Ecological Thinking, Applied

The idea in my previous post is that complex systems have niches, and it's often more productive to think about behavior as a property of the niche rather than the actors who inhabit the niche.

In this post, I'd like to explore what kinds of systems yield to analysis by the ecological mindset.

Let's start with biological systems.

It's unintuitive at first, but in biology, ecological thinking doesn't concern itself with individual organisms, but rather with entire species. This is because individual organisms aren't adaptive enough to change their behavior in meaningful ways. A tree, a shrub, a weed, a bacterium, even a snail or a bat — these individual organisms have behavior that is mostly determined by what their genes have programmed for them. If its environment changes, a single tree is going to keep doing what it’s always done; it can’t adapt (much). Species, in contrast, can significantly change their behavior, at least over evolutionary time scales.

What does this tell us about the kinds of systems that are 'ecological'? Let me propose this criterion:

A system can be analyzed as an ecosystem if it has independent, competing agents who can change and adapt to their environment.

So, besides the biosphere, what other systems fit this definition? Lots, it turns out. Communities of all sorts: corporations, agencies, committees, the student body at a high school, nations, online communities (think Reddit). Markets, where firms compete against each other to win resources (employees, customers, investment dollars). Financial markets, where traders try to outwit each other. The dating market. Academia. The media.

A corporation is both an ecosystem unto itself (where employees are the agents who inhabit the ecosystem), and an agent within a larger ecosystem (the market, where it competes against other firms).

The Internet is an ecosystem of staggering complexity, in part because it supports so many different types of agents. Twitterers compete for followers. Web pages compete for PageRank. Spammers are agents competing with legitimate content for eyeballs and mindshare. At the very lowest levels of the Internet, packets 'compete' for delivery, and a variety of mechanisms have been developed to incentivize processes to play nice with their packets.

But in all these cases, remember, it's not actually about the agents. It's about the system and its niches. In ecological thinking, agent behavior is a property of the system, rather than the other way around.

What's powerful about this way of thinking is that it abstracts away from individuals, and allows us focus on the properties of the system that are causing different types of behavior. In the process, it suggests completely different types of solutions to a lot of big, thorny problems. It asks us to stop thinking about the players, and get to work reforming the game.

Let's see some applications of this new mindset.

Office politics

To reduce the amount of politics in your workplace, it's not enough just to say, "We'll fire people who play politics." Instead you'll want to think about how, where, and why politics occurs. What gains do people feel they can achieve by playing politics, and how can you change the system to make politics less rewarding?

If politics thrives where there is restricted information flow (secrets, back-room deals, information brokerage), work on increasing feedback and transparency. Use open floor plans, encourage CC habits, and force people who are avoiding each other to talk. Maybe your office ecosystem has room for politics because there's a leadership vacuum, in which case you should fill the void.

Of course the threat of punishment (firing) is one way to narrow the niche, by increasing the costs associated with politicking. It's just not the only way.

National politics

If you want to change the world and politics is your scene, lining up on your 'side' and playing tug-o'-war is the least valuable thing you can do. Instead, you can have a much bigger impact if you work on changing the game.

Three examples of game-changing levers in politics are laws, technology, and education. Laws include things like campaign finance reform and constitutional interpretation. Technology includes ubiquitous video and web sites that make it easier for people to organize. Finally, education affects politics by changing the minds of the voting public — the substrate on which politics is played. A more educated public narrows the niche for bad behavior by politicians. The fact that so many Americans have read 1984, for example, has a huge effect on what kind of political behavior is possible, at least out in the open.

Drug violence

Ultimately, drug-related violence isn't caused by drugs or even by drug users or dealers, but by drug policy.

Because there will always be buyers, there will always be a niche for selling drugs. Putting pressure on that niche, by criminalizing it, isn't making it go away (this is an empirical fact I hope we can agree on). Instead, it's only making the niche riskier. Since the stakes are so high — jail time or death on the one hand, big big money on the other — the 'drug dealer' niche can support only people who are desperate and/or ruthless. Inevitably the result is violence.

And like all niches, the drug-dealing niche exists independent of any actors who might be filling it. Take out the kingpin and the niche is unfazed; someone else will soon step in to replace him. The only way to win is to change the game — treat drug abuse as a medical issue rather than a criminal one, for instance.

(I realize I'm not saying anything new here. I'm just showing how each of these problems can be analyzed with the lens of ecology.)

Product development

Proposed by Venkat Rao, the Milo Criterion states that products must mature no faster than the rate at which users can adapt.

To understand this, you have to see the product (Google+, say) as inhabiting a niche in the marketplace. Users are the medium of the market, much like voters are the medium in which politics is played.

If you dropped the Google+ of 2012 into the Internet of 2002, it would probably fizzle. Users (of 2002) would take one look and recoil in confusion. Circles, streams, hangouts, +1 — what does it all mean? Fortunately, Google+ was launched into the Internet of 2011, where users had already been trained, by Facebook, to understand most of the features immediately and intuitively. In other words, by introducing one feature at a time and training users slowly over 7 years, Facebook carved out the niche that allowed Google+ to launch, somewhat successfully, in 2011.

On the other hand, if a product is criticized for being "ahead of its time," sometimes it's because the users just aren't ready — the product is launching into a niche that doesn't fully exist yet. Google Wave may have failed for this reason, although there were plenty of other factors as well.


I don't know where all this is going, but here's what we have so far:

In complex systems, niches exist. A niche is a property of the system, independent of any agents who happen to be filling the niche. Agent behavior is explained by the niche, rather than by properties of the agents themselves (at least when using the ecological mindset). And finally, when looking at problems that arise in a system, it's often more productive to think about solutions at the niche level rather than the agent level.

Originally published April 3, 2012.