Brad Fitzpatrick, creator of LiveJournal, has an interesting blog post about portability of the ’social graphs’ of social networking sites like Facebook and MySpace. What is a social graph? In maths, a graph can refer to a network of ‘nodes’ and the connections between these nodes. That’s what social networking sites create, through the networks of friends that people build up. The members of the sites are the nodes, and they’re linked to other members through their sets of friends, and the whole thing is the ‘graph’ of the site. This graph is the basic asset of social networking sites, since belonging to such a site is next to useless if it doesn’t allow you to ‘connect’ to people.

As Brad explains, it’s a pain for people to have to build up their networks on many different sites. It’s tiring to add all my friends on Facebook, and then have to add them again on some other site for some other purpose. Brad suggests that this is preventing new social networking sites from getting established, because the cost to build up a network on multiple sites is too high for users, and because a site with a small graph (ie few members) isn’t very attractive to new users.

In economics terms, social networking sites are classic examples of services that exhibit ‘network effects’, where the value of the service increases with the number of its users. In such markets, entry by new firms can be difficult, because established firms have already built up their networks, making them more attractive to users. The newcomers can try to overcome this barrier by providing a superior service, or differentiating themselves from the existing firms in some way, but everything else equal it’s arguably going to be more difficult to enter the social networking market than, say, the market for pencils.

What Brad is really talking about is making the networks that the sites have established ‘compatible’ between sites. So if you join one site then you’ve effectively joined them all. This removes the size of the network as a differentiating factor between sites and will make it easier for newcomers to enter. There are a couple of economic effects that will underlie the decisions of the social networking sites to go along with this or not.

First, compatibility may be desirable because it weakens competition between networks, resulting in higher profits. Social network sites don’t (at the moment) charge prices for membership, but the way they compete is in the features that they provide to users. It’s expensive for the sites to develop nice user interfaces and other features that people want. More features equals a lower price, in some sense. Reducing features will result in losing some users, but will save the site some money. With compatibility, if a site reduces its features and loses some users to a competing site, its remaining users will still be able ‘connect’ to those who switched, and so the value of the site to the remaining users doesn’t change. Thus the ‘punishment’ that the site faces for reducing its features is weaker with compatibility. This will lead to compatible sites implementing fewer features, and presumably making higher profits. That’s not necessarily bad for users though, as the benefits of being able to connect to a larger network that compatibility brings might outweigh the losses from reduced features.

So compatibility may be desirable to existing competitors. On the other hand, an existing firm may not want to make its network compatible with a new entrant. Compatibility will weaken the competition as described above if the new firm does enter the market, but no entry may be an even better outcome than entry of a (weak) competitor. However, it’s not always desirable to keep competitors out, if they provide innovative products that expand the size of the market as a whole. In a different context, this perhaps explains why Adobe chose to make the PDF format an open standard and allow competitors to produce PDF-compatible software.

Overall, Brad’s proposal may be a good idea from the point of view of users of social networking sites, provided that the benefits of belonging to a bigger network outweigh any losses from reduced intensity of competition in terms of the features provided to users. But I’m sceptical whether the larger sites like Facebook will accept it. They’ve spent effort and money building up their networks, and unless they see big benefits from market expansion then I doubt they’ll want to establish compatibility with new competitors.

Standards are becoming ever more important in IT, communication and entertainment industries. The very fact that you can view this web page is thanks to standards. Your browser knows how to display the standard known as hypertext markup language (HTML), which is a set of rules for specifying how text and other elements display in your screen. My web server communicates with your browser thanks to the HTTP standard, and other standards are used for the pictures and the font formatting, and so on.

Standards are useful because they allow competition. Different web browser programs are available and can compete with each other because they all understand the HTML standard and other important web standards. Without these standards, we would have competing proprietary systems for web servers and browsers that would only work together in specific pairs. Either one proprietary system would emerge as dominant, destroying the benefits of competition, or a number of separate incompatible systems would exist, making life difficult for web publishers and web surfers.

On the other hand, standards themselves exhibit network effects. A standard becomes more useful the more potential users of the standard adopt and support it. This means that it’s often the case that a single standard dominates in a given area. HTML is the only dominant standard for displaying web pages. Similarly there’s one standard format for CD-ROMs and DVDs. Because of these network effects, ‘owning’ a dominant standard can be a profitable thing for firms. This means that firms tend to compete hard to establish their preferred standard in the first place. At the moment there’s hot competition between Blu-ray and HD-DVD to become the dominant next-generation DVD standard. Because of this intense competition ‘for the market’, the fact that a single standard eventually dominates might not be so bad. Consumers might benefit from the intense competition in the beginning, even if there is a lack of competition later. Also, history has shown that standards tend not to be dominant forever. Technological improvements mean that standards become obsolete and get replaced.

To me, the process of standards formation is quite interesting, especially when a standard involves multiple parties. Such standards are known as ‘consortium standards’ and depend on more than one key technology. Establishing such standards requires cooperation among firms that normally compete. They do so because they recognise the benefits of standardisation even if it means that competition between them once the standard is established will be intense. In other words, intense competition within a market can be preferred to winner-takes-all competition for the market.

ConsortiumInfo.org is an interesting website and blog about the process of forming consortium standards. They discuss in detail the events in standardisation processes and how the various parties use different strategies to try to manipulate the process. There are a number of strategic considerations that come up when establishing a standard requires the agreement of multiple parties. Essentially, the inputs of the parties are complements in the production of the resulting standard. What each does affects the outcome for the others. In the next post I’ll discuss some of the important strategic considerations.