I like this topic for much the same reason I like horror movies. I cannot look away, even though the action turns my stomach. And, though I am not a sadist, I want to see the plot play out — I am just a little curious to see who gets an axe and who escapes the chain-saw.
In short, watching the splintering of the Internet is a version of bird watching for nerds, with a bit of Hitchcock thrown in.
Ben Kunz suggests a label for the phenomenon — calling it the rise of the “Splinternet” — and he tried to popularize it recently in a Business week article. He borrowed the phrase from Josh Bernoff, a well known analyst at Forrester (and Kunz gave Bernoff ample credit).
Any short and self-explanatory label beats a long and confusing one, so the splinternet has much to recommend it as a label. There also have been many less snappy labels proposed for this phenomenon. This one is snappy.
But, if you look closely, this one is a bit confusing too. My goal is to unpack the label in this post.
Let me rephrase that, and signal where I intend to go. The Internet should no longer be called a “network of networks”, as it was called two decades ago. That era has passed. Commercialization has brought with it a new structure, a “network of platforms”. The splintering of the Internet describes the results of platform competition on the Internet.
Sometimes it resembles a horror movie. It is not all bad, however. This requires explanation.
Platform competition on the Internet
The latter already exists. There is a company called Splinter.net, or Splinternet holdings, and that firm sells security software. That is something altogether different from what Josh Bernoff had in mind.
Splintering of the Internet, however, is where Bernoff wants to focus attention. OK, let’s go there.
Splintering is a natural byproduct of platform competition on the Internet. That sentence is a mouthful, so let’s take it one piece at a time.
Platforms are reconfigurable clusters and bundles of technical standards for supporting increasing functionality. From a user perspective platforms usually look like “standard bundles” of components, namely, a typical arrangement of components for achieving functionality.
In down-to-earth language…. Windows is a platform for PC software. So are the Intel architecture and related mother boards. Oracle databases have their own platform, and so does SAP software. The iPod/iPhone platforms seems big enough to deserve this status too. The cluster of services sold by Google as a search engine/adsense/adwords (along with many beta services) serve as a rather effective platform too.
Platform sponsors have to make a living just like everyone else. Their conduct displays recurring patterns. To oversimplify, these firms have to generate revenue in excess of cost, but it is tricky do that well as a platform sponsor, and for a number of reasons — whether the sponsor lives inside the Internet or not.
Perhaps the biggest problem arises from the inherent mismatch between the timing of investment and the timing revenue. Most platforms get built far before their deployment. If they work as intended they yield revenue much later, long after the investment, and over long periods of time.
There are a couple tools available to platforms sponsors. Without going too deeply into it, let’s review a few of the tools. (If you want to go deeply into it, see some of the books listed here).
* 1. Switching cost: Sponsors of platforms do better if users do not frequently flit from one platform to another, but, rather, stay at one place for a while, generating revenue. Accordingly, firms engineer switching costs to make it harder for users to switch.
* 2. Migration: This is related to switching costs, but is not quite the same thing. Smart platform firms design their products to entice new users to migrate to their platform, leaving other ones, even those with high switching costs. There are many tricks for easing migration, but let’s just leave it there for now because I will mention several below — good firms think about motivating users to migrate and lowering their costs of migration.
* 3. Manipulate ecosystems: Smart platforms sponsors nurture a series of relationships with other firms so the sum of all their efforts raises the value to users from remaining loyal to the platform. Many executives like to call these ecosystems, and many economists like to call these coalitions of complementary component providers. Whatever it is called, a platform sponsor needs commitments from developers and application software to make the platform valuable.
* 4 .Constituent-oriented pricing: Smart platform sponsors price to subsidize those participants whose presence confers benefits to other types of participants. This is occasionally given the label “two-sided markets”, though often the pricing involves more than two sides. This topic is very complicated, and takes a lot of careful strategic thought to do it well.
Look, I am not saying anything that has not been said by others. This post lists books by Cusumano, Gawer, Shapiro, Varian, Evans, Hagiu, Schmalensee, and Zittrain. If you are curious about these topics, go read the books. And there are plenty more than these.
