided to put this arithmetic—this notion of relinquishing ownership and of letting go—at the very heart of a new venture.

Magnifying Human Potential

As I see it, models of equity ownership fall into three categories:

No Equity.

This is the most common state of affairs: employees work for a salary and have no stake in the overall welfare of the enterprise.

Token Equity.

This is equity sharing as it’s usually talked about: employees have some ownership, but the bulk of their compensation still comes from salary. It’s token ownership, amounting mostly to a vote on the proxy statement. Every once in a while, an employee may attempt to save the company some money somewhere or think: “This is my company, I’m hurting myself if the company doesn’t do well,” just as a citizen who litters might think, “Maybe I shouldn’t litter because it’s my own tax dollars that pay for cleaning this up.” But seldom does this abstract notion translate into a visceral conviction.

Significant Equity.

In this scenario, the monetary value of employees’ equity potentially dwarfs their salaries. This is radical ownership. People won’t throw trash on the lawn because it’s their lawn.

Significant equity, as I’ve come to define it, is owning 1% or more of a company. In one sense, that’s a completely artificial mathematical distinction. Nonetheless, I believe that this figure represents a psychological tipping point. Employee performance, I’ve observed over the years, tends to seesaw atop this fulcrum; with few exceptions, the people turning in truly stellar performances are those with at least 1% equity. The problem with 1% owners, of course, is that you can have only 100 of them. To keep the scale small enough, projects must therefore be spun out as a matter of course.

That is the principle underpinning Idealab, a company I founded in 1996 as the result of my Worlds epiphany. (I sold KA later that year.) A business incubator, Idealab systematizes the process of turning projects into companies. In other words, it systematizes the process of letting go.

The process works like this. In our Pasadena offices, I work with a core staff of 20 people to develop ideas for new Internet-based businesses. At any given time, four or five embryonic companies will be in development under Idealab’s roof. Those that seem flightworthy are then hatched into stand-alone enterprises, whereupon they find their own office space but continue to receive operational assistance from Idealab. About 20 companies have left the nest so far. Among them: eToys, an on-line toy retailer;, a search engine; WeddingChannel, which lets people use the Internet to plan their weddings; and IdeaMarket, an on-line market for intellectual property that survived only a few months before folding last summer.

Idealab’s fund of $10 million comes primarily from individual investors. We’ve deployed about $7.5 million of that so far, roughly half for Idealab’s operations and the other half for investments. Our policy is to give each spin-out no more than $250,000 in seed funding and to take no more than a 49% equity stake—and that stake is usually much less than 49%. Indeed, I insist that Idealab’s interest in its offspring be noncontrolling. Typically, the spin-out’s management team and employees get 30% of the equity, with the chief executive taking 10% of the total. Everyone on the payroll gets something. I also insist on a salary cap for the CEO, usually in the neighborhood of $75,000.

What those equity stakes end up being worth obviously depends on the individual trajectory of each company. But the objective in all cases is some sort of “liquidity event”—going public or being acquired—at a valuation of at least $100 million. The Idealab company CitySearch recently filed to go public at a valuation of $270 million. At that valuation, CEO Charles Conn would have a stake in the company, on paper at least, of $13 million.

The goal is to create a family of independent yet interdependent companies—and a family of owners. The ultimate results of this experiment, of course, remain to be seen. But early returns are promising. Last year, according to the National Venture Capital Association, the rate of return for the venture industry as a whole was 29.5%. In its first two years of existence, Idealab has generated a return of 155%. And most of its offspring have managed to raise follow-on rounds of investment from outsiders.

Tribes of 100

Most discussions of equity invoke the antiseptic language of compensation science. But my experience suggests that the potential for financial gain is actually only a small part of why equity motivates employees. Rather, I believe that once people are fed and sheltered, the rest is basically about fulfilling their fantasies. And equity is a wonderful tool for harnessing the power of fantasy because it involves a story. There’s a protagonist (the company) and an antagonist (the competition); a struggle; and a victor and a hero. Equity means drama. Annuity, by contrast, is boring because you already know the ending. It’s a dull story.

