by Jeff Stibel
To survive in a growth phase, you must own the market. That means putting aside profit, revenue opportunities, and anything else that could starve the network of needed oxygen. The goal is to consume all of the oxygen, which both allows your technology to grow and blocks any others from sprouting.
III
Anyone who drives in cities such as London, Los Angeles, or Atlanta must wonder why cities don’t increase the size of the highway system to reduce traffic. The answer is that highways are networks, and this means that bigger is not always better. While size matters during the early development of a network, eventually it becomes a hindrance. Solving the traffic problem by adding more lanes is an urban myth; it doesn’t really work. Writing in The Atlantic, Stephen Budiansky explained how scientists have observed “an interesting paradox about the flow of vehicular traffic.” When engineers and city planners add new roads or lanes, under certain circumstances, they may actually reduce the whole traffic network’s car-carrying capacity. “It turns out that the properties collectively exhibited by large numbers of cars moving over a network of roadways have many mathematical features in common with the behavior of other things that flow over networks, such as data carried by telephone lines and the Internet.”
The internet, just like the brain, can use TCP to slow traffic when things get congested. But there really is no parallel for highways. Cities have tried all manner of ideas: adding lanes, creating tolls, adding ramp stoplights. Nothing has relieved congested highways during rush hours.
Recently, the city of Stockholm tried a novel approach to deal with its traffic congestion problems: they started charging drivers a tax when they drove on certain severely congested roads at certain times. This was the first time someone thought of taxing driving at specific times. The idea was conceived in part by Jonas Eliasson, professor of transportation at Stockholm’s Royal Institute of Technology, and it was considered radical when first introduced. Most people in the community were violently against it. Yet it worked.
Almost overnight, there were 20 percent fewer cars during peak traffic hours. Twenty percent doesn’t seem like a huge decrease; after all, 80 percent of the cars were still on the road, and in many cases these hotspots were at old, narrow bridges, as there is lots of water to cross in Stockholm. It turned out, however, that removing 20 percent of the cars resulted in alleviating 100 percent of the congestion. It also had the unintended effect of increasing air quality, as smog must be evenly distributed throughout the day to avoid build up. Dr. Eliasson explained it this way: “Traffic happens to be a nonlinear phenomenon, meaning that once you reach above a certain capacity threshold then congestion starts to increase really, really rapidly. But fortunately, it also works the other way around. If you can reduce traffic even somewhat, then congestion will go down much faster than you might think.”
A city’s network of roads has a natural breakpoint—a point where the road can no longer handle the number of cars without causing traffic backups. When that breakpoint is exceeded even a little, the roads go from flowing smoothly to massively congested in an instant. This is no surprise because we’ve seen the same phenomenon with biological networks, whether it be ants, reindeer, Easter Islanders, or the brain.
While the traffic change in Stockholm was dramatic, the psychological effects, or lack thereof, were equally profound. Despite the fact that people either had to pay or change their driving habits, few complained. More than that, few even noticed. Those who had to pay were happy to do so because there was less traffic, and surveys showed that those who changed their behavior didn’t actually realize they had changed. The shift was so subtle that it went virtually unnoticed.
IV
Money is highly disruptive, even in small amounts. Against the backdrop of an overloaded network, money can be used to manipulate volume. For example, airlines constantly adjust ticket prices for air travel in order to keep the volume of travelers at the exact carrying capacity of their network. Too few travelers on an airplane and potential revenue is lost; too many travelers and the system is completely overloaded. Stockholm’s congestion taxes were small amounts—between 0 and 20 SEK (1 SEK is roughly .10 Euros or .15 USD)—that had low economic impact for most, but the introduction of these costs created a disruption that changed the network dramatically.
Adding money into the equation after a breakpoint can serve two purposes: it can reduce congestion on a network filled to the brim, and it can help companies convert their valuable networks from free to revenue generating. The former can be used effectively right after a breakpoint, but the latter should be avoided until the network is more stable.
Take Facebook as an example. There are many places where the Facebook network has now exceeded capacity, and money could be used to curtail this. We learned how there are currently too many friend connections on the site to be valuable. What if Facebook charged users for additional friends above a certain number? It could be a nominal amount, perhaps a penny per extra friend. The impact would likely be similar to that in Stockholm after the congestion tax. Some users would happily pay for the extra connections, but many would naturally reduce their list of friends to a manageable number. The net effect would be reduced congestion and a more efficient and more powerful network. Interestingly, in 2013, Facebook experimented with a similar concept, introducing a pay-per-message system. It offered people who weren’t friends with Mark Zuckerberg the chance to send him a message—for $100. The Zuckerberg fee was only an experiment, but it shows the extent to which Facebook is willing to charge users for communicating outside of their inner circles.
