Not All Network Effects Are Created Equal

Part of the excitement drivingLinkedIns IPOlast week comes from investorsassociating social media with network effects. You know, theprinciplethat the value of a network increases dramatically with the number of its participants. Thats the engine that drove Facebook and eBay, not to mention the public telephone network. Markets with network effects tend to have explosive growth and can end up with winner-take-all market share.

But not all effects are equal, and assessing the valuations and competitive positions of social media companies depends on knowing which network effects are actually at work and how those could play out.

First, lets look at a few of the different network effects currently in social media:

Core network effect utility:Theres a difference between economies of scale and the magic of adding connections to a network. Groupon is building a user base, a sales force and relationships with thousands of merchants, but until it uses its sales data to offerpersonalization, targeting and other marketing programsto merchants, it wont achieve much beyond scale. Even then, oncecritical massis achieved,additional connections dont add as much value.
Viral growth:LinkedIn and Facebook initially grew their networks the old-fashioned way: Users invited other users to join. Viral pass-along is a key growth driver for social commerce and games, but now services and apps canhitch a rideon existing social networks, leveling what was once a steep playing field.
Business model that reinforces the effects:While there are minimal network effects for its search users, there are huge ones for its advertising network. Googles$25 billionin extremely profitable search advertising depends on attracting advertisers to its dominant search audience and insuring a liquid marketplace via bidding and enforced relevance to create anunbeatable paid search business. Plus, Google lets developers using its services and APIs tap into that revenue stream with minimal effort.
Participant lock-in:Technology platformscreate positive business opportunities for developers. But they can also achieve customer lock-in for their originator by making those same developers dependent on APIs. End users can be locked in,! too, vi a familiarity (e.g., the QWERTY keyboard) and data storage (e.g., contact info, photos, message repositories) that raise switching costs for members.

As illustrated in the table above, heres how network effects are shaping competition in selected social media markets:

Social graph:Though there are network effects aplenty, consumers tend to belong to multiple networks (Facebook, Twitter, Foursquare), meaning would-be data miners must target multiple data sources. And the industry is only just beginning to harness thatcollection of big datainto reliable revenue streams.
Likes and log-in networks:Facebook was smart to hang Likes and Sign-ins off its Connect network, as each feature complements the others and assists in distribution; nowLinkedInandGoogleare trying to do the same. Bolting anad networkon top of those networks could provide missing revenue reinforcement.
Social commerce:As noted, most social commerce is more scalar than social. Without the additional services for consumers and merchants previously mentioned, single-market entry barriers and switching costs will remain low.

To read about more network effects and how those are shaping the current crop of social media companies, please see my latest Weekly Update at GigaOM Pro (subscription required).

Image courtesy of flickr user anselm

Related content from GigaOM Pro (subscription reqd):