Anatomy Of A (R)Evolution

You say you want a revolution
Well, you know
We all want to change the world
You tell me that it’s evolution
Well, you know
We all want to change the world

You say you’ve got a real solution
Well, you know
We’d all love to see the plan
You ask me for a contribution
Well, you know
We are doing what we can

Don’t you know it’s gonna be
All right?
All right?
All right?

“Revolution,”
John Lennon

If you’ve read to this point, well, you know: We all want to change the world. All right?

But how?

Big subjects generally don’t have tidy, easy answers. And changing the world qualifies as a big, Big, BIG subject if anything does. But we’ve already seen some of the moving parts underlying societal change—organizations and people.

We have considered organizations bent on “fixing” poverty, health care, education, and the environment. Working alone or in partnership, such organizations draw a bead on a problem and attack it. Though the details of their approaches vary, we have noted important similarities among them. Big Picture Design helps ensure that their focus does not narrow to the point that they lose sight of the large set of interlocking factors that must be addressed to succeed. It encourages a healthy (but legal) dose of stealing ideas as well, to build upon the experience of others. Making It Appropriate avoids cookie cutter solutions when what is required is a social pastry created to meet the nuanced needs of a particular problem. Making It Stick takes into account the difficulty of implementing a new idea—even one that can provide undisputed societal benefits. Much like a catalyst, Making It Stick can help a nascent solution get over the hump and then help embed it among the people or society it is aiming to serve. Making It Bigger means serving more people, more communities, in more ways.

If organizations drive societal change, what drives organizations? Certainly, a big part of the answer is people. In an earlier chapter, we peeked at snapshots of the lives of five individuals. Each of these change agents/social entrepreneurs/world beaters—it doesn’t matter what we call them, and they probably simply think of themselves as “people”—is like a gear that meshes with other people to create the kind of organizations that we’ve been talking about. Or look at it this way: People plan, take action, evaluate and do all the other things necessary for “stuff” to get done. And, by their actions, bigger stuff gets done at another level, the level of the organization. Companies don’t decide, people do. Schools don’t teach, teachers do. Pharmaceutical companies don’t discover new drugs—that comes from human effort. Yet, companies, schools, and pharmaceutical companies surely exist, and they get (big) stuff done.

Is this sleight of hand or mumbo jumbo? No: it means that things happen at different levels of aggregation. Individuals do; their actions in aggregate define the larger outcomes of organizations. But what about organizations? Is there a way to view their activities in aggregate, interacting to produce even grander outcomes?

We turn to that question and answer yes. If we want a positive societal “revolution,” at rock bottom it will require concerted action by people whose efforts are channeled through the right businesses and other organizations. In turn, certain organizations play a role in a larger “dance” where movement, at first, may be harder to notice, but once you do, you see a massive tango among interacting organizations—rich and complex footwork and legwork, combined in rhythm to form a powerful pattern of societal change. Or, using the language of “complex systems” we can describe the situation this way: the behavior of a beehive is emergent rather than being the simple sum of individual bees’ behavior. This means that things that happen at the level of the hive (such as dances telling other bees where to find food) are of a different from what happens with an individual bee and serve a different, higher-level purpose. So, too, can we view organizations’ behavior as being emergent from the actions of individual workers. And, large-scale societal revolution can emerge from the actions of different organizations—again producing effects at a scale that is different from what any single organization would be able to.

One more thing:

You say you want a revolution and
You want to change the world?
I’ll tell you that it’s evolution

Stay tuned:I’ll explain how evolutionary processes can produce emergent, revolutionary changes in society.

What do ligers, seedless watermelons, and YouTube have in common?

Each represents a kind of genetic “mash-up.” A liger is the interspecies offspring of a male lion and a female tiger. (Switch the relationship sexually, and you get a tigon.) Crossbreed two incompatible varieties of watermelon, and the result is a melon without seeds. And as we’ve seen already, just like mammals and fruit, organizations can be hybrids, too. It is just as possible for one organization to borrow traits from another organization as it is for a lion to “borrow” traits from a tiger. Consider YouTube, which strongly embraces elements of a hybrid: it borrows from nonprofits the trait of engaging a vast army of volunteers, who then create, evaluate, and market YouTube’s content for free—an idea borrowed, it seems, from Tom Sawyer whitewashing a fence. Yet YouTube is a for-profit company (which Google purchased for more than $1.5 billion).