(However, this is the moment for me to blow my own trumpet. Tim Bresnahan and I said some of these things in this academic article, published way back in 1999, and our first published article on the topic appeared two years earlier (in this article). And, also, let me give credit to Joel West and Jason Dedrick too, who have many of the similar ideas in their writing. In particular, Joel West has been on this subject for a long time. But I digress.)
Still, this theme is loads of fun to revisit. Here we ask: How does splintering the Internet arise? In a nutshell the answer begins thusly: no law compels any firm to interconnect with any other. Every firm always has the option to opt out of the Internet, or do something slightly less dramatic, such as opt out of commonly used standards.
It helps to understand why. The answer lies in the four behaviors described above.
As it turns out, battles between incompatible platforms arises from combination of strategies to embed switching costs (1 above) and migration strategies (2 above) in products. Firms compete to migrate users from one platform to the other. Indeed, such competition almost always takes place in anticipation of the emergence of an ecosystem (3 above) and multi-sided pricing for it (4 above), so platform sponsors will do quite a bit of work to build an installed base if they see large gains down the road from it.
For example, well in advance of the actual sales, the sponsors of HDDVD and Blu-Ray tried to line up enormous coalitions to build the platforms. It was a mean fight, and expensive. Incredibly, the competition was almost nascent. One side gave up rather quickly, when less than 5% of potential adopters had bought anything, partly to stem anticipated losses.
Multi-sided platforms, point (4), are fascinating to watch too. In these platforms the sponsors present distinctly different interconnection faces and fees to different participants in the ecosystem, all designed with an eye towards improving the situation for all participants.
For example, the presence of Google’s users confers benefits to its advertisers, so Google does not charge the users. It operates a word auction that gives advertisers incentives to be relevant to users. That led Google to invent all types of innovative features in the auction. Upon close look, the pricing is quite subtly done, with advertisers paying different prices, tailoring their services to their own needs.
All the features above generates strong incentives for platform sponsors to try to “leapfrog” each other with big technical splashes that generate migrations. Steve Jobs has become obsessed with this conduct, for example, insisting that product launches always come with big splashes.
For example, it is well known that he delayed the iPad for years because he did not think it had enough functionality to make a big splash. Deferring revenue on a new product in order to invest in launching one that might surpass others is classic leap-frogging behavior.
Related, any reasonably strong and established platform sponsor inside an ecosystem (as in point 3) will try to raise prices (point 1) and try to deter others from leaving (by using migration strategies, as in 2). The incentives for doing this are especially strong after making big investments to attempt to leap frog others. Ditto if the sponsor is also creating a new ecosystem, and sponsoring a new multi-sided platform at the same time.
Less abstractly….This is a long way of saying, for example, that back in 1995, just after Microsoft had released Windows 95, it had an enormous incentive to raise the switching costs of users and programmers, deterring any migration to an alternative platform. That also gave Microsoft and Bill Gates strong incentives to play hardball in the browser wars.
(BTW, we are now 15 years into that and most of this has played its way out. Oh, and Microsoft sure made a lot of money in the meantime).
Once again, nothing new. My only broad point is that during the browser wars Microsoft exhibited behavior that had arisen before (and continues to arise) and for reasons that continue to reoccur.
(Hmmm….the Internet did not splinter as a result of the browser wars…. why is that? Judicious government intervention perhaps? But I digress once again….)
And that is also a long way of saying that Google has a particularly interesting approach to its ecosystem and to developing its multiple-sided platform. Among the many things it wants to do, it wants to raise switching costs to advertisers (point 1), and generate incentives among users to migrate to Google with beta services (point 2) by nurturing an ecosystem of interrelated providers (point 3) using open source software. It is an interesting strategy, though it also has its tensions.
Similarly….Steve Jobs has developed such a large installed base of users that he seems close to approaching a point where he could be happy to raise switching costs (point 1) in his iPod/iPhone platform, preventing migration to Googles’ new one (point 2). The ecosystem (3) around the iPhone is especially large and impressive, and will be hard to dislodge going forward, so it is anybody’s guess what will happen next.
For those of us who watch this stuff, these type of fights are endlessly fascinating. We tolerate the horror show because the plots play out in so many different ways. To a platform afficionado, this sport combines the strategic complexity of chess with the psychological depth and risks of poker. Many little games add up to a very large meta-game, and no two meta-games ever look quite the same.
