This article is about a number I call the optimal crucible size. I’ll define this number — call it C — in a bit, but I believe its value to be around 12. This article is also about an argument that I’ve been unconsciously circling for a long time. Chris Anderson’s Free provided me with the insight that helped me put the whole package together: economics is fundamentally a process driven by abundance and creative-destruction rather than scarcity. The reason we focus on scarcity is that at any given time, the economy is constrained by a single important “bottleneck scarcity.” Land, labor, factories, information and most recently, individual attention, have all played the bottleneck role in the past. I believe we are experiencing the first major bottleneck-shift in a decade. “Attention,” as an unqualified commodity is no longer the critical scarcity. Collective attention is: the coordinated, creative attention of more than 1 person. It is scarce and it is horrendously badly allocated in the economy today. The free-agent planet under-organizes it, and the industrial economy over-organizes it. That’s the story of C, the optimal size of a creative group. There are seven other significant numbers in this tale: 0, 1, 7, 150, 8, 1000 and 10,000. The big story is how the economy is moving closer to C-driven allocation of creative capital. But the little story starts with my table tennis clique in high school.
A Table-Tennis Story
R and I played table-tennis nearly every day in high school. We were regular partners in a loose clique of serious players at our club, comprising approximately a dozen players. The score in nearly every 3-game match would go something like 21-14, 21-7, 23-22. It wasn’t that I was getting creamed every time; I’d occasionally take a game off R. He was only slightly better than me, in just about every department, but that all added up to him beating me nearly every time. He knew his strengths (defense/offense, forehand/backhand) enough to always pick a better strategy for each game. He selected his shots better and executed them better. The net result was that I was beaten mentally and physically. Errors would accumulate, and I’d invariably choke.
Then one day, I managed to convince S, whose father had been a state-level champion, to practice with me (there was no point playing, he would have beaten me 21-0, 21-0, 21-0). S was the sort of calm, unflappable guy who simply cannot be psyched-out or forced into error. He had an almost robotic level of perfection in all basic elements of the game. S put me through half an hour of very basic forehand-to-forehand top spin practice rallies, and it completely changed my game. After that, I still mostly got beaten by R, my regular partner (who was fundamentally more talented than me), but I actually began winning the occasional match, and all games were a lot closer.
Fast-forward 15 years. At the University of Michigan, I organized an informal tournament at the residential scholarship house I was living in at the time. Out of the field of about 8-10, I came in second. Most Americans in the house fared as well as you’d expect; since they view “ping pong” as not really a sport, most of them lack basic skills. I beat most of them relatively easily, but was beaten pretty handily by a Korean-American guy.
A final data point. About 2 years ago, with rather foolhardy confidence, I joined in a Saturday afternoon group of serious Chinese players. The result: I was beaten comprehensively by everybody. In particular, by a bored, tired-looking 14 year old (clearly first-generation) who looked like he hated the game and had been dragged there by his immigrant father.
Collective Attention and Arms Races
Now step back and analyze this for a moment. Table tennis is primarily information work. It is not among the more physically demanding games except at the highest levels. My serious table-tennis clique in an apathetic-to-the-game country, with a lousy athletic culture (India) got me to a certain level of competence: enough to beat many casual players in a vastly more athletic country (the US). But a disengaged kid from the diaspora of an athletic country that is crazy about the game (China) was able to beat me with practically no effort, despite being far less interested (apparently) in the game than me.
This little story captures the most essential features of collective attention. It exists at all scales (from small clique to country to planet). Within a group that is paying coordinated attention to any information-work domain, skill levels rapidly escalate, leaving isolated individuals far behind. I call this the arms race effect, and it is a product of a fertile mix of elements in the crucible: competition, mutual teaching, constant practice and sufficient, but not overwhelming variety. This is a very particular kind of attention. It isn’t passive consumption by spectators, and it isn’t performance for an audience. It is co-creation of value: that same dynamic that is starting to drive the entire economy, blurring lines between producers and consumers.
So our challenge in this article is to answer the question: what is the optimal size of a creative group? Is country level attention the best (China and table tennis) or clique (my high school)? Is it perhaps 1 (solo lone-ranger creative blogger)? Our quest starts with the first of our supporting-cast numbers, 10,000. As in the 10,000-hour rule studied by K. Anders Ericsson and made famous by Gladwell in Outliers.
