Metcalfe’s “law” has become the normie basis for valuing social networks. Robert Metcalfe conjectured that the cost of a network is proportional to the number of nodes N, but the network’s value is proportional to the squared number of nodes N^2. In human, he figured that networks tend to exhibit massive economies of scale when connections to other users matter more than connections to a mainframe (example: telephone v. electric grid).
A common way to translate this intuition into a metric is to take some function of the possible unique directed connections N^2-2 as the “natural” upper limit to a network’s value. There also are many empirical ways to measure network connectivity all the way down to digraphs, cluster maps and node valences. The core use case for each of these valuation metrics typically is derivative of the insight behind Metcalf’s law. They rely on counting nodal links (even if those are repeat connections over time).
Such metrics and ideas made Google and Facebook possible – and dominant. They also put them to sleep in the bed of Procrustes.
Waking Up to Value
Counting nodes, clicks and likes is easy. Grasping value is difficult.
Remember print and TV? Using Metcalfe’s law to measure the value of a social network is the digital equivalent of print media’s charging advertisers on the basis of circulation and placement. Social networks which substitute nodal metrics (including “engagements”) for value will earn old media’s fate. Yet, that is the path of least resistance, and some are taking it.
https://platform.twitter.com/widgets.js
Social monetizes network value through “eyeballs” (clicks, impressions, purchases etc.) and “meta” (what the company knows about users’ habits and identities). Eyeballs are easier to monetize. They are an incremental improvement on old media (for example, Google’s charging for clicks on ads rather than for displays or impressions). Meta are more valuable overall. They are a game-changer which increases eyeballs’ revenue stream, helps build moats and commits users.
While eyeballs’ value is linear and one-off, meta’s value is exponential and recursive. An extra unit of time spent on the network increases the value of each previous unit of time. Or: an extra data point increases the value of each previous data point (for that user). (The more I know about a user, the more I can charge for that user, including by direct-selling user meta to a third party. The more I know about a user, the more the user commits eyeballs and meta, the more I know about the user…) You can easily see a future where social has virtually complete control on access to a user because of meta. Even a mainframe network can get that kind of control.
https://platform.twitter.com/widgets.js
Social’s uses of eyeballs and meta have a common foundation: they are relatively easy to “measure” or “capture”. Use value isn’t.
Captain Obvious: “Users can find a wide variety of value on social.”
A powerful frame for use value is “proactive versus reactive”. Proactive value is derived from “hunting” for the new, reactive value is derived from “herding” around tried-and-true concepts and dopamine paths. Herding makes the network measurable and monetizable, hunting makes it… alive.
https://platform.twitter.com/widgets.js
Yes, abstract categories are just mental constructs, and yes, this distinction is not easy to handle. Which is exactly why I will mostly show it rather than explain it.
Reactive drives the dopamine-fueled metrics behind eyeballs’ monetary value and the quarterly report to shareholders: likes, clicks, impressions. Reactive use(r)s are behind the “network” value metrics of the network à la Metcalfe’s law because they are the most numerous. On the reactive channel, dopamine is fed from the social network into the user.
Proactive is hard to measure even at the meta level because it is by definition out of the ordinary or established. Proactive helps build the network and increase user value by adding new types of content or new use cases for the network. For example, trolling is a “new type of content” that wasn’t feasible at scale before the advent of social. Trolling is an especially good example because it is perceived as inherently discomforting, but can be very entertaining and – heavens forbid! – someone might even end up learning something.
From the network’s perspective, proactive “hunters” are testers and innovators and reactive “herders” are test subjects and consumers of those innovations. The testing bit of this social structure can feel onerous and untoward. Most of us don’t like being in the position of guinea pigs. However, innovation suffers when testing is restricted. (Yes, this is another quote from Captain Obvious.) There are a couple of much less obvious and much more frightful consequences though.
First, restrictions (such as “fact-checking”, language policing and other forms of censorship) may have predictable and linear effects on testing behaviors, just as they do with regard to reactive uses. These effects are easily captured by eyeballs metrics. However, the effects on testing outcomes are nonlinear because disruptive innovation does not occur around the mean. Those effects become apparent to decision-makers only when it is too late.
The second hidden consequence is about how hunter types tackle changes in the environment. And this one is the real kicker. When they face restrictions, hunters will not just test less, but test less outlandishly because what use is it if you get banned or muted? This narrows the possible set of chance innovations much further than the measurable censorship filter. But hunters, by definition, see chancing as the driving value in social. So when you restrict chancing, the hunters don’t just stop testing – they move their testing elsewhere, to your competitor’s platform. When they leave, content innovation dies. Without content innovation, the herd leaves for greener pastures. Gradually, and then suddenly.
https://platform.twitter.com/widgets.js
Herding & Hunting
The herd is structured on the dimension of agreement/disagreement – about goals, definitions, protocols, knowledge, technologies, values. The hunt runs on the tension of obsolescence/innovation – in skills, practices, vantage points, heuristics, applications, intuitions.
