The arc of social

The era of social media feels like it’s coming to an end, but it shouldn’t because I’m not done with it yet. Recently Eugene Wei’s Invisible Asymptotes and Kevin Kwok’s The Arc of Collaboration gave me the language to decipher why I feel like there’s still a ton of white space. This essay is my attempt to break down that feeling and articulate what I think is missing. 


Social products are powered by a web of connections that you construct via following, adding or connecting with other accounts. In Silicon Valley, we call the result of these connections graphs and we use it to explain why products succeed or fail. Facebook succeeded because we used it to build our “friend graph,” which is one of the most powerful graphs because everything is better with friends. However, as Facebook grew, our moms joined, and our friend graph became less about friends but more just our…generic social connections. The friend graph turned into the generic social graph. We went from wanting to share everything with our friends to becoming much more careful about what we put on Facebook because our moms and a bunch of miscellaneous connections can now see it. 

This would’ve been Facebook’s asymptote, a term Eugene Wei used to describe a ceiling to growth that are often difficult to spot. To Facebook’s credit, they actively spotted it and acquired Instagram and Whatsapp for their smaller and more meaningful friend graphs.

Our friend graphs on Instagram are smaller and more thoughtfully composed because 1) it was built in the context of sharing photos where it’s more accepted to keep your account private 2) usernames as opposed to real names enabled you to hide more easily and 3) it was a new network with a new chance to reject follow requests just in case you regretted friending someone on Facebook. Instagram needed Facebook to co-exist to take the pressure off of friending for it to get to be the smaller, more intimate graph. 

Whatsapp approached it from another direction by starting with your phone’s contacts as the basis of your friend graph. For most people, that was pretty representative of who their real friends are. 

Of course, as Whatsapp and Instagram continue to grow, they run the risk of diluting the meaning of their friend graphs as well, if they haven’t already. Instagram users are fighting back with finstas (i.e. private “fake” Instagram account often used alongside a person’s main account but with a more carefully curated follower list) and by taking advantage of the “Close Friends” feature.

Graphs are important because social products live and die by the graph. Facebook lost an entire generation of young people to Snap because they were turned off by Facebook’s big friend graph. Facebook never successfully squashed Twitter because Twitter has a powerful interest graph. Facebook does not have many advantages competing with LinkedIn and Slack because they don’t own the work graph. Yes, Facebook has a ton of data about what you’re interested in and who your co-workers are, but their generic social graph limits their ability to create meaningful contexts for other graphs.

Social products and their graphs (sorry can’t draw because I don’t own an iPad)


Building products is crafting contexts for your users to operate in (b2b SaaS products are essentially “opinionated professional workflows”). In other words, context is the answer to your users’ question, “what’s this thing about?” and you answer with your product decisions. For example, if your app opens first to the camera like Snapchat, your users will know that photos is the standard message and your app wants them to create first before they consume. Social products built on top of graphs need to have contexts that fit the underlying graph.

Facebook changed its context several times throughout its history. If you don’t use Facebook as much today, it’s not only because you’re older, but also because Facebook today is not the Facebook from ten years ago. Ten years ago, we used to bounce around different profiles and write on our friends’ walls. Profiles and walls were semi-private spaces by requiring a few more clicks of effort and intention from the user in order to visit. Because visiting took effort, profiles and walls felt like they belonged to our friends, which led us to write more personal messages on them.

Then Facebook introduced the News Feed, which turned “pull” into “push.” We no longer needed to intentionally visit our friends’ spaces because their updates were pushed to us and everyone else. This was a rational decision to help engagement keep pace with the growth of the friend graph, but what we lost along the way was the ability to share personal thoughts because we knew the feed would amplify everything. What used to be intimate updates only to friends who cared enough to come see them became basically public broadcasts. 

With a bigger, more generic social graph, the contexts on Facebook that were meant for the smaller friend graph had to take a backseat. Decisions to default sharing to public, prioritize viral content, and introduce features like Facebook Marketplace and Facebook Watch can all be understood as the consequence of Facebook moving on from the original friend graph and needing to create new contexts to engage to a much bigger graph of people who are only loosely connected to each other. I miss the old Facebook.

