This is a post that I wrote back in 2020 reflecting on my own startup that I never published. I’m publishing it now because is suddenly a lot of energy in the space. Threads! Bluesky! Mastodon! Let’s go!
Unpopular opinion — social media is ripe for disruption.
The massively successful platforms have been around for a decade and some. Many of these platforms are being weighed down with problems inherent to the design of the platform. Platform innovation is grinding down to a halt. While revenues are booming, they are lagging indicators to audience attention. New successful platforms tend to be no where and then suddenly everywhere at the same time.
Social is hard, very hard. Like a pot at the end of a rainbow hard. For a while, there was seemingly a new social media platform every other week; Today we barely hear them. At this point, VCs are publicly disavowing from investing in them. Ben Horowitz, in an interview with Kara Swisher in 2019, states that he thinks the opportunity for social networking is pretty much over (starting at 05:33). (That said, A16Z funded Clubhouse in May of this year)
The world is more connected than ever. There is a Goliath in the marketplace, yet, it is still possible to build the next big thing. Let us find out how.
Part 1: What is social media
I find it useful to first lay down definitions. A social media platform that we are discussing here ranges from a messaging application like iMessage to a broadcast platform like YouTube. While there are clear differences that separate the two, the shared commonality is that both platforms are used to communicate ideas (within its capabilities) between multiple different people. iMessage is mostly text and intimate; YouTube is video and performative.
Back when Snapchat was emerging, Ben Thompson proposed a framework to categorize social media applications between the axis of public/private, symmetric/asymmetric and durable/ephemeral. In organizing my thoughts, I found it useful to generalize further. Social media platforms can be thought of as containing two main constituent parts: the content (format) and the network (distribution).
The content part is self evident, and the obvious candidates are video, audio and text. It is useful, however, to add nuance to this definition. In my definition, content is the least constrained format that the platform is designed to deliver. For example, TikTok has a 90 second limit and is viewed in portrait. Traditional movies or TV shows will be ill suited to be delivered over that platform. In a separate hypothetical scenario, a platform focused on art and drawings (either through technological or moderation means) has a different content format than one based on photography. This is important and we will discuss this at length later.
The network is how content reaches the people the creator intends to reach, in other words the distribution channel. There are two properties that are useful to discuss here. First, is the design intent of the distribution mechanism, which ranges from the spectrum of 1-to-1 to 1-to-N (symmetric/asymmetric). Second is the reach of the network and the user base of the platform. Networks have to reach a critical mass for it to be useful; however it doesn’t mean that bigger networks are better, it depends on the design intention of the social platform.
People maintain multiple networks effortlessly. Since applications are the vessel to these networks, in practice this means that most people maintain multiple applications and switch between them. In my research on messaging apps, I found that most people maintain at least three different apps for the same purpose. One for close friends and family, one for work or school and one for acquaintances. More interestingly, the app choices are dictated by how the app took hold in the relationship more than anything else. While David may use Facebook messenger to talk to close friends, Jane may reserve this app for people that she knows less well.
Social platforms can be built off either one of the two factors. Examples of content first social platforms include iMessage (text), Instagram (photo), SnapChat (ephemeral video), YouTube (durable video). Examples of network first social platforms include LinkedIn (professional), Xing (macro-geographical) and NextDoor (micro-geographical).
Part 2: the business of social media
It would be an incomplete discussion if we omit the physics of survivability—monetization. Even with perpetual calls to shift to subscription business models, current successful networks monetize access to the network. Facebook with advertising, LinkedIn with recruiting. Even with national sized networks like WeChat, revenue is generated through facilitating financial transactions through the platform instead of access to the platform (maybe they should just institute a service fee through the tax system).
Past performance has indicated that the most lucrative opportunities lie in maintaining a captive audience attractive enough to build services on top of them. In this case, because the additional user provides much more upside than the marginal cost of use, providers are incentivized to price entry into the network to as low as possible. In almost all cases, that number is free.
Because the medium of exchange is media, the resource that is being granted access to is attention. To monetize attention and make it effective, it has to be something that is native to the platform. I suspect this is why most social networks eventually build broadcast products (e.g. Snapchat Stories). It is an interruption seeing advertising when in conversation; however, if you are watching a series of short videos, an ad is just another video in rotation.
Part 3: opportunities in Social Media
There are three ways that social media platforms can be started:
The first and most obvious are platform transitions. When a new platform starts to take off and people are adopting the platform en masse, it creates a void in which these users have a lot of appetite to grab applications to fill needs. The rise of the smartphone obliterated many venerable communication applications for the desktop era. However, these transitions are few and far between. It would be unwise to build a new company in anticipation of an event like this.