In this chapter, we put nature under a microscope and look at just what traits and borrowing mean. A quick genetics lesson will help us understand that identical concepts pertain to both hybrid organisms and organizations, and that a common evolutionary process governs both biological change and organizational change. Then, in the next chapter, we look in more detail at the processes by which evolution wages a competition to winnow more fit from less fit organisms—whether those organisms are plants, animals, or organizations. This will help us understand how we can create large-scale change at the societal level.

Putting Nature Under the Microscope

Entities in nature—plants and animals, wheat and corn, ligers and llamas—share a basic genetic architecture: a genome. But for every kind of entity, the genome is different. For instance, there are different genomes for bacteria, viruses, dogs, cats, birds, bees, and every type of plant.

It may be useful to think of each genome as a long string of information. This string is divided into major subunits, called chromosomes. Along the entire length of the string are various codes, or genes. Two horses will share the same structure. Their strings will be the same length, divided into identical chromosomes. And, position for position along this string, there will be the same genes.

In actuality, the situation is a bit more complicated. (And even what I’m about to write is far from the complete story.) First, the genome for most entities is actually a pair of strings with identical structure. At identical positions along these two strings lie the same genes. For a particular horse, a specific gene that controls its color may be a code for “black tail only” on one of these strings but “black pigment everywhere” on the other. It is the combination of these two different “alleles”—or forms of the gene—that give the horse its actual color.

Finally, the string that I have described as being unbroken from end to end is, in fact, broken into separate chromosomes. In the case of the horse, there are thirty-two paired chromosomes, or sixty-four in all. But the same genes will appear at the same position on both members of a pair, exactly as we described for a single unbroken string.

Genes are made up of the hereditary material DNA. DNA is itself made up of four chemical building blocks. Thus the genome of an organism can be thought of as its complete set of genes, or its complete set of DNA, in a particular sequence.

Though there is a specific genome for one type of plant or animal, individual plants (or animals) of that type will differ. Some people have blue eyes, others are bald, some need glasses, and others need braces. These differences, as we’ve seen, depend on which specific alleles, or form of a gene, that individual has: genes for blue or brown eye color, a gene for baldness or hair, and so on.

Let’s think for a moment about how genes might change. Suppose one of your friends has the gene for “knock-your-socks-off-itis,” a condition where his socks won’t stay on no matter what he does. Teeth chattering during the winter, he expresses the desire for his children to be free of this condition (that is, if he overcomes his “cold feet” and ever has children). What might he do?

Well, through genetic engineering he might try to have his knock-your-socks-off-itis gene removed via “gene knockout.” Or he might get it replaced with the assistance of a virus that is carrying the “socks-stay-up” allele for this gene. Or maybe he could just sit under an X-ray machine for hours, hoping that the electromagnetic radiation produces the desired mutation.

As you can tell (I hope!), this is a made-up example. But gene knockout and replacement via genetic engineering and genetic mutation are quite real. Each can alter the DNA, and thus the genes, of an organism. In this way, we tend to get modified versions of organisms in future generations.

Of course, your friend with knock-your-socks-off-itis might try the mating route: have a child with someone who does not have this gene and hope for a “socks on” offspring. The “math” here, as we’ve seen, comes from the fact that, under Mendelian inheritance each parent contributes genetic material to his or her children, thus allowing the possibility of a gene from one parent to counteract a gene from the other.

Even this stylized description should help us see that it is possible to start with a certain organism and then begin to introduce genetic variation, which ideally, but not necessarily, creates improvements. Indeed, we can consider a hybrid organism to be one with traits from different sources. With this broad definition, we include the variability that comes from mating a lion and a tiger or from crossbreeding different varieties of watermelon. But we also include organisms that have been “tweaked” by altering their genes such as is done to make certain mosquitoes malaria resistant (thus preventing them from spreading the disease among a vulnerable population), just as we said that tweaking can eliminate knock-your-socks-off-itis. As we’ll see, these same ideas apply to creating cross-sector social hybrids.

Why are these ideas important to us, given our interest in improving the world? First, the notion of introducing variation and then letting successful variants proliferate can be thought of as a model to better society. Through societal “genetic engineering,” we can alter specific traits of an organization, or possibly combine the traits of two different parent organizations.

