[Freedombox-discuss] Network effect

[Boaz]

One of our most vicious enemies in the beginning will of course be this
thing called network effect, which I'll define as that property of both
communication protocols and walled-garden style communication services
wherein its value to an individual prospective user is increased the more
people already use it and decreased the less people already use it.  This
will be our enemy in the beginning because in the beginning no one will have
a freedom box and many people will already use unfree walled-garden
communication services.  Why would anyone get a thing which won't help them
talk to anyone?  On the other hand, if we clear the initial hurdle, in the
end game the network effect will be our most powerful weapon.  This post
contains my thoughts on how to overcome network effect in the beginning.

This is the framework in which I view our goals.  With respect to all the
forms of communication that people use in their lives, we desire all of:
1. That they use a free and open protocol, that they could run their own
servers.
2. That they do run their own servers.
3. That they encrypt their communications from end to end.

As an obstacle to the achievement of our goals, the network effect
specifically attacks that first goal.  The nonachievement of the first goal
would very effectively stifle our ability to work toward the other two, and
it is therefore quite imperative that we overcome network effect to achieve
the first goal.

One of our first orders of business will be to create and describe in
totality the set of protocols that makes up the way that freedom boxes
communicate, which for lack of a better term I'll call for now the freedom
protocol.  Looking over my own and other people's wish lists, it seems to me
that a large portion of the necessary protocols already exists.  SMTP
handles mail, HTTP does pages, SIP/RTP for voice, XMPP does instant
messaging and a sort of friend request and approval transaction, FTP for
file transfer, RSS does feeds though perhaps is not as rich a language as
Twitter or Facebook news feed and etc.  So as I see it we're most of the way
there on the front of creating the protocol suite.  And there are projects
already working on filling in the gaps and integrating into a unified whole
of a "social networking" protocol.

So once the protocol exists, our task with respect to the network effect
then, is to promote the use of this protocol.  How can we do this?

We're fighting an uphill battle, so we need to be not just equal but
superior to our unfree competition.  There need to be significant aspects of
the freedom box which are materially better in terms of utility and
experience than Facebook.  Also, many users will weigh into their decision
process degrees of importance of freedom box's privacy superiority ranging
from slight to compelling.  Best will be if people get diverted away from
our competition by censorship.  If it gets to the point where some small but
significant minority are already using freedom protocol, and then some group
for some controversial cause gets their page on Facebook shut down, bonus
points for us.  Likewise if Facebook as a whole is blocked in some country.

Clearly there will be a small number of people, like me and I imagine many
others on this list, who refuse to use the likes of Facebook no matter how
many people we know use them, and will buy one of the first freedom boxes
before knowing anyone else who uses one.  This can provide an initial jump
start, but this will not be a very large number of people.  Why should
someone who knows only one or two or worse yet zero other freedom box users
buy a freedom box?

The first thing that I foresee will be enormously helpful to us will be the
way in which an owner of a freedom box will be able to communicate with
preexisting adapters of individual parts of what make up the freedom
protocol.  Straight out of the box, the owner of the first freedom box will
be able to use his freedom-box-served email to write people, his
freedom-box-served website to make announcements, his freedom-box-served
XMPP account to chat with, for example, users of Gmail's chat feature, his
freedom-box-served FTP server to send people files, and so on.  He can buy
the freedom box for these uses, but by doing so he now also supports the
entirety of the freedom protocol and is adding to our network effect.  This
is why I see it as so crucial that we incorporate those protocols which are
already in heavy use.  We would be shooting ourselves in the foot to reject
SMTP for example, because we would be killing a powerful ally in overcoming
the network effect.

Just one single, really useful, really compelling use of a freedom box could
propel the freedom box to prominence in no time.  Just as an example, if
people have trouble sending large files to each other (no problem, i'll just
email it. oh wait, the size limit. okay, i'll use one of those web services.
oh wait the size limit. damn) and the freedom box presents a powerful, ultra
convenient solution to this problem (imagine right clicking on a file in the
file browser and clicking "create FTP link and copy to clipboard"), people
might pay $29 just for this one feature.  But by doing so, they would become
supporters of the entirety of the freedom protocol and add to our network
effect.

