Caveat: IANAL
In cyberspace, such barriers are basically imaginary. "Network effect" originally referred to physical phone lines, i.e., hardware that can't be modified/upgraded remotely. Within the internet itself, there is no such limiting factor.
People could decide to switch to another social client easily if they could access the same data, and the only reason they can't access that data is not an incontrovertible "network effect", but our own conception of intellectual property and server access rights.
Facebook's provided personal data export is slow and designed for use in relatively-rare periodic backups. It is noisy and difficult to access. Users cannot use an alternative client to trigger this on their behalf afaik (I don't see an API function for it) and programmatically controlling a web browser to access Facebook's site is explicitly forbidden by the ToS, making it a CFAA violation as well as a breach of contract.
If we're being realistic, the functionality of the data export represents an impractical barrier for Another Social Client to overcome, but let's say they try to go with it, exporting your user data every 24 hours so there is only a one-day lag time between the networks.
It would probably be illegal for this Alternate Social Client to process the files output in this archive, as they likely constitute a copyrighted work owned by Facebook, and copying content owned by someone else without authorization, even if you just copy it into RAM for processing/modification, is likely to be infringement. This is known as the "RAM Copy doctrine".
Thirdly, the Alternate Social Client is unable to provide a competitive window into the user's social graph because every other user would also have to go through this contrived process before their data would be visible within that network, because the CFAA makes it illegal to gather this data out of Facebook directly.
This is mostly not theoretical. Check out Facebook v. Power Ventures, where a small entrepreneur who tried to provide a better Facebook export interface, only collecting data which the user owned, was sued into the ground by Facebook and left holding a bag of $3M in personal liability. This entrepreneur violated the CFAA by contacting Facebook's servers in a manner that was contrary to their wishes, and he violated the Copyright Act by loading their HTML into his software for the purpose of extracting data owned by his knowing users who had purposefully asked him to perform that function on their behalf.
We need to divorce data from applications, such that Facebook is a Social Graph Browser, and users are free to move around and select a different Social Graph Browser, as they wish, just as My Word Document Collection can be viewed without legal penalty in either OpenOffice or MS Word (though this is probably legally dubious too and just hasn't been pressed; lmk if there's been a case about interoperability between proprietary file formats).
Using a remote browser to access the Social Graph shouldn't be any different than using a local browser to access the larger network that Graph lives on, the Internet.
In the earlier days of the internet, where everything was a document without a login barrier or remote database lock, users were free to select the browser of their choice to present those documents; they were free to decide how that data would be processed and presented to them.
Through some legal sleight-of-hand involving moving this stuff onto server backends and accessed remotely instead of being installed locally, the tech conglomerates have taken that data and put it under lock and key, within a walled garden, in a closed web atop the normal internet, or whatever you want to call it, and the people who attempt to liberate it are prosecuted under laws such as the CFAA (see: Aaron Swartz). In doing this, they have taken away our ability to choose how we access that data. [Sidebar: this approach has also negatively affected copyleft licenses like the GPL, giving rise to the AGPL.]
We won't have competitive SaaS until we get to the point where choosing our own client experience is possible again. Many entrepreneurs have already done the plumbing to make it possible without incurring any development overhead on Facebook or any other data source; the legal system prevents them from realizing that commercially.
Facebook, Craigslist (see Craigslist v. 3Taps), et al are terrified at the prospect of losing their monopoly on user data, because they know that means they could lose users to people who provide better clients with friendlier user experiences, less spyware, or whatever. They won't have their mechanism of lock-in anymore (other than the conventional arms race thing where they try to screw up the format so that competitors have a hard time reading it).
Reminder caveat: IANAL