Contexts have to fit the graphs


I yearn for a social product with a smaller, more meaningful friend graph and a variety of contexts to help me deepen my connection with friends. I want to feel like my friends are online and hanging out with me like I once did when I opened my AIM Buddy List. I want to know and engage with my friends’ vulnerable thoughts like I did when we all had LiveJournals. I miss knowing that my friends updated their favorite movies and music like they once did on their Facebook profiles. Wow, way to date myself, but I miss those days. Snapchat is the only product that does this well today, but their graph mostly consists of 18-24-year-olds and it doesn’t seem like they can age up their appeal without alienating their base [1]. So who’s it gonna be?

I feel old

Facebook probably can’t do it. They introduced the “close friends” list, but it was underutilized and now buried because there is very little context left on Facebook that is meant for us to use with our real friends. It’s near impossible for Facebook to serve two masters at the same time (i.e. generic social graph and friend graph). Although Facebook’s strategy to launch new standalone apps to create new contexts off of a clean slate could be viable.

Instagram might be able to. People my age (early-30s) are already on Instagram. However, Instagram suffers from having begun its life as a public photo sharing product, and much of its older user base is still using it that way. Stories, Direct, multiple account support, and the Close Friends list have shifted the product in the direction of more intimate contexts, but that’s going to come to a head with the public use cases, especially the IG influencers who drive significant growth. Snapchat famously separated the social from the media in a controversial redesign, which was a move to clearly delineate the friend graph context from the more generic social context. Instagram might have to do something similar. There’s only so much context you can pack into a product especially on mobile.

A wild card candidate could be organized chat products like Slack and Discord. Their graphs today are the people you work and play games with, respectively, but chat is a very flexible and the lowest common denominator type context, so users already use Slack / Discord for purposes other than work or gaming. I’ve set up a Discord server with a small group of friends and I know other people who have done the same. You can essentially build your own graph simply by inviting people and set your own context through moderation. The downside is that it puts the onus on the user to create a fulfilling experience. Barring significant product changes or improvements in their bot ecosystem, that extra friction could be the asymptote for Slack or Discord.


Maybe we need new entrepreneurs to take on the challenge and re-think graphs, contexts and growing a social product.

Maybe the problem of creating and maintaining a list of friends is not the right way to think about building a graph to begin with. Google Plus tried to make creating lists of friends fun with Circles and Path tried to limit the graph to 50 friends. Both companies understood the problem of mixed contexts, but their proposals still revolved around lists, and deciding which friends to include / exclude on a list will always be a deeply uncomfortable exercise. Once a list is made, it is also difficult to maintain. Real relationships are much more fluid than our ability to manipulate lists. We need more fresh thinking here.

Maybe the right context for real friends is email. I use Substack to write a personal newsletter and it’s been extremely rewarding for me to share thoughts that are closer to my chest directly with a smaller number of friends, without any performance pressure. Email as a medium already screams “thoughtful” and “personal.” It brings back the feeling of semi-private communication delivered from my inbox to yours and will most likely stay just between us. The assumption that your friends care is built right in, so you don’t need “likes” to be validated. 

Maybe we need a brand new way of understanding how to grow a social product generally if we were to build products with smaller graphs. Snap kept the user experience relatively inscrutable to discourage moms from moving in, but they are pursuing growth via other means like taking bitmoji and stories into other apps. An innovative product like is experimenting with the Tree Explorer, which is how you can discover content without relying on social signals like “follows” and “likes.” 

I feel underserved by social products, and I’m sure many of you do, too. The possibilities feel limitless. We’re still in the early innings.

[1] Snapchat is the company that has had the most opportunities recently to build new contexts for an intimate friend graph, but unfortunately their product simply doesn’t appeal to my (old) friends yet, and I wonder if it ever will. Snapchat’s ambition is much more expansive than how companies like Facebook have thought about it in the past. They don’t confine you to within Facebook’s walls but instead wants you to take your real friends with you to the real world through the camera. I wouldn’t count them out. 

Thanks to David Tran, Wil Chung, Shravan Reddy, Jessica Tsai, Ryan Witt and Joey Shurtleff for reading drafts of this essay. Also thanks to the essays that inspired me: 

Invisible Asymptotes by Eugene Wei
Arc of Collaboration by Kevin Kwok
Status as a Service by Eugene Wei
The Form of Formats by Drew Schorno

2 thoughts on “The arc of social

  1. Thanks for giving us language to discuss this void. Facebook is now 15.5 years old, and despite having 2B+ users, doesn’t seem to have the right graphs or contexts for the types of interactions we crave with close friends. So, the question is, who’s got next? Is there something better than all of the disparate friend group chats across iMessage, Messenger, Telegram, Instagram, Whatsapp, Slack, and Discord?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.