Hence, I believe that equity exerts a pull disproportionate to its actual financial potential. That is, if you gave employees salaries that were equivalent to—or even greater than—the net present value of their equity holdings, they would be less motivated, even though a purely mathematical analysis would not justify their feeling that way. (After all, for most start-ups, the likelihood of a big payoff from going public or being acquired is slim.)

I believe that equity exerts a pull disproportionate to its actual financial potential.

Likewise, belonging to a small team exerts a basic emotional pull on employees. When a company has more than 10 people and fewer than 100, it feels like a tribe—that primordial unit of human organization. The CEO knows everyone’s first name, has met the spouses or partners, and knows who has kids. Fewer than 10 people is a family, the most basic unit of organization and, in many ways, the ideal one (the bonds are the most intimate, the communication the most immediate), except that there are too few people to accomplish much. When you have more than 100 employees, the bonds begin to recede to a level of abstraction.

That’s why I advocate breaking down economic endeavor into tribes of 100, tribes of owners. Within a tribe, people still feel like one clan fighting a common enemy. No one has to be packed off to seminars on team building. We’re ten families congregated around a common campfire or watering hole, and when we look up, we know everyone’s name and recognize each face. There’s something immensely dramatic and immediate—and human—about that. And the shift from physical manufacturing to an information-based economy, along with the decreasing importance of economies of scale that attends that shift, means that small tribes are more viable economic entities than they used to be.

Thus harnessing the power of fantasy is a way to overcome purely economic restrictions. It not only unlocks potential in employees but also helps a company recruit and retain talented people. And compensating employees largely through equity, not salary, can serve as a self-selecting filter: it tends to attract employees whose mind-sets are well suited to high-risk, high-reward endeavors.

Take CitySearch, one of Idealab’s offspring. Its employees are relative paupers—yet its turnover rate is extremely low. The sense of personal involvement in the unfolding drama (in this case, CitySearch versus Microsoft’s makes it difficult for employees to exit the story. Or take CitySearch’s chief executive, Charles Conn. Previously, he was a partner at McKinsey & Company. It is unlikely he would have left that job had CitySearch been just another business unit inside a corporation. What lured him there was the opportunity to direct his own drama, to own his own destiny.

Best of all, radical equity sharing involves few concomitant trade-offs for the founders. (Though those trade-offs—equity dilution and, in the case of stock options, claims against future earnings—are issues that need to be weighed carefully.) The same person who habitually litters can be converted into someone who cares about her lawn. And conversely, someone who cares about his lawn can be converted into someone who doesn’t. Even me.

In retrospect, I learned this lesson even before my Knowledge Adventure experience. In 1986, my brother, Larry, and I sold our software company, GNP Development, to Lotus Development and spent the next seven years as Lotus employees. But we felt like owners because, unlike other Lotus employees, Larry and I earned royalties on the products we developed for the company. These included HAL, a software program that simplified Lotus 1-2-3, and Magellan, which allowed PC users to quickly scan the contents of disks. We worked frantically to promote our products, at one point pulling an all-nighter, shirtless and sweat-drenched, to assemble a trade show booth from scrounged materials. Meanwhile, we beheld in wonderment the relative indolence of our fellow employees, who seldom remained at trade shows a minute longer than required.

Just before Magellan’s release in 1987, Larry and I wanted to be sure the product made a big splash in the market. So with $1,500 of our own money, we bought a used disk-duplicating machine. Wandering down to the Lotus manufacturing facility, we inveigled 15,000 floppy disks that were on the verge of being discarded due to incorrect labeling. (Back then, floppies went for about $1 a piece.) We printed up 15,000 new labels, placed the duplicating machine under my desk, covered it with a black shroud, and set it to work producing 15,000 copies of Magellan over a period of days. (“I’m not even going to ask what that is,” said our boss, noticing a wheezing sound and a red LED light emanating from under the shroud.) We packed the disks into 30 boxes and boarded a flight for Chicago, bribing the baggage handler $100 to overlook the two-parcel check-in limit and load all 30 boxes.