And that is my point. Splintering happens all the time in this market. But it is a symptom of something bigger, not to be understood in isolation.
Drilling down to the Splinternet
The Splinternet is a result of the complex combination of these competitive forces.
* So, for example, the industry is abuzz with Steve Jobs’ decision to not support Flash on the iPad, and to foster Apple’s own proprietary standard.
Look, this little decision is not so hard to understand. Jobs wants to design the platform, and he believes that Flash is going in a direction that does not serve his platform’s interests — either because it is getting to be too powerful in the Internet ecosystem, or because it is not optimized to the iPad’s architecture (uses too much of the battery), or it allows users to migrate too much content on and off the iPad. Whichever it is, and I am not sure, it does not matter whether Jobs is right or wrong about this decision. It just is not very surprising that he wants to build his own tools. It is normal behavior in platform competition.
* Many advertisers are also abuzz with all the efforts they expend to make their landing pages consistent with Google’s advertising policies. Many of them have invested heavily in understanding the set of key words to use and what they must do to succeed on Google’s platform.
Well, it should surprise nobody desired what happened next. Now that advertisers have gone to all that effort they are not particularly interested in doing it again for another search engine. Few seem to be switching to Bing, for example (even though it has actually gotten reasonably good). Think of it this way: Advertisers would like there to be a bit more competition, but each advertiser would like someone else to be the one to move and bargain down prices. Google anticipated that, so it aspired to build a big ecosystem (and succeeded). Hence, it is not really in any individual advertiser’s interest to change the system. And Google gets my respect for being clever.
* Here is another version of a related phenomenon: when AOL controlled the largest IM network a decade ago, there were plenty of firms trying to interconnect with those buddy lists. AOL kept stopping them.
Look, has life changed much? Not really. AOL’s behavior is not much different than Skype’s today. Skype refuses to interconnect with any other software today. Nobody comments on it because we are so used to it. (Of course, if AT&T refused to connect the calls coming from Verizon that would be a different matter….but I digress again.)
Put it all together. We started with this question: Why does such behavior result in the Splinternet? We have come to this answer: Almost by definition, platform sponsors faces strong incentives to tweak interconnection with other platforms. Firms want more switching costs and resist letting their users migrate away. Firms want the business partners to commit to one platform and stay there. Seamless migration across the network between platforms never can be — indeed, never will be — an attribute that results from these forces, because, in short, firms opt out of commonly used bundles of components.
But, wait a minute. Before we leave this horror show with a stomach ache, let’s notice what else has happened at the same time. If platform competition has been vigorous and robust, then users have received more functionality at lower prices. That is not a bad reward for sitting through a hatchet job.
Let’s summarize. What does result from platform competition? A splintered mess with lots of innovation.
At close inspection the Internet looks confusing. Its participants are fractured in their aspirations, vibrant with energy, evolving with starts and stops on a technically hilly domain, trying to out-maneuver one another in an endless quest to build an installed base that another rival cannot steal.
A slothful legacy of prior investments intertwines with an uncertain future of technical possibilities. The sum of many uncoordinated initiatives stresses the connective tissue of the network, resulting in a whole so disorganized beyond recognition that none of its designers know what to call it other than the mother of all Gordian knots.
Ah, but at a slightly higher level of observation the Internet looks different. It flows, ever onward, toward expansive functionality. Getting better. Yielding new services and at lower prices. It progresses in stops and starts, to be sure, but at least it progresses.
Sure, no law makes the Internet work, and no law compels all firms to interconnect with all firms. But there is something healthy there too — as long as nobody gets too big, and the switching costs do not get too large, and the ecosystem continues to generate new entrants.
It is fragile, to be sure. Frankly, it is not too hard to see a number of things that could go wrong. I am not always so upbeat. I am quite willing to say that the Internet has tendencies that might slouch toward a dystopian future. Indeed, I have said so here. If you want an alarmist viewpoint, read that.
For now, let me end this horror picture on a quieter note. If nothing else, the splintering of the Internet sure is fun to watch — sort of like bird watching for nerds, except not as leisurely. It is more like a Hitchcock version of Gorillas in the Mist.
And I can’t look away. Can you?