Sunday, March 20, 2005
5 Years Later: A Short History of the Bubble (and the Next One)
I wrote this piece for Om's 5 year bubble anniversary compendium (it's a bit late :) . It's waay too long, so I divided it up into two parts. It's also kind of messy - will edit as time permits. Enjoy!
Five years on, it's easy to see in retrospect what went wrong during the bubble. The Net turned out not to be such a great 'channel' - which is what hordes of fresh-faced MBAs toting business plans hoping to sell pet food, lava lamps, and expensive yet shabby looking clothes were taught.
Instead, the Net turned out to be great at, well, what you might think something called 'the Internet' would great at: connecting people. All of the most succesful business models on the Net are, essentially, vast switchboards.
Sometimes, they connect consumers with other consumers, creating huge gains to consumption - this is Amazon's model. Imagine Amazon without reviews and user info - I don't think it would have been very succesful, no matter how vast an inventory it offered. We only have to look at numerous distribution-focused competitors that crashed and burned to see the power of Amazon's demand-side economies of scale.
Sometimes, they connect intermediaries with consumers - this is eBay's model. Sometimes, they connect producers with consumers - this is the revolution that Google and Overture jointly, quietly unleashed on a baffled, dumbstruck advertising industry. Sometimes, they just join people, and worry about strategy later: this is Blogger, Bloglines, Flickr, del.icio.us, and Metafilter. Sometimes, they try and do it all: consumers, producers, intermediaries, and fun - this is LinkedIn, MySpace, and, to an extent, Yahoo.
I went to b-school because I wanted an answer to a very simple question: if the Net was revolutionizing my life, my family's lives, and my friends' lives, why did the massive value that was created go up in flames?
The intuitive answer - and I think the right one - is that the Net was never a great place to sell dorm furniture, pet food, or lava lamps. Most people who need those realize a greater surplus by shopping at local retailers. But the Net does a very special set of activities much better than the real-world can. Now, you probably know what I'm gonna say, so you might want to skip to the last third, where I discuss the future, not the past.
The Net is a much better place to buy and sell things that are costly to find in the real world. Not just 'information goods', or things made out of bits. Any goods for which search costs are high. This was the basis of a fundamentally new economics: what my old prof Gary Hamel called the transition from distribution economies to search economies. Those firms that could create search economies were the winners, because search economies are essentially massive gains to consumption, which create huge advantages over real-world firms dependent on real-world resources which create real-world consumption costs.
All of the models that I've outlined above are variations on this theme. Amazon doesn't succeed because it sells books more cheaply, but because it makes it easier to find books you'll derive greater utility from. Google/Overture's advertising revolution didn't succeed because it grabbed audiences by the throat better, but because it offered people ads they began to derive utility from - ads that were matched with things they were already looking for. eBay didn't succeed because it made it cheaper to get secondhand stuff - in fact, you can easily observe the winner's curse in action on eBay - but because it made a market in things people wanted to pay something for, but whose search costs were prohibitive.
Now, in 2005, the true power of search economies is about to rip most of the media industry apart. Publishing is on it's knees; the music and film industries are stumbling. This is what Chris Anderson has now famously called the Long Tail. I won't explain the idea again - but note it's fundamentally an expression of search economies. At the limit, search economics becomes Long Tail economics: markets for everything, everywhere. Because you can finally cheaply find the stuff that more closely matches what you really want, it's now becoming reasonably economically viable for others to make, market, and sell it.
This is like mass personalization - but given to consumers under a much different production regime and industry structure than predicted during the bubble. It's not monolithic firms that are going to personalize their goods for micromarkets, as the digerati thought. In fact, it's the inverse: microproducers with the ability to produce and sell goods crafted with passion to macromarkets.
The LT is, as many people suspect, going to define the next economics of the Net and traditional media. In fact, if you follow my argument so far, you understand the dynamic mechanisms behind it: as more efficient search mechanisms have been built, and as the size of the network has increased, the Tail has been lengthening. Because they don't have to pay high search and transaction costs to find and acquire them, consumers can now purchase goods from which they'll derive the greatest utility. Those gains to consumption are exactly equal to the fall in search costs.
Want vintage shoes? Now you can check out more than you can handle. But only because Google can slash your search costs, while making it economical to do so with ads that don't make you want to cut your eyes out. This really is an economic revolution: think about the size of search costs slashed in just one vintage shoes search, multiply that the universe of searches, and you get an idea how massive search economies really are.
Search economies are going to follow the same dynamics innovation usually does. The dominant design's been established, and now the race for incremental efficiency will begin. The dominant design will begin to mutate into offspring which resemble it, but are more efficient at surviving in the environments they inhabit. So search economies themselves will verticalize: travel search will begin to look very different than video search will begin to look very different than blog search will begin to look very different than people search.
The important thing to note is that this is going to be incremental change. Models will change a little bit, but not much - because the dominant design is hyperefficient in the current environment. Search economies are king. Something very important did happen during the bubble - the revolutionary transition from distribution economies to search economies. A huge amount of value was created - it only disappeared during the time the market's been sorting the winners out from the losers.