Arriving at Comdex, the computer industry’s semiannual convention, my brother and I each staked out a taxi stand where attendees were in line waiting to leave the exposition. “Free Magellan demo!” we shrieked, managing to unburden ourselves of all 15,000 disks by show’s end. (Security guards asked us to suspend our activities several times, and by the time we returned to Lotus offices in Cambridge, word was that Bill and Larry had been arrested hawking disks at Comdex.) At least one person observed that, running down the aisles of the taxi lines, we looked more like two wild-eyed entrepreneurs than the corporate drones we actually were.

Several months later, Lotus decided it wanted to bundle Magellan with its flagship 1-2-3 software. Rather than continue to pay my brother and me our 5% royalty on sales of the new product, Lotus offered to buy us out with one flat payment. It was a generous offer—equivalent to more than the probable future value of our royalties—and we accepted.

And the next day, we felt like employees.

That may sound somewhat hyperbolic, even absurd. But the transformation was undeniable. It was as if the new chemistry in my body could not have conceived of going to Comdex and hawking diskettes in line. It’s not that I would have thought of it and decided not to do it; the idea just wouldn’t have come up. I had become constitutionally incapable of such a notion. And this happened after six-and-a-half years at Lotus—six-and-a-half years of my brother and me snickering (perhaps unfairly) about our colleagues’ lack of inspiration and commitment. With those six-and-a-half years flashing before my eyes, I turned to my brother and gulped: “Now I get it. This is what everyone else here must feel like.”

Math Lessons, Part Two

At Idealab, mastering the new math has been an ongoing experiment, full of trial and its inevitable handmaiden, error.

For instance, I’d hoped that the owners of the various spin-outs would actively share ideas and knowledge with one another. They didn’t. But that figured: they didn’t have ownership interest in one another’s companies. So I gave them all shares in Idealab itself, meaning that they effectively hold equity in each other’s companies. The result: Toby Lenk, CEO of eToys, shares his most proprietary information (for instance, the cost of acquiring customers and the most cost-effective places to advertise on America Online) with Tim Gray, CEO of the Wedding-Channel. Lenk says he would not do so if he did not have an indirect financial stake in WeddingChannel’s success. Some recently hired CEOs also tell me that the equity stake in Idealab played a role in their decision to accept the job.

But perhaps my biggest mistake occurred when the owners of one of the fledgling companies, IdeaMarket, ran out of money and asked me to dip into Idealab’s pockets. I obliged them, violating our policy of investing no more than $250,000 in each spin-out. Several months later, I obliged again. And again, until Idealab had poured a total of $810,000 into the struggling company.

I came to regret this generosity sorely, and not simply because we’d invested so much money in what proved to be a losing cause—though IdeaMarket’s eventual failure (one of three we’ve had so far) was painful enough. Rather, I was more concerned with the message I had broadcast to other Idealab companies. By pumping so much money into IdeaMarket, I had partially shattered the impression that each spin-out was expected to fly under its own power.

Why is constructing this impression, and making it as convincing as possible, so important? Because if the managers of the spin-out perceive that the corporate center will not allow their company to fail, it deprives the power of ownership of its other, darker, component: fear. That is, the promise of the upside alone will not drive people to their highest potential. There must also be the lurking fear that a competitor will snuff them out of business, rendering their equity worthless.

Of course, it is difficult to make your offspring live in fear. Like any parent, I instinctively want to shield them from danger and make their lives easier. Yet I’m now convinced, from my limited experiences, that too much help early on will ensure failure later on. If companies don’t know that they can run out of money, they won’t be thinking of ways not to run out of money. To that end, I now try to stand firm on that $250,000 limit, though I suppose my earlier profligacy renders my Procrustean pronouncements a little less credible.