A deeper examination, however, will show that this is not only a model for how to better society, but is how organizations really do adapt and change to become more effective. The evolutionary forces that control the natural world are every bit as applicable in an organization-in-society context. In fact, in the next chapter we will look carefully at how to understand and leverage these effects in a societal setting. But let’s begin with a rough understanding of how this works.

Putting Organizations Under a Microscope

There is nothing special about wheat (a natural hybrid occupying more farmland than any other plant), seedless watermelons, or ligers. And one organization borrowing traits from another is just as possible as a Labrador retriever “borrowing” traits from a poodle, or a lion “borrowing” traits from a tiger.

Consider a normal, neighborhood bank in Minnesota that makes loans to its customers. If there were a “banking genome,” it might have genes for describing the following: the amount of money a bank has available to lend, who its loan clients are and where they live, and how it obtains the money it uses to make loans, among dozens, if not hundreds, of other features. By understanding and manipulating this genome, we could change the characteristics of our bank in Minnesota. We could take the gene for “amount of money available to lend” and change its value (that is, change its allele) from $10 million to $50 million; replace the allele representing “loan to local citizens” with one representing “make loans to the developing world”; and similarly adjust the value of the gene for “source of funds” from “local depositors” to “online lenders.” In essence, by making these adjustments, we have changed the description of a bank that makes loans within its own community to an online microfinance institution that attracts and distributes money around the world.

“But do banks have genomes?” you might ask. Fair question. In truth, a genome is nothing more than a language for creating a precise description. Though there are more than 6.5 billion people on the planet—no two of us identical—a single genome is sufficient for describing us all. That is because the genome specifies, position for position in every chromosome, the type of information allowed, rather than a specific value. For instance, there is a position on the genome for a hemoglobin gene, but some individuals may have “sickle genes” (actually, alleles), whereas most people have normal hemoglobin genes. Similarly, the genome indicates on which chromosomes, and where upon them, eye color is coded for—and different values there produce different eye colors.

This idea of a descriptive language can be extended in a natural way to describe businesses or other organizations. For instance, we can imagine “corporate genes” that indicate whether an organization is a for-profit or nonprofit, and whether a corporation’s stock is publicly traded on a stock exchange or privately held by a small group of people. If we really wanted to create a corporate genome, we might begin to argue about how many genes to include, though it is unlikely that there would be anywhere near the estimated 20,000 to 100,000 genes in the human genome. But for now, all that is really relevant is that we could create a corporate genome. And because we can, we have a description (you might think of it as a way to spell out instructions) powerful enough to account for all the variation in organizations.

Example: Multi-Organization Hybrid and Its DNA
The Global Fund

Recall that our definition of a hybrid organism is one with traits from different sources, and that this definition applies both to biological organisms and artificial organisms such as organizations. Such hybrid organizations are likely to become more and more common, even assuming the form of strange hydralike creatures with multiple heads: part business, part government, part charity, and part multilateral. In fact, we see such multiheaded organisms attacking some of the world’s severest problems. For example, the Global Fund to fight AIDS, Tuberculosis, and Malaria is a partnership between governments, nonprofits, businesses, and affected communities to help raise and disperse funding for international health issues. The “hybrid power” of the Global Fund is extended further when it combines forces with other organizations, creating what we might view as an even more complex uber-hybrid.

The Global Fund has targeted Tanzania, an East African country relying mostly on agriculture, as a recipient of its aid. Like most sub-Saharan African countries, Tanzania is poor, the average citizen living on approximately three dollars per day in terms of U.S. purchasing power. Ninety percent of the population lives in an area where malaria can be contracted. Malaria is a highly preventable, treatable disease in the developed world that kills at least one million people a year—or one person every ten to thirty seconds—in countries too poor to prevent or treat it.

To combat malaria in Tanzania, the Global Fund needed help. There not only was a need to combat malaria, but also a need to come up with a solution that was affordable and appealing to those who needed it most. The Global Fund made a grant to the government of Tanzania to provide antimalaria bed nets in an effective, fair, and sustainable way. What was required was to manufacture this product, distribute it, and sell it in a way that would not destroy local economic activity. A voucher system was managed by a faith-based nonprofit, Mennonite Economic Development Associates (MEDA), to help at-risk individuals buy bed nets from local businesses. The bed nets were made locally by A–Z, a company in Dar es Salaam using special chemicals produced by a Japanese chemical company, Sumitomo, which also licensed its proprietary technology. Because many Tanzanians were not accustomed to using bed nets for protection against malaria, the job of educating citizens about why and how to use them effectively fell to the international nonprofit known as PSI (formerly, Population Services International), Tanzania, headquartered in Washington, D.C. UNICEF helped distribute some of the bed nets that A–Z made elsewhere in Africa. The entire “multiheaded” network was orchestrated and partially funded by the Acumen Fund, a hybridlike venture capital firm that helps launch organizations with potential for strong societal returns (traditional venture capitalists seek strong financial returns).