So I've spent a long time explaining what I think happened during and after the bubble. The more interesting question, I suspect, is: what now?
My answer is that an equally big, if not bigger, economic revolution is lurking just around the next corner. We can already see the first glimmerings of it in things like tags, CC Mixter, MMORPGs, SecondLife, blog communities, and even fabs.
This is what I call (inelegantly) the transition from search to coordination economies. My hypothesis is simple: when you can find anything you want costlessly (ie when hypercompetition erodes search economincs based b-models' margins), what becomes valuable? My answer is pretty simple too: doing something with all the cool stuff you've found.
But doing stuff, right now, is pretty hard. Sure, we blog, and podcast, etc. But these are solitary, atomized pursuits. It's hard to create a great deal of value on your own, unless you're Jenna Jameson or Stephen Hawking. There's only so much we can each do on our own. Collaborating to produce, market, distribute has massive economic implications - just like it did in the days of Adam Smith's pin factory.
My point is that there's a huge imbalance: one the supply side, we now have a huge set of new resources unlocked by better search mechanisms with which to create new business models. But on the demand side, there's still few way to do things with these resources: to combine, remix, retouch, refinish, reproduce, mutate, select them into newer, cooler, faster, smarter, sleeker, innovations.
Think about it this way. Imagine you're a musician. MP3 + Google creates the opportunity for the world to costlessly find your immense musical talent. That's the 2005 way of seeing you. But, I suspect, during the next bubble, we'll see you this way: as a musical resource with value creation ability going massively untapped because there's little way for you to coordinate your value activities with those of others.
Smart entrepreneurs are arbitraging the imbalance by creating coordination economies: things that make it cheaper for people not to find things, but to get together and produce things. Think of O'Reilly's Make. Think of selling things you've made in SecondLife. Think of Zopa. Think of the Gaming Open Market. Arbitraging the value creation potential of search resources by coordination economies is the heart of open-source/open-access/peer production models.
My favorite example is probably threadless, because it's so far removed from tech in general, and software in particular. It's a bunch of hipsters who realized that coming together as a community was cool, but that in collaboration, they could unlock relatively massive amounts of value. All they needed was a coordination mechanism. So they set up a simple website where anyone can submit T-shirt designs, the community rates them, and the winning designs get produced.
Now, this doesn't sound like a world-changing idea. But I think it might be. That's because threadless absolutely eviscerates the traditional economics of of all the industries involved in making, marketing, and selling T-shirts simultaneously.
Right now, most of your T-shirts are designed by a professional designer in the 1st world, approved by a committee of suits, made somewhere in Asia, packaged and distributed by BigCorp, and sold and marketed by BigBox. Threadless, and peer-production models in general, vaporize this traditional value chain. T-shirts threadless style are designed by amateur designers, approved by the community, made somewhere in Asia, and distributed, packaged, and marketed by, well, nobody. Maybe Fedex. The point is that half the value chain disappears.
But what's really important are the economic effects of coordination economies and peer-production models. I'm still wrestling with these, but here are my initial takes.
First, they hugely expand the market size. I might not buy a T-shirt from the mall, or even from tshirts.com - but I have from threadless. Second, coordination economies make it viable for pro-ams to make things only firms could make before. This is what I call distributed economies of scale, and it's about coalitions of pro-ams having a much lower opportunity cost of production than firms do, as well as access to factors of production. Yochai Benkler thinks it's about lumpy goods, which is probably also important.
Third, and this is the important one, by stripping out the half of the value chain we really don't value - the half that gives us cheesy designs, bad advertising, oscillating prices, and three whole colors to choose from - threadless aligns rent capture with value creation an order of magnitude better than before. This is bad news for Wal-Mart, Ralph Lauren, and your local mall, to be sure, but it's great news for T-shirt lovers, 3rd world suppliers, passionate designers, local economies, you and me - and, of course, potental innovators. And an entire generation of kids my kid sister's age - 11 - are gonna grow up spending their pocket money at places like threadless, Zingtone, and SecondLife.
At the limit, the transition from search to coordination economies changes the consumption landscape from a Long Tail to...something else. I haven't named it. Maybe we should call it All Tail (although that's a terrible name). Whatever we call it, it's my candidate for the next economic revolution: prosumers making stuff more efficiently and innovatively than firms. It's gonna be pretty cool, not least because whoever gives them ways to make it is going to get as rich as sin.
But (and this is what's really cool to me) the huge disruptions I suspect coordination economies - Web 3.0 - promise are evidence of Asimov's old dictum: we overestimate the impact of emerging technologies in the short term, and underestimate them in the long term. In other words, the bubble was a case study in irrational exuberance. But the innovation crash of 2001-2004, I think, will be remembered as (massively) irrational pessimism.
Thank you for that. I'll have to read it several more times to get the entire thing to sink in, but it was outstanding. -Dave Gallagher