I got a first-hand lesson in the pace of technological change when I was writing my paper on the DMCA. I wrote the first draft last July, went through Cato’s editing process from October to January, and then the paper was released in March.
Between the initial writing of the paper and its publication 9 months later, the online video marketplace underwent a revolution. When I was writing last July, I described the market for streaming video as stagnant, dominated by three mutually incompatible, vertically integrated video platforms controlled by Apple, Microsoft, and Real, respectively. By press time, YouTube had burst onto the scene.
The remarkable thing about YouTube is that from a technical perspective, there’s nothing especially remarkable about it. Instead of inventing their own video streaming format, as Apple, Microsoft, and Real had done, they built a video player atop Adobe’s popular Flash technology. Flash was originally designed as a platform for lightweight web animations, but it has evolved into a full-featured application platform. Luckily for YouTube, Flash was already installed on the vast majority of PCs and Macs, meaning that YouTube didn’t have to go to the trouble of getting its software installed on millions of PCs.
Here, I would argue, is an example of a case where we had at too many video platforms that tried to do too much. Had Microsoft, Apple, and Real been able to agree on a unified video format, and allowed other third parties to develop for that format, it’s likely that streaming video would have taken off much earlier. But because each company was more focused on making sure its format prevailed than on ensuring consumers had the best possible video experience, the streaming video market stayed in a kind of stalemate for close to a decade.
It didn’t have to be that way. The 2000 Streambox decision provided an ideal opportunity to transform Real’s video format into a de facto open standard the same way that Phoenix and Compaq turned the IBM PC into a de facto open standard in the early 1980s. It’s impossible to tell exactly what would have happened, but it seems likely that a Streambox victory would have caused a proliferation of copycat video players and servers built on the Real format, and that the network effects from those interoperable products would have built up the critical mass needed to make Real’s format the industry standard. Instead, we had to wait another 6 years until Flash–a technology that wasn’t even designed with video playback in mind–matured to the point where it could support YouTube and its competitors.
There’s an important lesson here: innovative products are rarely built the way NASA built the space shuttle. A few of the largest companies, such as Microsoft, sometimes build everything from the ground up, but the vast majority of innovative products are a thin layer of proprietary technology atop a big stack of commodity hardware and software. (Flash isn’t truly an open standard, but it’s open enough for our purposes: Adobe didn’t try to restrict what kind of technologies you could build on top of it)
MikeT offers another example of the same phenomenon in comments to my last post:
The single biggest blunder that [advocates of platform monopolies] make, at least in terms of IT, is assuming that the practical results of R&D here are something more than derivative, let alone evolutionary. The XBox is, for example, a stripped down PC. AFAIK, it even uses the exact same DirectX APIs that a Windows PC uses for games. Most of the cost associated with it was simply creating a stripped down PC and absorbing the cost of selling it at a loss.
A lot of the resources spent on developing new platforms don’t go into making a device work. Often, that can be done quite quickly using off-the-shelf components. What’s expensive is building closed, vertically integrated platforms, because that frequently requires re-inventing various wheels. Companies tend to build vertically integrated platforms because those are easier to monetize, not because they’re necessarily technically superior.
All of which is to say that the assumption that large profits are required to induce the creation of innovative devices and platforms is usually false. If there are large profits to be had, companies will focus their development dollars on proprietary platforms that will make it easier for them to capture a larger share of those profits. But if there are only modest profits to be had, companies will focus their development efforts on incremental improvements to existing technologies. Each step might therefore be less innovative than the introduction of a new proprietary platform, but because the incremental improvements can be produced so much more quickly, the overall pace of innovation is likely to be faster.
Perhaps more importantly, each generation of open technologies becomes a foundation on which a subsequent generation of open technologies can be built. Closed technologies tend not to work as well as foundations for subsequent generations of innovation, for two reasons. First, there’s the obvious point that licensing requirements impose considerable frictions. Secondly, closed technologies are more likely to be special-purpose and incompatible with other technologies, making it harder to build a new technology by combining several existing ones. Generally speaking, closed technologies are technological dead ends. They might be massively profitable to the company that created them, but unless they’re opened up they’re not likely to form the foundation for subsequent generations of technology as open technologies often do.
Thanks to all three of you who made it this far. In the next post I’ll wrap things up with a succinct summary of the argument I’ve laid out.
Update One nuance that I forgot to mention above is that whether products are tightly integrated is a distinct question from whether the platform those products comprise is open or closed. Obviously, some vertically integrated platforms, such as the iTunes/iPod combination, clearly are innovative and of benefit to consumers. However, there’s no reason to think that opening the platform up legally (that is, permitting other companies to reverse-engineer the devices) would undermine the advantages of tight integration.
Abolishing platform monopoly rights wouldn’t mean the end of vertically integrated platforms. Companies would still vertically integrate if and when doing so led to better products. What they would stop doing is vertically integrating their products simply because doing makes it easier for them to sue competitors.