Turning a free network into a profit center is enticing but risky. Because Facebook is past its breakpoint, it could use this technique to make serious money, and the network could bear it. But smaller networks should be extremely careful. Networks that attempt to make money from their users before the breakpoint risk a complete collapse. Before it is exhausted, carrying capacity can ebb and flow, potentially leaving room for competitors. Charging too much for a feature that many users find valuable isn’t smart.
The same is true with advertising. It should be well noted that very few successful networks allowed advertising from the beginning—it wasn’t present on Google, Facebook, LinkedIn, or Twitter until the sites reached maturity. Even then, advertising was introduced slowly, and there were numerous times at which a company had to pull back certain programs.
Consider how Google began to make money. For the first few years, it didn’t. There were no user charges and no ads. The company began toying with revenue models only after they hit scale but did nothing significant until 2000, when Google launched AdWords, a targeted advertising system unlike what was typical on the web. The ads were designed to act like content: they consisted of text that was relevant to the user’s search. AdWords wasn’t scaled until 2003 with the launch of a new technology called AdSense, at which point Google had largely beat out every other search engine. And from there, Google reaped the rewards, with revenues growing from a respectable $83 million in 2001 to almost $1.5 billion three years later. In 2012, Google’s revenues were over $46 billion.
Facebook’s dominance is as clear as Google’s. They are well past the breakpoint, and users are more than willing to pay in terms of time, distractions, and cost. Unsurprisingly, Facebook has recently made a major push into advertising, with much of it being invasive, targeted, and highly personal. Facebook is using people’s content, friends, and even their habits to sell products on behalf of advertisers. The lines have blurred to such an extent that in some cases it is hard to tell friendly content from ad content. Facebook has even converted comments from users into real ads. One user posted a joke about a personal lubricant and Facebook converted it into a Valentine’s Day ad. But even at this extreme, users in a network past breakpoint are willing to put up with these tactics. The New York Times quoted the lubricated user as saying, “I was mildly annoyed, though not to the poin
t of deleting my Facebook account or throwing a hissy fit.”
Of course, not everyone can be Google or Facebook. But the point is that even these companies held off on making money before they had stable dominance across the internet. It is tempting to try monetization early, but the rewards for long-term patience are great.
The message is equally applicable to smaller companies. Often, small businesses cannot easily dominate a large market, nor do they have the resources to subsidize costs. The trick for a small business is to redefine the market so that it is small enough to own. Here again, Facebook provides a great example. When Facebook was launched in February 2004, MySpace was the dominant social network, and there was no hope for Facebook to compete. In a stroke of brilliance, Mark Zuckerberg limited his network to Harvard students. That allowed Facebook to dominate a much smaller market. Within a few months, half of Harvard’s student body were Facebook users. Only after hitting a breakpoint at Harvard did Facebook open to students at other Ivy League schools. Slowly, it allowed all college students and then high school students to join. It took another three years before Facebook opened the network to the world. And perhaps most importantly, it was a total of five years before Facebook began making any money.
All businesses can learn from the Facebook story. For the largest businesses with the biggest goals, this “network of networks” approach is a great way to move into a market. If your sights are broad, first tackle a smaller market and then use that network dominance to move into other markets until you reach your ultimate goal.
A smaller business must focus on a market it can truly own. If Zuckerberg had been thinking small, he could have limited his network to Harvard and then quickly monetized to reap the rewards of a small but dominant network. That would have worked perfectly well, and his dominance at Harvard would likely still persist today. The idea for a small business is to focus on an appropriately sized market, geography, or category and dominate that niche. A company that has grown to lead the network for skydiving in Boise, Idaho, for example, can reap the rewards and make money. Networks of underwater basket weavers, fountain pen fanatics, or F-16 pilots can be equally vibrant.
There are countless success stories in this regard. Edmodo, created by two Illinois school district employees, is a site limited solely to teachers and their students. It’s both a social network and a virtual classroom. The Huffington Post reported that over 25 percent of all teachers—inside and outside of Illinois—are already using the site, which is replacing more general social networks for teachers and students. That number is up from 3 percent penetration only a few years ago. Once Edmodo fully dominates the market, the network will be extremely valuable, though admittedly smaller than Facebook or Twitter.
Businesses don’t have to create their own network to benefit from network dominance. Most businesses aren’t in the network-creation business, but almost all can leverage other networks. Naked Pizza used Twitter to build its business and expand into markets outside of its native New Orleans. The business collected followers market by market; once Naked Pizza had critical mass in a given market, they opened a store and used the network to generate excitement in their customer base. Entrepreneur magazine noted that the buzz for Naked Pizza was disproportionate to its size, and the New York Times highlighted Naked Pizza as one of 11 companies whose Twitter strategies should be emulated. The company’s founders perhaps stated it best when they explained that rather than running a pizza company, they are actually “running a social media operation that happens to sell pizza.”