Got all that? (Catch your breath.) The main point is that a business genome would be capable of describing this hybrid structure, even if we quibble about whether what we’ve got is really a hybrid, “partial dominance,” or even a chimera (a rare but real occurrence in nature where the genetic materials from different individuals are fused, allowing you to share your fraternal twin’s DNA!).

We can even conceive of a business chromosome robust enough to describe a complex networked organization involving multiple parties to prevent malaria in Tanzania (described in the box). And because we can describe such a hybrid organization and see how it could grow from less developed parts, we can also imagine how similar approaches might be developed to address world hunger, poverty, lack of education, and so on. Once we free ourselves from thinking only about “what is,” we can begin to imagine powerful new approaches to deeply entrenched problems.

What qualities do successful hybrids possess? They can have traits more desirable than their parents (hundreds of advantages per fruit for seedless watermelons). They can be well adapted to changing environmental conditions (hybrid Galapagos finches’ “inferior” beaks became an advantage after a storm in the mid-1980s introduced new plants to which their beaks were extremely well suited). And they can be very long lasting (like the hybrid wheat that evolved millennia ago and today occupies much of the world’s farmland).

But what does a successful societal hybrid look like? Such an organization should, of course, help solve tough societal problems. It can be the proverbial better mousetrap. By combining the traits of two successful organizations, it can compensate for the flaws or deficits of either one. In nature, the resulting improvements are sometimes called “hybrid vigor.” If in nature, why not in organizations? Hybrid solutions to societal problems can be essential, too, in times when societal “storms” are gathering and threatening the environment or our social order.

Evolution Under the Covers—Sexual and Otherwise

The same evolutionary forces that operate in the natural world operate in the world of organizational genetics. We can interbreed two crops; we can also interbreed two organizations. Mutations introduce variety in the natural world and in the world of organizations as well. Genetic engineering? Though GMOs currently stands for genetically modified organisms, how about genetically modified organizations, too?

There is vast expressive capability within a genome. There are some three billion DNA molecules, or nucleotides, in the human genome. By altering just 4 percent of this structure, we end up with the genome for a chimp instead.

In terms of organizational genetics, we see the same potential. First, by making small changes in organizational architecture, we can produce large changes in outward appearance or performance (as with chimps and humans). And by changing a little bit more—even just 5 to 10 percent—the changes can be that much more pronounced.

Because both organisms and organizations possess underlying genomes, both can evolve over time. But the processes by which they evolve differ in important ways. The hallmark of biological evolution, of course, is “organic” change. Sexual reproduction mixes up the genetic material of two parents, producing offspring different from either one. In certain situations such as plant breeding, parents might be selected with the express purpose of producing a certain kind of offspring. For instance, a high-yield (or high-nutrition) variety of vegetable might be crossed with one that can better withstand the elements in the hope of providing very-low-income farmers living in harsh conditions with a better food source. Whether the exchange of genetic material takes place “au naturel” or as the result of intentional crossbreeding, the phenomenon is one that is most clearly revealed empirically. Plant-1 (a tall red flower) + Plant-2 (a short white flower) = Plant-3, a plant whose height (tall) and color (pink) we can easily observe. The changes taking place “under the covers,” so to speak—changes taking place at the structural level of the genome—remain hidden from view.

Organizational evolution, in contrast, can be far more targeted and precise. Two companies don’t just casually “mate,” hoping their union produces cute little baby companies. Instead, it is as if corporations have state-of-the-art technology for analyzing and sequencing DNA. If we examine an organization with great care, we can see the “instructions” that it follows. What are its hiring policies? How much does it spend on research? Does it have a chief technology officer? Which technologies does it use, and how does it use them? How much money has it raised by issuing stock? Or bonds? In which countries are its products manufactured? Does it own these plants or outsource the responsibility? The list of instructions is huge, but it is finite. By reading about an organization, doing business with it, talking with past and present employees, and carefully studying it in many other ways, we can obtain a very detailed understanding of the instructions that dictate how it will behave in the real world. We can discover (with considerably fidelity) its “genome.”

