Building Another Social Network

Don’t be encumbered by history. Go off and do something wonderful. — Robert Noyce

I argue that founders interested in building a new mass-adoption social network should focus obsessively on the problem of “keeping up with your fifteen closest friends and family”.

I’ll then argue that the social network game is not over: players like Apple and telcos are still in the running.

Finally, Facebook, by letting its Messenger app cannibalize the original Facebook site and app, may already being using the above strategy.

Why We Used to Join Social Networks

The average person now has 338 friends on Facebook, and 50% of users have 200 or more friends. Interestingly, the median user’s friend count already exceeds Dunbar’s Number (150), which is hypothesized to be maximum number of people with whom a person can typically have high-trust relationships.

How did we get here? In the early days of social networking (2004-ish), the very idea that you could stay in touch with lots of people, living far away, whom you didn’t meet regularly, was amazing. Reconnecting with school friends from whom you’d drifted away over five or more years was a novelty.

So, like a starving person at a buffet, we “friended” a lot of people.

One Ring Doesn’t Rule Them All

Facebook is the single social graph for our close relationships and our more-distant ones. It’s intrinsically hard for the same system to serve both sorts of relationships well, because we present different versions of ourselves to our close circle and our more-distant circles.

IMHO Facebook at present does a better job of serving our more-distant relationships, which opens an opportunity for new social networks to better serve our close relationships. Snapchat and Whatsapp are good examples.

Notwithstanding the racy image Snapchat cultivated with their disappearing images, I’d argue that essence of the service was the ability to better nurture close relationships. Whatsapp started out a cheap way for users to stay in touch with people close to them.

Snapchat has since diluted their essence by added the ability to follow brands and broadcast to the public. Whatsapp added groups, that drove huge growth. Groups is mailing lists redesigned for a mobile-only, low-bandwidth world: sending group SMS’s is prohibitively expensive but free with Whatsapp.

I expect Whatsapp and Snapchat to double down on such public features, to increase the average time users spend in their apps. The bigger the pool of content that a user can be shown, the more content they will be shown, resulting in more time spent in the app.

The Opportunity

As these networks add public-focused features, their original role of nurturing close relationships will get diluted. That’s the weakness that a newer social network can most easily exploit to gain traction.

Since the number of close relationships is very small, switching costs for people are low: you only have to convince a few people to install a new app. And since these relationships are important, people can convince others to switch.

For instance, if you sister uses some super-private, not-very-popular service to share her kids’ pictures, you’ll probably install that app.

What Should a High-Trust Social Network Do?

I’ll call social networks geared to serve your close circle “high-trust social networks (HTSNs)”. To be useful right away, a new service probably needn’t do much else than provide decent versions of two features: photo and video sharing, and instant text messaging.

To stay true to it’s HTSN mission, it should have a disciplined UX that discourages people from adding more-distant relationships.

Here’s one idea: your entire friend list will be shown on many screens. Each friend gets a square icon with their picture. If you have many friends, their photos will get smaller. Your friends will literally become fuzzier the more people you add.

And another: perhaps users should be disallowed from belonging to more than, say, six high-trust networks, analogous to how a single person is unlikely to be a member of too many extended families.


Since users’ networks are small, the average user is unlikely to spend too much time on this service. While the info from this network is likely of high value, the number of messages shown to a user per day will likely be relatively low.

On the bright side, due to these same reasons, the service will also be inexpensive to run. Perhaps so inexpensive that requiring just a single network member to pay a modest subscription fee will be sufficient.

The fact that users pay will pay very close attention to notifications from this app (since they’re from close relationships), means that advertising minutes from this app can be sold for a high price. That might result in a virtuous cycle of junk ads getting priced out, and the ads becoming compelling content themselves (e.g., the photos in fashion magazine ads are interesting even if the product is uninteresting).

Another obvious approach is to not monetize. A company could provide this as a perk to their existing customers (e.g., Amazon and telcos), or businesses could build such a service to to keep their brand in people’s minds (e.g., Coca Cola).

Facebook’s Recent Moves

Facebook probably recognizes that interactions with one’s close circle suffer when all relationships are represented in a single single social network. Splitting off core, frequently-used functions into separate apps (Messenger, and recently, Photos) may indicate this.

Each specialized app essentially interfaces with de facto different social networks. The effect is very pronounced in Messenger’s design. You always choose whom to talk to. Grouping friends is very easy to do, unlike in the regular Facebook app. By simply using that app and having multiple group conversations, you solve many of the problems with a single social graph.

Facebook seems to letting these specialized apps, more geared keeping in touch with one’s close circle, cannibalize the regular Facebook experience. If that is indeed the case, Facebook will have to solve the monetization problems listed earlier. Current ad formats, such as ads between News Feed items, or promoted stories will probably seem too intrusive.


Building a new mass-adoption social network is not a fool’s errand. It’ll be fun to build, useful when finished, and who knows, maybe even profitable.