The second opportunity comes from building off existing networks. This can be connecting college students online or having a platform for tech elite to pat each other on the back. In these cases, the network is often the means and ends are the platform and technology that facilitates the member interactions. Students want to socialize, elites want to brag. The best opportunities here build of new, emerging use cases with a good appetite for tools tuned to the use cases. Discord building on the backs of gaming turning social would be an example.
The last opportunity lies in inventing new formats. The most interesting innovations in this space has been counterintuitive. Photos that are ephemeral. Videos with short time limits. Many new entrepreneurs in this space tend to want to specialize, for example, building photo sharing for families or ephemeral text for businesses. However, unless the scenario has specific qualities that can be targeted at, a lot of these scenarios collapse down to the most acceptable generalized version of the tool. It is immensely difficult to activate and move a network into a new platform, and to do so, the new platform must be unusually attractive.
Part 4: Designing for Social Media
On one hand, designing for an established network and a known format is straightforward; existing design methods work. By deeply understanding how users use the platform, designers can create new features that will either extend the format to enable richer interactions or build new tools to complement the core format. Existing users will be better for it, the deeper and richer content may tempt people who are on the fence about joining the network, and a broader tool set lowers switching needs and lock in the network.
On the other hand, designing a new social media platform is maddeningly difficult. When I was working on Twist, we leaned heavily into user research methods. We started with doing deep ethnography into potential audience groups. With insights from the research, the team generated ideas and tested them with audiences. During every phase in development, we brought prototypes and put them into the hands of potential users, including setting up dyad and triad studies. Throughout the process, we followed the signals that we thought were strongest, yet when the project launched it fell flat.
With the benefit of hindsight, conventional user research methods were unsuitable for designing social media tools. Participants in social media sessions are placed in an artificial environment where they get to dig into and learn a new app; it is easy for people to learn how to play with an app and have fun. The limited time and participants makes it difficult for us to understand if there are durable usage effects and any built in virality in the design.
Instead of thinking of social media platforms as tools to connect people, they are more accurately experiences that connect people. Social media platforms take flight when the early adopters get strong emotional feedback from using the platform. For people who like to express their opinions or creativity it can be in the form of validation from peers or the broader public; for others it may be the dopamine hit of new content or closeness with their peers.
Social media design is experiential and emotional. It requires at least two parties on each end to make the connection work. Instead of helping people achieve specific goals, they are tools that help people express ideas to others. Social media design also needs to be novel. People are already very good at using existing tools to talk to each other; the bar set for anything new that needs to have an impact is impossibly high as a result.
The only design method to design for social media platforms is to build and iterate in the open as quickly and nimbly as possible. The goal of this method, if you may even call it that, is to first pique the interest of potential users–people who may be attracted to your format or novel ways of doing things. If the design elicits amusement or good emotions, the social media innovators will start soliciting her friends to join her network. If the design has staying power, these people will then bring their friends along.
In practice, the amount of resistance is exponential in every step. While the innovator audience may be willing to give the app a try, the people within her social network may be reluctant to download and set up yet another app. Even if the application takes hold within a specific network, perhaps buoyed by the strength of the friend group for example, the possibility of the app stopping at the margins of the group is very high. The goal of this design process–building and iterating in the open–is to get in front of this problem before the interest and energy of any new potential users fade.
A common pushback is a fear that if the product is not mature enough, newcomers will not be able to see the product vision, have a bad first experience and thus be “burned” by it. A similar concern from the engineering standpoint is that imperfect features will lead to bad user experience and thus prevent the product from succeeding. When we were building Pulse, we were so worried about the battery life being a drag that we spent much of our time improving our battery usage. A great technical achievement no doubt, but hardly helpful to finding our product market fit.
Iterating in the open stands in contrast with typical product development process that waits until the product is mature before releasing it. The fear of imperfection is overblown and the trade off is immense. Working with limited sets of participants restricts the speed and quantity of feedback that the team may receive. Social media apps are similar to bands in their early days. To be successful, the band will have to go in search of their 1000 true fans. Because social media apps tend to follow the 90-9-1 rule, doing so openly is the only way to guarantee that you can find your way to the next phase before the band breaks up.
Put another way, a team only has so many energy credits they can put into product development before the team gets burned out and the venture fails. As responsible team members, the only way to tackle a social media app discovery process is to put as much energy credits into this process before the clock runs out.
There is no guarantee that this method will lead to riches. Like I said in the beginning, starting social media platforms is really hard. Often times it may feel like the current landscape is permanent and change is futile, impossible even. The big blue has too much ground and too ruthless to copy.
Remember however, the antithesis to size is agility. Given where we are in the cycle, I suspect there are opportunities hiding in plain sight, in between the cracks and places where we are not looking. It is utterly wonderful to invent culture, so seek feedback often, chase the fun, and enjoy the ride.