10,000 Hours and Gladwell’s Staircase
Gladwell is a jump-the-gun trend-spotter. He nearly always finds a uniquely interesting angle on a subject, and nearly always analyzes it prematurely in flawed ways. That’s a story for another day, but let’s talk about his latest, Outliers. The basic thesis of the book is that there are all sorts of subtly arbitrary effects in the structure of nurture (Gladwell’s way too smart to play up a naive nature/nurture angle) that make runaway success a rather unfair and random game of chance. In particular, Gladwell focuses on a key argument: that to get really good at anything, you need about 10,000 hours of steadily escalating practice, with opportunities to “take your game to the next level” becoming available at the right times. For instance, due to some weird cutoff-date effects, nearly all top Canadian hockey players are born in winter (thereby, Gladwell implies, unfairly penalizing burly talents born in warmer months). This basic argument is just plain wrong for the simple reason that no human talent is that specifically matched to particular arbitrary opportunity paths like hockey. No talented human being is starkly “hockey star or schmuck.” There are presumably other things demanding strength and athletic ability available in Canada and other parts of the world, that have no winter bias (or perhaps, complementary summer biases). As Richard Hamming put it eloquently in his famous speech at Bell Labs, You and Your Research, “There is indeed an element of luck, and no, there isn’t. The prepared mind sooner or later finds something important and does it. So yes, it is luck. The particular thing you do is luck, but that you do something is not.”
But that said, Gladwell is on to something. The pattern of increasing opportunity stage-gates he spotted is real, but most of the arbitrary effects he talks about (being born at certain times, your university having one of the first computers, and so forth) are red herrings/minor elements that confuse the issue. But one effect is not a red herring, and that is the fact that the staircase of opportunity puts you in increasingly intense crucibles of collective co-creative attention.
The Distillation Effect
Start with 1728 (12^3) people and let them learn widget-making in144 groups of 12, for 3000 hours. Then take the top talent in each group and make 12 groups of 12, and again let them engage in an arms race for 3000 hours. Then take the final top 12 and throw in another 4000 hours. With two levels of distillation, you’ve got yourself a widget-making dream team. Or a fine scotch. A team that will be leaving the remaining 1716 far, far behind. You can watch this process accelerated and live today on America’s Got Talent and American Idol. Imagine the same process playing out more slowly over 20 years. What does that transformation look like?
That is what is scarce. Collective attention. That’s what creates the 10,000 hour staircase-of-opportunities that Gladwell talks about. Information may want to be free, but live attention from other humans never will be (AI is a different story).
A note of irony here: Gladwell was also among the first to stumble across the importance of such dream-team crucibles, in The Tipping Point. Today, researchers like Duncan Watts have pointed out that viral effects don’t necessarily depend on particularly talented or connected “special” people (the sort Gladwell called “mavens” and “salesmen”). But “special” people do have a special role in shaping culture. It is just that their most important effect isn’t in popularizing things like Hush Puppies, but in actually creating their own value. New kinds of music, science, technology, art or sporting culture.
This is the signal in the noise, and here is the lesson. Information work in any domain is like weight training: you only grow when you exercise to failure. The only source of weight to overload your mental muscles is other people. And the only people who can load you without either boring you or killing you are people of approximately the same level of talent development. And that leads to the question: what happens when you hit the top crucible of 12 in your chosen field? Where do you go when there are no more levels (or if you’ve reached the highest level you can, short of the top)? That brings us to the next two numbers in our story: how you innovate and differentiate as a creative.
1 Free Agent and 1000 Raving Fans?
I’ve hated the phrase “raving fan” since the day I heard it. If you are not familiar with the argument, Kevin Kelly, who originated the idea, claims that an individual creative — blogger or musician say — can scrape along and subsist in Chris Anderson’s Long Tail, by attracting a 1000 raving fans who buy everything he/she puts out (blogs, books, special editions, t-shirts, mousepads; 1000 raving fans times $100 per year per fan is a $100,000 income). Kelly’s original adjective is a less-objectionable “true” rather than “raving” but “raving has caught on, and the intended meaning is the same.