That’s financial assistance. So what about operational assistance? Namely, which operational functions do we centralize in the corporate center, and which do we leave in the hands of the spin-outs? Again, my reasoning was that while some assistance is invaluable, too much is counterproductive. So I figured that Idealab should maintain only the barest of supporting threads.

In the end, I arrived at this solution: accounting, legal services, payroll, some design work, and some technology platforms (for example, databases and Web-building tools) would be centralized in Idealab. Product development, sales, marketing and other core functions would be left to the spin-outs. Idealab also provides shared office space, and here again shared equity proves advantageous. Because we don’t have two classes of citizens—salaried-only and salaried-plus-equity—we don’t have the jealousy and contempt that commingling the two can breed.

Putting Radical Ownership to Work: a Provocation

Earlier this year, I was standing in the boardroom of Korn/Ferry International, the executive recruiting firm. For several hours, an associate and I had been trying to convince the board to launch an Internet-based recruiting service, not as an internal business unit but as a stand-alone enterprise to be owned partly by Korn/Ferry, partly by Idealab. We weren’t making much headway. The board members had a perfectly understandable objection: Korn/Ferry was in the business of recruiting, and this new venture was about recruiting too. Didn’t something so tightly coupled with Korn/Ferry’s core mission belong under its corporate umbrella, not outside it?

The board members were right, of course. This was their business. And, desperate for a breakthrough, I told them so. I told them that this on-line venture was not at all peripheral to Korn/Ferry’s mission. And I told them that, despite this fact, they should let go of it. If they did not set it free as a separate business, I argued, the venture would not be able to attract the best talent or get the most out of the talent it did have. Nor, for that matter, could it raise more capital on the open market or be free to cannibalize Korn/Ferry’s existing business.

I relate this story not to celebrate my powers of persuasion (though I’m pleased to report that the board eventually succumbed to my pleading) but to underscore the resistance with which the corporate world traditionally greets these ideas. As I’ve said earlier, this new math of ownership is counterintuitive. Executives are accustomed to spinning off their underperforming units, but letting go of their best people and ideas and technologies—well, that’s something else entirely, something deeply discomfiting, something that feels like corporate apostasy. Moreover, in a world where executive pay and prestige are still largely linked to the size of the assets under one’s control, the idea of relinquishing control represents a direct threat to a CEO.

Perhaps that’s why corporate America has had such a dismal record trying on these ideas—despite the notable success of companies that have. Consider Thermo Electron, the Waltham, Massachusetts, industrial giant that has achieved stellar success with this model, spinning out more than 20 companies around its core intellectual competency of electrical and chemical engineering. Hundreds of corporate managers have met with the founders, George and John Hatsopoulos, to discover the virtues of carving out projects into separate companies. Yet, as John Hatsopoulos complained to me not long ago, few ever try it.

That’s a shame. I believe these ideas could be applicable, in one form or another, in every industry for which intellectual capital is the key asset. If R&D-intensive companies such as 3M or Xerox were routinely to spin out new-product-development projects (especially those that are peripheral but even those that are threatening to the companies’ existing lines of business), they might overcome their historic disadvantage in bringing new, disruptive innovations to market.

There will be those, I’m sure, who dismiss this new math of ownership as so much voodoo, as the microeconomic equivalent of the Laffer Curve (the now-discredited theory that lowering tax rates would actually increase tax revenues). And certainly, there are bound to be instances in which giving away much of your ownership stake would end up lowering your net return.

Employees won’t think and act like owners unless you make them into owners.

Yet corporate CEOs are always pining for ways, as General Electric chairman Jack Welch put it, to “get that small-company soul and small-company speed inside our big-company body.” I submit to them: you can’t really create a small-company feel unless you create a small company. And you can’t expect employees to think and act like owners unless you make them true owners.

A version of this article appeared in the November-December 1998 issue of Harvard Business Review.