Once these genetic instructions are understood, they can be adopted by another organization. The small, local bank in Minnesota that we talked about earlier can therefore study the “instructions” for other organizations and decide which ones it wants to adopt. By identifying and adopting instructions that allow it to interact with banking clients over the Internet and to make small loans to the developing world, the bank can evolve into a microfinance institution.

At this point, you might be having some objections or at least some questions. Are an organization’s genomic instructions really written down? Are they known completely? Can another organization really adopt them?

The answers to these questions are related. Think for a moment about technology. Only a bit more than a decade ago, the Internet was dismissed as a toy or a fad, even by big companies such as Microsoft. Companies that even dared to mention their own websites on television still spelled out their complete, clunky web address: http://www-my_company_name.com (usually saying “backslash” when they meant “forward slash,” but hey!). Today we just assume that all companies are on the Internet, and we can guess their web addresses or use Google to find them in an instant. But back then, just a few organizations were marketing over the Internet, and it seemed to be working. When this was observed, it was imitated and eventually became common practice. You see the same phenomenon today with blogs (few people ask what they are anymore) and comment areas where customers can review a product’s features. Twitter, anyone?

Think about organizational structure. What do an organic food co-op and a street gang have in common with a large Wall Street firm? Nothing? What about the fact that all of them have a set of officers who direct the organization, all have a board of directors, and all have bylaws that govern their operation? Each of these things was at some point “new,” but over time, it was observed, inspected, and found to be useful to the point that we more or less take it for granted. In truth, these successful “genetic instructions” were simply being widely copied and adopted across quite different segments of society.

Or think about workflow, the way a company gets its work done. Fairly recently, U.S. companies discovered they could have certain nonessential activities performed more cheaply overseas, especially in India. They began to outsource operations such as payroll, record keeping, and customer help-desk support. After this came slightly more ambitious efforts. Doctors in the United States, for instance, could voice-record medical notes and send them overseas, where they could be transcribed cheaply and effectively before being shipped back to the doctor’s own office. And if medical record keeping could be shipped overseas, why not aspects of medical practice itself? We have begun to see this. Massachusetts General Hospital in Boston now has licensed radiologists in Bangalore, India. They receive X-rays and other digital images over the Internet, interpret them, and return their findings to the United States. Is it a coincidence that help-desk services, medical transcription, and now specialized practices in medicine are all being outsourced? Of course not. Bits of DNA from one activity (and firm) are being injected into another.

Companies are always trying to steal other companies’ DNA. By night, they may pick through their competitors’ garbage. (Really. Yuck.) By day, they “benchmark.” Companies such as Toyota and L. L. Bean are as well studied as the Talmud. Toyota’s DNA has genes that make it “lean,” and other organizations—whether they manufacture a product or run a hospital—want to acquire them. L. L. Bean’s ability to get orders out the door is legendary, prompting companies such as Gillette and Chrysler to study and imitate its order fulfillment (to help them get things out their own doors), but also the New York Times (which studies Bean to “fulfill” its customer service requests). Yet best-in-the-world companies do benchmarking themselves. When Bean found that it couldn’t keep up with Internet speeds and volumes using its mail-order methods and facilities, it studied others it considered on the electronic cutting edge. Companies want genes that fit well and make them look good, no buts about it.

And now companies have begun benchmarking environmental activities. Companies that never knew the difference between Styrofoam and compost suddenly want to be green. There are green standards for them to follow (Leadership in Energy and Environmental Design—LEED—for example), and environmental benchmarking companies to advise them. And then there are leading organizations, including Interface (textiles), TerraCycle.net (can you guess the main ingredient in its Worm Poop lawn fertilizer sold in a recycled pop can?), and Oberlin College (uber-environmental) that are inspiring other organizations to change their colors . . . to green, naturally.