This basic model of creative capital is just not believable for two reasons. First, it reduces a prosumer/co-creation economic-cultural environment to a godawful unthinking bleating-sheep model of community. I try to imagine my blog, for instance, as the focal point of a stoned army of buy-anything idiot groupies, and fail utterly. I would not want to serve such a community, and I don’t believe it can really form around what I do. I certainly refuse to sell ribbonfarm.com swag.
The second problem is the tacit assumption that creation is prototypically organized in units of 1. The argument is seductive. The bad old corporations will die, along with its committees of groupthink. The brave new solo free agent, wandering in the woods of cultural anarchy, finds a way to lead his tribe to the promised land of whatever his niche is about. “Tribe” is a related problematic term that Seth Godin recently ran amok with.
The reason Kelly (and others like Godin) ends up here is that he answers my question “after the dream team, what?” with “individuals break away, brand themselves and become individual innovators.” Kinda like Justin Timberlake leaving N’Sync. A dream team of 12, in this view, turns into 12 soloists. Not that he ignores groups, but his focus is on the individual.
Individuals vs. Groups
That’s not what happens. You cannot break the crucible rule. 12 is always the magic number for optimal creative production. The reason people make this mistake is because they draw a flawed inference from the (correct) axiom that the original act of creativity is always an individual one. I’ve talked about this before: I am a believer in radical individualism; I believe, as William Whyte did, that innovation by committee is impossible. Good ideas nearly always come from a single mind. What makes the crucible of 12 important is that it takes a group of competing/co-operating individuals, each operating from a private fountainhead of creative individual energy, to come up with enough of a critical mass of individual contributions to spark major revolutions. Usually that’s about 12 people for major social impact, though sometimes it can happen with smaller crucibles. These groups aren’t the deadening committees of groupthink and assumed consensus. They are the fertile, fiercely contentious and competitive collaborators who at least partly hate the fact that they need the others, but grudgingly admire skills besides their own.
What happens when you exit the dream team level in a mature disciplinary game is that you get out there and start innovating beyond disciplinary boundaries; places where there are no experts and no managed progression of levels with ritualistic gatekeeper tests. But you don’t do that by going solo. You look for crucibles of diversity, multidisciplinary stimulation and cross-pollination. But you still need the group of 12 or so, training your brain muscles to failure.
This gives me a much more believable picture. As a blogger, I am the primary catalyst on this site, but I am not creating the value solo. If I try to think of the most valuable commenters on this site, I can think of no more than 12. My best writing has come from trying to stay ahead of their expectations, and running with themes they originally introduced me to. But that’s far from optimal, since I still am the dominant creator on this blog. The closer I get that number to 12 via regular heavy-weight commeters, guest bloggers and mutually-linked blogroll friends (I’ve turned my blogroll off for now for unrelated reasons), the closer I’ll get to optimum. Think of all the significant power blogs: they are all team-acts. Now, I may never get there, and there’s multiple ways to get to 12, but the important thing is to be counting to 12. At work these days, I am pretty close to that magic number 12, and enjoying myself a lot as a result.
So the important number for the creative of the future is 12, not 1 or 1000. But what about money and volume? Don’t we need a number like 1000? Not really. As the creative class matures, you won’t really ever find 1000 uncritical sheep-like groupie admirers. That is a relic of the celebrity era. The real bigge- than-crucible number is not 1000 but 150. Dunbar’s number.
The Dunbar Number and $0.00
Why 150? That’s the Dunbar number. The most people you can cognitively process as individuals (the dynamics are entertainingly described in the famous Monkeysphere article). That’s the right number to drive long-tail logic. By Kelly’s logic though, I have to get to, say, 100,000 casual occasional customers before I find my 1000 raving fans (1% conversion is realistic).
Face it: there’s no way in hell most of us will get there. If I accidentally did, through this blog, I’d probably erect walls to keep the scary crowds out somehow. That picture makes sense for almost nobody. I write long, dense epic posts and don’t bother to be accessible. I look to attract readers who can keep up with me. Unapologetic intellectuals in fact, whose own eclectic interests overlap sufficiently with mine to create the right mix of resonance, dissonance and dissent. In terms of Geoffrey Moore’s classic pair of business models: complex systems (a few high-touch, high-personalization customers) and volume operations (mass-consumption stuff), this blog is a complex-systems play. I can (and have) written posts entirely with one reader-muse in mind. I have more chance of making a living off 100% of a base of 150 powerful micropatrons than from 1% of a base of 100,000. The question is: which is actually the right type of model for the individual creative (in a crucible of 12 similar-minded others; not selling to each other, but collectively representing a high-value-concentration crucible)?
