Boids and Kenny Rogers

"Arguing in My Spare Time" No. 2.15

by Arnold Kling

Sept. 18, 1999

May not be redistributed commercially without the author's permission.

"Today, however, we are trying to straddle the fence--to maintain the traditional mind-set, in which capital is the key resource and the financier is the boss, while bribing knowledge workers to be content to remain employees by giving them bonuses and stock options."

--Peter F. Drucker, "Beyond the Information Revolution" Atlantic Monthly, October, 1999.

The quotation above from Mr. Drucker is a concise, eloquent statement of a problem that I have struggled to articulate in my aimst essays. For example, From Allocating Capital to Allocating Talent.

It is disappointing, therefore, to see Mr. Drucker arrive at a relatively tame and unimaginative solution. He concludes by saying that knowledge workers can be dealt with by "turning them from subordinates into fellow executives, and from employees, however well paid, into partners."

What Mr. Drucker’s approach fails to address is the fact that executive decision-making is not a scalable solution to the problem of management. Even if we turn the 15th-floor conference room into a football stadium, that does not make it feasible to accommodate knowledge workers in large organizations into the process for setting strategic direction.

The question becomes: what management processes can work in big organizations with a large share of knowledge workers? There may turn out to be no answer--such organizations may be destined for extinction.

What sorts of institutions will arise to solve the problem of allocating talent? Let me propose a principle and try to trace out its implications. The principle is this:


Suppose we posit that talented individuals must find a way to work on the right issues in a coordinated way without the direction of "top management." What sorts of problems need to be solved in order to make this work? As I see it, there are two classes of problems that require solution. These I call "boids problems" and the "Kenny Rogers problem," respectively.

The "boids" problem

Steven Levy, in his book "Artificial Life," describes how Craig Reynolds developed a program to simulate the flocking behavior of birds.

"To Reynolds it appeared that there was no drill instructor bird, that the phenomenon of flocking was a decentralized activity, where each bird followed simple rules." (p. 76)

Reynolds programmed some "bird-oids," or "boids" to follow simple rules. Stay close to other boids, try to match the speed of nearby boids, and try to avoid bumping into things, including one another. With these rules, which each boid could apply in relation to its local environment, his simulated birds behaved as a flock.

To solve the talent allocation problem without a manager, knowledge workers are going to need to develop behaviors that solve the "boids" problem. That is, they will need to develop local rules that enable networks of small companies co-operating with one another to act as if they were a large integrated organization.

Like the "boids," knowledge workers will need to be able to operate by using rules that require only local information. For example, they might need rules of thumb for deciding when to solve a problem yourself and when to get help.

A "boids" problem that has come up in my business is renegotiability. Possibilities for acquiring or being acquired by other companies have caused my partners and I to tear up several agreements in the past three years. We always write an agreement with long-term incentives, but when the new opportunity arises the original provisions of the agreement are impossible to implement. As a result, we end up settling our old agreement on the basis of "Well, if we had stuck to our original plan and met our objectives, the present value of the payout to A would have been x and the payout to B would have been y, therefore. . ."

My guess is that for knowledge worker "boids" to collaborate they are going to have to develop rules of thumb that make it possible to build in long-term incentives for one another. However, they will need to be able to untie the strings quickly and cleanly as opportunities arise. If my experience is representative, the ability to renegotiate agreements quickly and fairly will be important.

It is ironic that so many web sites seem to be springing up with the idea of creating markets to allocate talent. Markets provide information for decision support in the form of prices. While this is helpful, in my view the "boids" problem is broader and more complex than one that can be solved by price signals alone. We need more than markets.

The Kenny Rogers Problem

One important point that Mr. Drucker makes is that with new-era businesses, "the financial fruits are likely to take much longer to ripen, if they ripen at all." If this is the case, how can knowledge workers allocate themselves effectively? How are they going to learn the difference between a project that will pay off eventually and one that is going to fail?

From society’s point of view, we want knowledge workers to stick with the winners and dump the losers. I call this the Kenny Rogers problem, after the singer’s famous line, "You’ve gotta know when to hold ‘em, know when to fold ‘em."

Large organizations tend to resist saying "hold ‘em" to new ideas. I have argued (Should Corporations Encourage More Risk-Taking?) that this is not likely to change. If companies were to adopt the Gospel According to FastCompany and remove impediments to middle-management risk-taking, they would end up drawing to too many inside straights.

Our current environment is one in which the critical decision is made by venture capital (VC) firms. Entrepreneurs show their cards (or napkins) to Kleiner-Perkins or Draper-Fisher or CMGI, and those VC’s say "hold" or "fold." If they say "hold," then the company eventually either goes public or is acquired by a large public company. If they say "fold," then the company dies.

It used to be worse. I like to tell the story of Christopher Locke, the gadfly who launched a project on the World Wide Web in the early Fall of 1994 that embodied many of the characteristics that ultimately became hot Internet buzzwords. His boss, Alan Meckler, took one look at his cards and said, "Fold!" These days, the adjective used most often to describe Locke is "cranky." The adjective used most often to describe those who received VC support while following in Locke’s footsteps is "billionaire."

Nonetheless, my confidence in the current solution to the Kenny Rogers problem is rather low. See Arithmetic in a Bubble.

The problem with the current system is that it does not build in enough patience. Concepts are being declared winners before they are fully implemented. It is as if drugs were approved for humans without even testing to see whether they work on animals.

"Internet time" has not changed the fact that it takes several years to assess the viability of a business. I would have more confidence in the VC business if it would return to the historical practice of holding companies for long incubation periods.

In the future, will a talented person work for just one company at a time? If so, then the need for diversification of risk will have to be met by the capital markets. Somebody will have to feed me while I wait to see whether my grand enterprise is going to succeed.

An alternative might be for a person to work on two or three businesses at once. Someone might work on some projects for a salary and on other projects for equity. There may be benefits to this internal diversification, including greater breadth of learning.

Mr. Drucker suggests that we are still in the early stages of the revolution that is being propelled by the Internet. Thus, it is difficult to predict the institutional changes that are likely to take place. Prediction is particularly difficult if we accept the premise that talent must allocate itself. Both the "boids" problem and the "Kenny Rogers" problem are challenging.