So, you’re starting to be convinced that there are organizational genes that are propagated throughout society. (Some might call them memes, which means “ideas that spread.”) And these are really just instructions that other organizations can follow. But can these instructions really be known completely? No, they can’t. There are trade secrets, of course, and certain organizational intricacies that may be impossible to untangle perfectly. But then again, there are no perfect, written rules for how to speak English either, and we seem to get along fine. We have no trouble understanding the sentence “I downloaded the MP3 file to my iPod over Wi-Fi,” even though four of its terms are so new that they aren’t in most dictionaries. Instructions don’t need to be perfect for us to interpret and then follow them. And, as we know, Making It Appropriate is vital to successful performance—no less so when we are talking about taking another organization’s ideas and tweaking them to fit your own.

In fact, the search for ideas that make your own organization perform better is really a search for effective building blocks. These building blocks are strands of corporate DNA that have been put to the test and shown to work in a variety of circumstances. Most animals have eyes (even though evolutionary forces led to the eye being “invented” independently forty times, because eyes are good building blocks that themselves contain other smaller building blocks such as lenses and retinas). The idea of building blocks is found throughout nature, even down to the level of the cell, where the Krebs cycle, for example, is an emergent and vital chain of chemical reactions supporting all living cells that use oxygen to “breathe.”

Recap: all companies are trying to invent new and better ways of doing things, copying others who they think have good ideas, and being copied by others who think that they, too, possess a “secret” (even “secret” with a very lowercase s). Question: What is the net result of these simultaneous attempts at getting better through trial and error and imitation? We turn to that question next.

Complexity and Emergence

A new, multidisciplinary approach to science is emerging: complex adaptive systems. Complex adaptive systems address problems that seem to have very little in common: How do fads suddenly emerge and just as suddenly disappear? Why do ants prefer one food source over another that is equally appealing? How do our brains work? How does the economy organize itself when no one is in charge? What unites these questions is the idea that relatively simple, easy to understand interactions among small elements of a system produce powerful and very surprising emergent effects.

One of the techniques in the arsenal of a complex adaptive systems scientist is the genetic algorithm. An algorithm can be thought of as a “recipe” that can direct a computer or any other individual or system that is capable of interpreting and then acting on instructions. As the name suggests, the genetic algorithm is based on biological evolution. As we will see, the genetic algorithm produces changes in the direction of increasing “fitness.” But “genetic adaptation” does not—repeat, not—only describe changes among living things. Any system that involves copying and that has a tendency for variation that can differentiate (short-term) winners from losers is a candidate. This even includes rocks, crystals—possible precursors to DNA—and certainly organizations. Any instructions that are specified in sufficient detail to be carried out reliably are subject to adaptation. The genetic algorithm allows us to examine what happens when we pit a set of instructions against each other.

Genetic Algorithm for Comparing and Improving a Set of Instructions

  1. The algorithm operates on a set of instructions, each designed to address the same “problem” or task. Each instruction can be thought of as a long list of features, a structure identical to a chromosome, that determine its “owner’s” genetic makeup. Equivalently, an instruction may be regarded as a very long and detailed set of directions—not a single step—just as directions for driving from point A to point B contain many details. Let the instructions “duke it out” by letting them guide work on the same problem or task. Evaluate the effectiveness of each competing instruction by using an appropriate means of measurement. For instance, if the instructions are supposed to guide you to a destination, their effectiveness may be judged by how close or how quickly they get you to it.
  2. Select some of the instructions that are performing the best.
  3. Reward these best-performing instructions by making copies of them. Some aspects—features—of these best-performing instructions must be above average; otherwise, the entire instruction wouldn’t be among those chosen.
  4. Create new (never seen before) instructions by letting the existing set (including those created in Step 3) “mate.” Mating involves creating a new instruction from two “parent” instructions by combining some of the sub-instructions of each parent. The new instruction may not make a lot sense. What’s worse, there is a two-way praying mantis syndrome at play, for once mating takes place, both parent instructions are history.
  5. Repeat Steps 1 through 4 many, many times. For convenience, think of each repetition as another “generation.”

sequence

The illustration above shows two “instructions”—each a string of ten symbols—before and after they exchange instructional sub-sequences.

  1. Evolution takes place in all kinds of systems, biological, mechanical, social, and more.
  2. A common mechanism, or algorithm, describes this evolutionary process—a “genetic algorithm,” so to speak. The mechanism is powerful and robust. When we say organizations can and do evolve, we mean this more literally than metaphorically.
  3. The genetic algorithm makes improvements over time by “tinkering” with the appropriate instructions (“DNA”) to produce better crops, vehicles, organizations, and so on.
  4. The “residue” from generation to generation is a set of successful building blocks, which tend to get more complex, perform better, and become more and more common over succeeding generations.