The benefits of a stable network are worth the wait for companies large and small. Stable networks are no longer at risk of losing users. Users will come and go, just as neurons come and go in the brain. But the larger network will remain intact. Economists call these “natural monopolies,” a term originally coined by John Stuart Mill. A natural monopoly occurs when it is most efficient for an entire market to be represented by a single company. You often see these monopolies in a utility industry, for example, but they also exist within networks. This is because in both cases, the value of an additional user is disproportionally greater than typical economies of scale. In these industries, user growth provides higher value to a single company than it would to multiple companies.
What business doesn’t want to be on the right side of a monopoly? Markets change, and that shift can cause a decline, but otherwise natural monopolies will persist in networks. (Of course, sometimes the government can get involved, which is always a concern. AT&T found this out in the 1980s, when the US government forced it to split because of its domination of the telephone network.) Nonetheless, the benefit of waiting until your network is dominant cannot be overstated.
LinkedIn, which recently surpassed 200 million users, has had considerable success navigating these waters. The site targets professionals and LinkedIn has done an impressive job of staying true to that niche. They have never encouraged users to add sparse connections; in fact, the site doesn’t allow people to see that you have a million connections (it will show “500+” in that case). This eliminates the popularity contests that abounded on MySpace and persist on Facebook. LinkedIn highlights degrees of separation, letting users know just how connected two people are.
LinkedIn also employs a freemium model, where the site is generally free to use but requires a subscription to access certain features. It charges users who want to dig deeper into the network. The pay features tend to be higher level, so there is no real risk of average users feeling slighted. The company does take advantage of advertising, but here, too, LinkedIn’s cycle followed the right path: it started ad-free, later employed content-rich, non-invasive ads, and only recently started displaying more prominent ads.
It shouldn’t surprise us that LinkedIn figured out a successful strategy, as the company was founded and led by one of the most prolific networking investors of our day, Reid Hoffman. Hoffman has invested in virtually every network imaginable and has likely seen the stages across each of them. Hoffman’s list is almost too long, but it includes early adopters and powerhouses alike, including SocialNet, PayPal, Facebook, Zynga, Wikia, Digg, SixApart, and Last.fm. Without a playbook for how networks unfold, the next best thing is to witness how it happens across a few dozen successes and failures.
With so much focus on growth, few people have seen what happens to networks in the long term. Growth is necessary when oxygen remains, but once it runs out, you have hit the breakpoint. At this stage, the carrying capacity has been consumed and the market is dominated. It is then, when it is almost impossible for formidable competition to arise, that there is an opportunity for a network to become a business. A network past its breakpoint is like a sea squirt who has found his lasting rock home: it’s time to reap the rewards and eat the brain.
Nine
Pheromones | Language | Mirrors
Despite their minimal brainpower and inability to speak, ants are remarkably good at communicating, which they do automatically through scent. Their bodies are covered in a greasy layer called cuticular hydrocarbon, and each ant carries a unique cuticular hydrocarbon pheromone, or scent, that is specific to the colony. When an ant encounters another ant, she touches her antennae to the other ant’s antennae or body, and she can instantly determine whether the ant is part of her own colony. If she finds that the other ant is a nestmate, she investigates further and can smell where the ant has been, what task the ant is currently performing, and, in some cases, what the ant has eaten recently.
Ants come in contact with many nestmates throughout the course of the day. The patterns of a particular ant’s interactions largely determine what task, if any, she decides to perform. When foraging, an ant follows scent trails left by other ants to know which direction to go and to predict how good the food will be. When ants encounter high-quality food, they leave a stronger scent trail, thus encouraging other ants to follow the same trail and bring back more of the good stu
ff.
Ants have a much more acute sense of smell than do other insects. A new study by biologist Laurence Zwiebel at Vanderbilt University mapped out a typical ant olfactory system and found that they have 400 different olfactory receptors. This is a huge number for an insect; honeybees have fewer than 200, and fruit flies have a paltry 61 of these odor-detecting proteins. “It’s a reasonable supposition,” says Dr. Zwiebel, “that this dramatic expansion in odor-sensing capability is what allowed ants to develop such a high level of social organization.”
The nuances of scent are vital to the survival of ant colonies. In fact, if you want to fool an ant colony, all you have to do is have the right scent. One type of jumping spider has evolved a unique survival strategy: they copy the scent of a particular colony and walk right into the nest. They steal larvae right under the ants’ noses, which don’t detect intruders that have replicated the colony scent.
Perhaps most interesting to study are the interactions between ants from different colonies. Most often, they avoid each other at all costs. As Deborah Gordon puts it, “Ants sometimes look like they jump apart after an encounter with an ant that is not a nestmate, recoiling from the unfamiliar smell.” If a harvester ant meets ants from a different colony while foraging, she heads in the other direction immediately. She does not lay down a scent trail on her way home, which keeps other nestmates from heading in the direction of the foreign ants.