I am going to make a prediction: personalization and customization will rule. Without that common prefix of the day, “mass-.” Mass customization/personalization is a good model for Enterprise 2.0, but individual creatives have a far better chance of creating an economically sustainable lifestyle by paying close individual attention to 150 people than by selling the same thing to 100,000 and hoping 1% of the sheep convert to your religion. This isn’t to say that volume games can’t succeed. But it isn’t the way most people will succeed, because the numbers will not add up. Can you really imagine a significant proportion of the world’s information worker/creative class being able to draw 100,000 unique visitors per month to their blogs, most of whom will be other creatives trying to build their own 100,000/1000?
The “100,000 base” argument can be safely ignored for most of us. And that’s what I’ve done to most of the 62,000 unique visitors Google Analytics tells me have visited this blog since I opened up shop in July 2007. An overwhelming majority of them bounced away before I could even say “Hi!” Some read one article and never came back, leaving only an IP address behind. In an age where superhits and celebrities are on their way out, that’s what any crowd of ~100,000 will do. Your actual goal as creative today is to find and keep your 150, to whom you pay individual attention. Pass-through crowds don’t deserve much attention. In fact, the monetary value of your transaction with them is exactly $0.00. Anderson hammered home the point that to the masses, the right price for your work is $0.00, but he didn’t address the flip side. They are also worth only $0.00 to you on average. Which means you should put no marginal effort into pleasing them. If one of them finds something you did for your 150 useful, let them have it. You get paid in word-of-mouth, they get free stuff. Small serendipitous barter transaction. Aggregate over 100,000 and net hard-dollar value is still 100,000x$0=$0. The barter is non-zero sum, but doesn’t pay your rent.
Personal Economic Neighborhoods
By carefully curating your Dunbar neighborhood of at most 150 (in practice, likely much less), in collaboration with your crucible of 12 (each curating their own 150-neighborhoods, with a good deal of overlap), through actual personal attention, you create the foundation for your life as a cultural creative and information worker. Free agency is an important piece of this, but don’t dismiss traditional economics: a good part of your 150 is likely to remain inside the formal organizations you are part of.
The Kelly number, 1000, is important, but not in his sense. If you and your crucible of 12 are creating value in a loose coalition, and each have a 150 circle with some high-value overlap, the total is probably near 1000. So that’s 12 people sharing a community of 1000, each of whom gets personal attention from at least 1 of the 12. The members of the 1000 get the overhead savings of finding more than 1 useful, personally-attentive creator in one place.
Count the 12 most valuable co-creators you work with. Now consider the overlap in your Dunbar neighborhoods. If the average level of overlap isn’t in the double digits (the actual set-theoretic math is tricky), you probably haven’t reached critical mass yet. Guess where you can still find such critical mass today? Inside large corporations. Any pair of people in my immediate workgroup of around 12 can probably find 20-30 common acquaintances. Our collective personalized-attention audience at is probably around 1000. Large corporations still allocate collective attention pretty badly (they hit the numbers, but get the composition wrong), but still do a better job than say, the blogsphere. But the free-agent nation is catching up rapidly. The wilderness is becoming more capable of sustaining economics-without-borders-or-walls every day.
So how will you create and monetize your Dunbar neighborhood? By definition, there are no one-size-fits-all answers, because the point of working this way is that you’ll find opportunities through personalized attention. Not a great answer, I know, but still easier for most of us than dreaming up ideas that can net 100,000 regulars of whom 1000 turn into raving fans.
8: The Maximal Span of Control
We’ve argued that the optimal crucible size must be greater than 1 and less than 150, but we still haven’t gotten to the reasoning behind 12 rather than 30 or 5. Another number will help get us there: 8, the upper end of the range of a number known as the span of control. The number of direct reports a manager can effectively handle, and still keep the individualized calculus of interpersonal relationships tractable.
What happens when you exceed the span of control? You get hierarchies. You cannot organize, complex coupled work (think space shuttle) requiring more than 8 people in a flat structure. But here’s the dilemma: between 9 and 15, if you split the group into 2, you may get high overhead and micromanagement by managers with too little to do, and other pathologies. So between the limit of a single manager’s abilities, and the optimal point at which to force cell division, ontogeny and organization, you get a curious effect: the edge of control. Single-manager structures fail, but team chemistry can take over. The whole thing is just barely in control, and teetering on chaos.