Brief Overview of What Genetic Algorithms Do

What happens when we let the genetic algorithm operate on a set of instructions? To start, in each generation the instructions that are performing better than the rest (Step 1) are identified (Step 2) and replicated (Step 3). But these instructions are not preserved; they are broken apart and combined with other instructions (Step 4).

As this process occurs again and again across many generations, the “duking it out” step usually becomes a competition among new instructions that never existed before. After all, they came from a “mating process” involving their “parent” instructions. But the building blocks of successful instructions can endure and be passed on from generation to generation. If you mate instructions for building a bicycle with instructions for building a lawn mower, you may end up with something resembling a motorcycle—a kind of motorized bike.

In the next generation, even though the original “parents” won’t be around anymore to cut grass or be ridden safely by six-year-olds, the building blocks of both the lawn mower and bicycle may be preserved in the new gizmo. The bike might contribute its frame and tires; the lawn mower might contribute its two-stroke motor. These same building blocks (by being parts of new configurations) now participate in the duking out/selecting/rewarding/mating cycle, even though the original, specific “parents” that produced them have been tossed aside on a heap of obsolescence. (Is this the way Harley-Davidson gets its inspiration? Probably not.)

So, why should we pay any attention to this process at all? Important question, to which there is an important answer. This model describes how any system evolves to better meet the demands (or expectations) of its external environment. And the “emergent” nature of what evolves means that what we end up with may be something quite different—and much more intricate and well suited—than what we began with. If the instructions we are considering (Step 1) are biological chromosomes and not instructions for building physical machines, then we have a competition in which individual biological organisms (think different varieties of wild grasses) might “duke it out” (Step 2) by competing to occupy the same parcel of land, fighting over available water, soil nutrients, and so on. Varieties of grass that best adjust to these conditions (the “environment”) will predominate (Step 3), more successfully reproducing and coming to occupy a greater percentage of the land over time. Occasionally, however, the pollen from one variety may fertilize another, giving rise to unexpected variation (Step 4). There is nothing that automatically confers an advantage to this new variety, but it has won the right to duke it out in the next generation. Its performance “against” other grass varieties, and its ability to replicate (Steps 2 and 3) depend on how well it takes advantage of the physical environment.

Muddying?—No, Generalizing—the Picture

I understand that right now you might be saying, “Yes, but what you’ve described resembles ordinary biological evolution. You said that the model would explain how any system evolves and improves.” Let’s address that issue by looking at a really tricky question: Heredity and biological evolution require the availability of DNA, but DNA is an incredibly complex organic molecule that, itself, must have evolved. So, where did DNA come from?

Amazingly, some scientists think that DNA could have emerged from an evolutionary process in which different kinds of clay and mud played the starring role. Clay and mud contain crystals that line up to create tightly packed, repeating patterns. In fact, two or more different kinds of crystals, with different atomic-level patterns of molecules, are capable of being formed from the same mud or clay. And once they form, each type of crystal tries to “replicate,” or grow, by gathering additional material from the clay or mud and attaching it to form an ever-growing crystal. So, replication using clay or mud is indeed possible.

So is a variation. Although each type of crystal is defined by a unique pattern, imperfections arise, too. As Richard Dawkins, from whom I am borrowing this metaphor, describes it, if you examine a crystal with a sufficiently powerful microscope, you’ll see something like a repeating herringbone pattern. But every once in a while, the pattern might twist, veer off at an angle, drop a stitch, or insert something extraneous. These defects appear unpredictably and unexpectedly, and that means they can convey information. Imagine that you are sitting at home and your water pipes are tapping out a rhythm with the regularity of a metronome. If, unexpectedly, this pattern were replaced with the sound of “dot-dot-dot d-a-s-h d-a-s-h d-a-s-h dot-dot-dot,” you might be tempted to conclude that  someone was trapped in your water heater and sending you an SOS to help them escape. (Please don’t blame the brilliant science writer Richard Dawkins for the SOS part of what I’ve just written. I loosely borrowed his herringbone description and his scientific descriptions of how clay or mud could be the predecessor of DNA. But someone caught in the water heater? That’s mine.)