The ultimate rite of passage into a herding community is about suffering through humiliation and other forms of credentialing (example: American college fraternities). The community values members’ taking on absolute loss just because the group demands it as a signal. Suffering ensures that the member has a strong identity (which makes conformity valuable), the humiliation/credentialing rite brands the group identity onto the personal identity. The herd’s survive & prosper strategy is “cohesion against adversity”. Never underestimate how powerful this strategy is.
The ultimate rite of passage into a hunting relationship is survival through pain and other potent reality filters. The relationship encourages members’ testing their own and each other’s limits and personal skills (example: US Navy Seals). Pain is a required byproduct of that testing (if you haven’t gone through a pain episode, you don’t have hunting cred), survival is reality’s passing grade for having developed the right abilities and combined them with the right relationships. The hunt’s survive & prosper strategy is “skill-testing against adversity”. Never underestimate how painful this strategy is.
https://platform.twitter.com/widgets.js
Herding is collectively fractal (herders are herded). Hunting is personally fractal (a hunter is hunted).
Herding is monotonic and convex in the number of sheep and the number of herders. Hunting is nonmonotonic, concave and discontinuous in the number of hunters and the availability of game.
Herding is about risk management. Hunting is about opportunity generation.
Herders apply the same standards and expectations to everyone – friends and strangers. Egalitarianism fosters cohesion, and the strength of the herd is in sticking together (herders coordinate not to overgraze pastures, to fend off predators etc.). Hunters are more exacting with family, friends, business partners and anyone deemed “high-potential” – higher standards apply to them than to the plebs. Filtering through repeated testing selects for hunting partners who are not just adapted to the current situation, but adaptable to changes in it.
Herders build master skills to avoid pain and loss. Hunters’ master skills are incomplete without ample pain and loss.
https://platform.twitter.com/widgets.js
A herder pens the same sheep tonight that were penned last night. The herder actively hopes to pen the same sheep. A hunter has to find new prey every time. The hunter must find new prey before starving to death or dying of hypothermia.
If the grass fails, the herder will have to lose some sheep and slaughter more than usual to get by. Herders’ effective response is to do more of the same. If the bison and geese don’t show, the hunter will go into the deep woods and into the deep waters for fish and elk. Hunters will even go to herding for a while if nothing else works.
Where the herder grazes cattle today, the hunter killed the predators some years ago.
Herders slaughter cattle. Hunters slay game.
The herder shaman sees the future on peyote. The hunter shaman seeds the future with blood.
Before jumping the shark on the herder/hunter frame, understand that the distinction is about emphasis, not exclusivity. We all do both. The insight is about the driving rationale for our decisions and strategies. One heuristic for distinguishing your frame for any one decision is to be clear about what you are seeking out. By making that decision, are you seeking out safety or are you seeking out opportunity? Is your question itself framed as safety-seeking or opportunity-seeking? How likely are you to get what you seek in the conditions you face?
The Social Network Lifecycle
To understand the lifecycle of social online, you need look no further than the typical lifecycle of an individual’s offline social network, and how the herding/hunting dynamic evolves over time.
When we are young children, we build our first networks around received structures – family, kindergarten, school. Even early signups for Facebook and Twitter happened through pre-existing networks like club memberships and referrals. At the childhood stage, although networking is rigidly structured, we can still enjoy a lot of chancing simply because it is a numbers game – when you change schools, neighborhoods etc. Everything is new, even the old and established.
The early days of online social are just like a child’s brain. You have a live exciting product and you have no flipping clue what to do with it. Every new stimulus (use case) is an explosion of excitement (new features, new content, new users). Stress levels are high and endorphin payoffs are stratospheric. You know that building the product is more important than monetizing, so you suckle off your angels (parents) until you figure things out. If you have good angels, they will help you structure your learning (make better use of experience) without restricting your chancing opportunities. The hunting mindset dominates amongst your founders, supporters and early users simply because the herd does not like this sort of environment and you are too small to get noticed by the collective.