Should sound familiar. Those are the conditions, complexity theorists have been telling us for decades, that spark creative output. More than 8, less than 16. Why 12, besides being a nice mean? Anecdotal data.
The Ubiquity of 12
I hope you are too smart to conclude that I am making 12 a number of religious significance. It is simply the mean of a fairly narrow distribution. Still, it turns up in a surprising number of “creative crucible” places in practice:
- The dirty dozen (alright, there were also the 7 samurai)
- Juries (creative judicial decision-making)
- Teams in cricket and soccer (~12)
- The number of apostles required to start a major religion
- The approximate size of famous cliques of mathematicians, scientists, engineers, philosophers, writers and so forth.
- Ideal class sizes in education
- G-8, G-12
- Ensemble casts (Friends, Seinfeld, counted with frequent regular side characters who appear often enough that you recognize them).
- Improv comedy groups (typical size of a generation of SNL regulars).
(I believe there is some research related to Dunbar number research that actually talks about how “small world” groupings where you really intimately know the others in the group tend to be around 12. All our craze for weak links has tended to distract us from the fact that the small world is small in 2 ways: the maximum distance on the social graph between 2 nodes being 6, and the fact that we cluster in small groups).
The Magic Number 7
Let’s get to the last of our big list of numbers. Seven. As in, Miller’s famous magic number, the number of unrelated chunks of information you can manage in your short-term memory at a given time, and a big implicit hidden variable in everything we’ve talked about so far. It’s why lists of 7 are effective. So let’s make up a list of 7 to summarize the key concepts so far.
- Collective Attention: a group of people paying attention to the same thing, with the group size varying in size from 2 to 6 billion.
- Arms Race: The effect by which groups paying collective attention to something force individuals within the group to rapidly improve their skills and separate the group from outsiders.
- Mental Exercise-to-Failure: The fact that only people close to your talent level can load your mind in ways that cause you to grow
- Crucible: The optimal-sized creative group. Stages of crucibles reach successively higher plateaus and culminate at the dream-team level, beyond which lies innovation.
- Innovation: The graduation of a creative from a dream-team level in a disciplinary game to a more diverse and unstructured type of crucible, with few rules.
- Dunbar personalization: The idea that you are more likely to succeed as a creative in the new economy by paying personal attention to up to 150 people, than by paying mass attention to 100,000 in hopes of harvesting a 1000 raving fans.
- Span of Control: The threshold group size for a crucible. Above this number creativity is possible. Below this, the group can be brought under the dictatorial and low-creativity control of a single individual.
That’s CAMCIDS for you acronym buffs. I know. I am a terrible person.
Managing Collective Attention Scarcity: The Dynamics of 12
A little paper math shows you why collective attention is scarce. Marketers will recognize a classic example: marketing sugary cereals to a kid, but closing the sale with Mommy, is much trickier than marketing and selling to the same person. You need the collective coordinated attention of both, and you need that attention to have “good” chemistry. But unlike individual-level attention scarcity and its complement, the myth of information overload, you cannot solve the collective attention problem through appropriate reframing and a good set of automated filters. Also unlike individual attention dynamics, the action isn’t in stuff like advertising, fame or too many emails. It is co-creation groups.
Individual attention economics was merely poorly managed, and is now technology exists to manage it well, even though it hasn’t diffused completely. Collective attention, even if it is optimally allocated, is fundamentally scarce, since it requires live people in the other 11 empty spots in the crucible.
Optimal allocation is hard because the numbers blow up in your face. A Dunbar community of 150 admits 11175 unique pairings. Most will be fights, divorces and toxic messes. A few will create great value. Searching that space for the Gates-Ballmers and the Jobs-Wozniaks is horrendously hard. That’s why it is a scarcity. Things do get simpler once a core of 2-3 start to attract enough to reach that critical mass of 12, but not by much. The math is much harder for more complex ways of thinking about groups, but you get the idea. Can you think of all possible ways to break down the world’s population of 6 billion into constantly shifting crucibles of 12?
But let’s at least work on the right problem.