So, replication is possible with mud and clay crystals. And now we see that variation is, too, meaning that variations can be considered to contain instructions or convey information. But what about the proverbial “survival of the fittest”? How do clay and mud compete (“duke it out”)? And what does it mean to say that one type is performing better than another? Even here, certain muds and clays have the “right (though messy) stuff.” Some are better than others simply because they win at the game of survival by better reproducing themselves. This is the sense in which insects far outperform humans in terms of total biomass. In this way, they lead us by a wide margin in the ultimate game of survival.

How can a particular type of clay or mud win at this game versus other grimy contestants? Possibly by affecting the flow of the very rivers or streams from which it came, altering the rate of flow so that the right kind of building blocks for creating its own type of crystals are more likely to beleached from the riverbed and deposited into the stream. More of the right kind of building blocks, more chance for the right kind of crystals to form, and more chance of a particular type of mud or clay “winning.” (Of course, there is no “intention” to win here, but it is not needed.) This particular clay or crystal might gain further advantage if it is flaky enough to dry up occasionally and be blown to other streams when things get very dry, or if it is “crafty” enough to cause its host river to split into diverging tributaries that spread in different directions. In either case, the “winning DNA” (excuse me, mud or clay) is beginning to slime the competition. Eventually, these inorganic clay molecules can even selectively support the “success” (survival) of certain carbon-containing organic materials, thus taking another important step toward the birth of (organic) DNA.

We’ve seen small Minnesota banks evolving into microfinance institutions, food co-ops and street gangs borrowing ideas from sophisticated businesses, and bicycles mating with lawn mowers; and we’ve just taken a look at how DNA itself might even evolve from river clay or mud. In each case, the adaptations obey the logic of the genetic algorithm (although I’ve presented only the trailer of the algorithm, not the whole movie, in my descriptions).

If we were to watch the entire genetic algorithm movie, including the context in which it operates, here is what we would see. Individually, different “organisms,” which we can alternatively view as different “instructions,” are fighting for their survival. To survive means to fare well at the task or problem at hand so that the organism participates in the rounds of replication and mating. In this sense, our bicycle and lawn mower must have been performing well enough against other contenders to be selected for replication and mating. How will our lawn mower–bicycle hybrid do in the next generation? If the task at hand is providing a means of transportation, it will probably be a hit, fare well in the genetic competition, and continue to pass on its own “genes” (which, of course, it got from its parents).

But as we zoom out with our cameras, we see that the action is really taking place among an entire set of characters and unfolding over time. Although our lawn mower–bicycle–motorcycle contraption will continue to evolve (unless its lineage comes to an end—“I knew it wasn’t a good idea to ride some sort of screwy lawn mower on the highway”)—it’s more instructive to look at the competing set of instructions (or organisms) duking it out generation to generation. On average, we’d see two things. First, certain features—and even combinations of features—would get more and more common over time. The same wheel that provided a performance advantage to a bicycle found its way to the hybrid bicycle–lawn mower, where again, this new means of transportation would likely be superior to anything with square wheels, triangular wheels, or no wheels all. So “wheel” emerges as a building block. Second, because it is the effective combination of features that gets more and more common, we would see performance improvements (again, on average) generation to generation for the entire set of instructions in force. That is, for an instruction to be duking it out in generation 1,000, it is likely to be a better performer than instructions in force 999 generations earlier.

Let’s step back from the clay and mud, broken bicycle chains, lawn mower oil, and community banks that have been reborn as microfinance institutions to recap (Yes, I’m repeating content from the previous box I used to preview what was to come. Put your hand up if you really read it.):

  1. Evolution takes place in all kinds of systems, whether biological, mechanical, social, or rivers containing mud and clay.
  2. A common mechanism, or algorithm, describes this evolutionary process—a “genetic algorithm.” The mechanism is powerful and robust. When we say organizations can and do evolve, we mean this literally, not metaphorically.
  3. The genetic algorithm can make improvements over time by “tinkering” with the appropriate instructions (“DNA”) to produce better-performing vehicles, grasses, mud, or organizations.
  4. The “residue” from generation to generation is a set of building blocks, which tend to get more complex, perform better, and become more and more common over succeeding generations.

In the next chapter, we’ll see that we can use the genetic framework I’ve described to create hybrid solutions that help produce a better world at the societal level. And by understanding the genetic mechanism, we can dramatically speed up evolution in the processes.