Adolescence to early adulthood is when the chancing numbers game best combines with an appetite for deliberate risk-taking (if not opportunity-seeking). This is peak chancing. High school, college and early jobs is where the hunter mindset peaks in a typical human life. We shift “career” paths and skill sets in rapid succession, meet new people as we take different courses, build new business relationships as we look for new jobs/skills and business partners. The number and diversity of people we seek out and meet – some of whom we add to a sustained network – reaches a crescendo.
https://platform.twitter.com/widgets.js
On online social, a lot of the crescendo comes from herders flocking in after hunters have cleared the space and populated it with content. Numbers are positive for chancing, so hunters and hunting behaviors continue to proliferate. The herds make hunters’ testing strategies much easier to run in both time and scale. Use and content bloom. Use value (the value of your network to users) peaks. This golden age fizzles quickly and stealthily.
Into middle age, two big changes have happened. One is mechanical. People settle into the nine-to-five, get married, have children, mortgages, car loans, etc. All these life choices impose severe time costs and restrict one’s risk-taking budget. They cull social networks and make them more rigid. Because all of this appears to be progress – and a lot of people do have happy family lives, happy homes and happy lifelong friendships – there is little mental resistance to our herding instinct. The mind drifts with the herd until game is forgotten. The hunter mindset withers in the slow boil of the Indian summer, not in a volcanic eruption.
The other change is related to how brains work. With experience, our brains weed out more and more of the environment as insignificant noise. Our “significance” filters become much more constrictive, and often get clogged with generic noise (for example, not getting the latest phone that you can’t use to make phone calls). This narrows the range of chancing events we register and the depth of their penetration to our deliberative apparatus. We are blinded to ever more of the chancings we encounter and we make ever poorer use of the ones we do notice.
To naïve minds caged in semantic systems, curation features such as “fact-checking”, news filtering, ICYMIs, push notifications users can’t disable and speech policing appear to be progress because they may boost eyeballs in the short run and limit legal liability. Such antics can stimulate dopamine-fueled herding. Such effect is transient because of adaptation. The permanent consequence is that they kill the dynamics that make the network great in the first place. Curation is the death of chancing. Without game, hunters pack up for the remotest wilderness. Hungry hunters skulk about fringe forums designing decentralized systems to escape Big Social. Much of this is invisible to the eye until it is far too late.
The late days of the curated middle age feel like a permanently high plateau, often until long after the sharp descent has started. The fences that herders build can create a false sense of security that lasts long after the fences have become pointless. Big Social like Facebook and Google seems invincible behind its moats of network effects, logistics, regulation and content death squads. While we fence ourselves in, the world moves along on its path. Friends begin to die, marry or divorce away; children move out; spouses drift apart; longtime partners go out of business; entire industries and skillsets are rendered obsolete. Hunting skills having atrophied, old age is mired in phantasies of the good old days.
https://platform.twitter.com/widgets.js
Chancing, My Love
The more you get into the hunter mindset, the more you tend to value chancing. Hunting breeds an addictive self-reinforcing cycle jacked on adrenaline/cortisol bursts and an endorphin high. Pavlovian conditioning pulls the intoxicating effects forward to the chancing stage. The opportunity itself becomes the reward. You become more willing to invest time and energy into opportunity generation. The more and better opportunities you get, the more selective you become about the types of opportunities you are willing to entertain. To you, the value of chancing – opportunities that dropped into your lap without much effort – goes stratospheric. The shift from risk management to opportunity seeking turns chancing from a bonus sideshow into the main event.
Chancing is about things and ideas that you never knew existed nor would ever know how to look for. The typical chancing event is a failure of your network filter. The correlation is especially strong in adulthood, and applies even in the rare cases when the social filter is intended to encourage chancing. The vast proportion of chancing value comes from encountering things which run contrary to your existing understanding; which are emotionally “negative”; which are offensive, ugly, useless, harmful or otherwise toxic; or which are simply wrong. Products and entire business models are invented because a future founder was fired up by need or annoyance. Chancing is about noticing the blank space, and the dumpster fire hiding behind it.
Your chancing opportunity set grows exponentially with the number of nodes on the social network even and especially if that makes the network noisier. When you value time, noise is error, and error is where chancing is. Curation/moderation/categorization or any other form of penning or censorship restricts your chancing opportunity set. This is true by definition because the core of chancing value is about finding new/unpredictable relations where your existing filter didn’t expect them to be.
Death
When Facebook implements another of its clickbait gimmicks it does introduce noise into the system, but that noise is uniform because it draws on averaging algos and pedestrian data mining. White noise is like rain pounding on the foliage while you're attempting to track a doe. And even if water-snake videos are something new and annoying which spurs you to activity and creativity, that effect fades quickly. Your filter adapts to it even if it doesn’t brush it off outright. What had chancing potential yesterday is old news today.
This is why Facebook is dead to me as a user.
I was holding off for a special piece on the levels of knowledge/decision making one can operate on, but I was tempted to jump the gun by a fresh example of this concept being put to work by Nassim Taleb. It was just too juicy to pass up.