Many people propose that as an aide to overcoming the network effect the
freedom box should provide an interface which aggregates communications
through unfree channels with ones through free channels.  Perhaps, but this
could also just encourage those unfree channels to linger like a plague.
 Imagine that a small but significant minority is starting to use the
freedom box.  What encourages someone who doesn't care about the goals of
freedom box, or does but is lazy, to get a freedom box to talk to the people
he knows who use freedom box if he can just keep talking to them, through
the same interface on their side, using Facebook?  In any case, for the help
of the person who already uses Facebook and now is trying to switch to
freedom box, nothing stops him from having the freedom box interface open in
one window, and a web browser opened to Facebook open in the next window.
 He's just using a new kind of communication, the freedom box, in addition
to the old kind.  Mike says:
>I'd be trying to "sell" them on a box which would allow them to talk
>to me, but cause it to be harder (or at least no easier) to talk to
>all the friends they have on Facebook.
But that's not true: the box in no way makes it any harder to talk to their
friends on Facebook, and as for making it easier, well, why should the
freedom box make it easier to use Facebook?  Also, asking someone who used
to use the Facebook web interface to switch to an aggregation of Facebook
and freedom box inside the freedom box interface is asking him to change the
way he uses Facebook.  Perhaps more useful would be to ensure that the
people who are already using social networking aggregators can integrate
freedom box into the aggregate.  Better, if possible, not to ask people to
change how they do things they're already doing.  And we would certainly
want to be very careful that no one finds himself encouraged by the freedom
box interface to use an unfree service that he was not already using.

I think it's much more important, rather than to make it easier for a
freedom box user to communicate through unfree channels, to make it easier
for a user of unfree channels to communicate with users of freedom box.

Imagine a hypothetical person who knows one or maybe two people who use
freedom boxes, and doesn't want to spend $29 to talk to one or two people.
 What for this person?  One feature that would really help there is a way
for someone who has a freedom box to extend an account for the freedom
protocol to someone who doesn't have a freedom box.  It needs to be dead
easy for both parties.  And for privacy, it needs to, when possible, encrypt
the person's traffic from the people he's talking to all the way to his own
computer, not just to someone else's box.  I see this as being a powerful
way to expand the user base and create network effect in our favor.  When
one of the first freedom box users who doesn't know any other freedom box
users is trying to connect to his friends in freedom, perhaps his closest
friends will set up accounts on his own box just to talk to him.

And, for a person in the position of the hypothetical person from the
beginning of the last paragraph who doesn't know any of those one or two
people very well or simply doesn't like relying on (even very small) favors,
imagine an account provider who provides an account for the freedom protocol
on their server.  This situation utterly fails at the second goal, and is
not at all what we want in the long term, but it is potentially (at least
partially) compatible with the third goal and does satisfy the first goal
and so can grow our network effect.  It allows this person to communicate
using the freedom protocol to that one freedom box user he knows, and
increases the value of a freedom box to anyone else who knows him and may be
interested in buying one.  This service, provided by the "businesses" that
Bjarni is so excited about, can be a stop gap measure which helps us to
achieve the first goal.  That hypothetical person might buy his own actual
freedom box at some later time.

So these are a few of the things that I think might help us overcome the
network effect.  And now I'm about to propose one more strategy that I think
could compliment those others quite nicely, and prove a powerful force in
our favor.

Enter the smaller closed garden social networks.  Of what interest could
they be to us, you ask?  In the US market at least, the network effect of
Facebook is so powerful that it's becoming hard for any other closed garden
social networks to compete.  As time goes on, this effect will only
increase.  They will feel the heat, and they will be desperate for their
very survival.  How can a small closed garden social network hold a candle
to the network effect generated by the behemoth mass of user base that
Facebook controls?  What I propose is that we approach them with our
protocol, and we make a simple proposition: "You, so and so closed garden
social network, on your own, will dry up like a worm on a sidewalk under the
heat of the Facebook sun, and so will all the other closed garden social
networks.  But you, throwing together your network effect with the network
effect of all the other social networking websites other than Facebook, have
a chance of competing head on with Facebook, surviving and flourishing.
 Adapt the freedom protocol and become interoperable with all the other
social networking websites that adapt the freedom protocol and with users of
freedom boxes."  Armed with our protocol, the swarm of small, currently
walled garden social networks will band together for their own survival.
 And in doing so, they will harness their existing user base to create a
mass of network effect which can feed the growth of freedom box.

The people at http://onesocialweb.org are trying this approach.  Without
commenting on the technical merits of this particular project, which I'm
perhaps not qualified to do, I find it noteworthy that they're pitching
their protocol to operators of walled garden social networks.

Imagine a world where a free and open social networking protocol was already
the standard.  In such a world, the task of moving people over to running
their own servers, while still facing significant challenges, would not need
to make enemies with the mighty network effect.  And it goes without saying
that in such a world the likes of Facebook wouldn't stand a chance.

What do people think of this strategy?


Boaz
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/freedombox-discuss/attachments/20110320/ad6a59cf/attachment-0001.htm>