I had a beer recently with Dave McClure of 500 Hats
. As is always the case when I get together with Dave, we had a long, rambling and enjoyable conversation about how the Web is changing the way businesses get built.
At some point, I said that the vector of innovation has changed. It used to be that innovation started with NASA, flowed to the military, then to the enterprise, and finally to the consumer. Today, it is the reverse. All of the most interesting stuff is being built first for consumers and is tricking back to the enterprise. I suggested that one reason this is happening is that the success of a web service is more often determined by its social engineering than its electrical engineering.
Dave immediately said he’d give me three months to blog that before he did. I thought that was generous even for me who doesn’t blog easily or often. But just to be sure I make the deadline, here is the post.
The basic insight that the flow of innovation has reversed has been out there as a meme for a while. Fred wrote about it
and referenced Esther Dyson’s Release 1.0 article. I took a shot at why it was happening; I focused on changes in the way services are built and their complexity
. The conversation that Dave and I had was more about how critical the user interface is in consumer facing web services and how that might influence the flow of innovation.
We have marveled more than once on this blog about the remarkable efficiency of Craigslist. That service is essentially a very lightweight governance system that manages an enormous collection of users who contribute all of the content and much of the oversight that makes the service work. It is because Craig and Jim focus on managing the efforts of their users instead of doing the work of those users that Craigslist is so phenomenally efficient. Many of the most interesting web services are like Craigslist, at their core, lightweight governance systems. Facebook and Twitter come to mind.
Even services that do more than mediate communications among their users often depend on users contributing data through their engagement with the service before they can provide value back to those users. Wesabe can only help users understand their spending and suggest ways to do more with less because users share their spending data with the service. Del.icio.us depends on users tagging the Web in order to be able to help users discover sites, services and memes on the web. Last.fm only works because users share their listening behavior with the service.
In the old days, electrical engineers focused on getting computers to work not on getting people to engage with the systems built on top of those computers. The folks that built enterprise software were vaguely aware that their systems had to be accessible to the humans that used them but they had a huge advantage. The people who used them did so as part of their job, they were trained to use them and fired if they could not figure them out.
Today, no one tells you to use Facebook. There are no employer sponsored training sessions on the use of del.icio.us. The burden is on the designer of the system to meet a need, entertain, or inform their users. They also have to seduce those users, hiding complexity, revealing one layer at time, always enticing, never intimidating, until the user one day finds they are intimately familiar with power and the pleasures of the service.
Designing a system that does that is not an electrical engineering problem. It is a social engineering problem. The best social engineers are working today on consumer facing web services. They understand that there is enormous potential leverage in those services. The creators of these services recognize that services like theirs will ultimately disrupt the economics of many, if not most, parts of the global economy in much the same way that Craigslist collapsed the multi-billion dollar classified industry into a fabulously profitable multi-million dollar web service.
So that, it seems to me, is one more reason the flow of innovation has reversed.