The Network Liquidity Map: Why Time Matters
Synchronous networks like Meerkat and Clubhouse are susceptible to liquidity challenges because they need users to be active at the same time — pivoting to emphasize asynchronous features is one way to overcome these challenges
In the startup world, time is primarily viewed as a hurdle to be removed — everything needs to be instant and real-time. This is certainly valuable as a general principle, and it can even be critical to the defensibility of certain types of startups, i.e. data networks. However, blindly applying this principle in all situations can create complications. Time-delayed behavior is sometimes a requirement to gain critical mass, in particular for interaction networks — ones that connect specific users to enable interactions, e.g. social networks.
I have previously explained how network structure influences the potential of network businesses — this includes the presence of network bridges, importance of user identity, nature of connections, and network density. However, this is irrelevant if a network cannot sustainably build critical mass or liquidity in the first place, i.e. it needs to have a minimum density of users that can interact with each other on an ongoing basis. While some elements of network structure can have an impact on liquidity, they are better viewed as secondary constraints. The primary determinant of liquidity is how “real-time” an interaction network is — whether it is asynchronous or synchronous.
On asynchronous networks, interactions do not depend on the time when users are active on the network. Users initiate and respond to interactions as and when they are available. Take Snapchat as an example. Snapchat popularized the concept of disappearing messages or the “ephemeral” social network. Ephemerality is certainly a core part of Snapchat’s value proposition, but it does not affect its relationship with time. This is because Snapchat messages disappear after users see them, whenever that is (with some restrictions). Unopened snaps are only deleted after 30 days in a direct message and after 24 hours in a group chat, leaving users more than enough time to view and respond. Even its widely cloned Stories are only deleted after 24 hours. So despite being ephemeral, Snapchat is asynchronous — its users do not need to be active at the same time to interact.
Since users do not need to be online at the same time, liquidity purely depends on the number of connections available for a user, i.e. the number of people someone can interact with. There are no other barriers. As a result, the vast majority of networks that have scaled successfully, whether B2C or B2B, happen to be asynchronous — this includes Facebook, WhatsApp, Linkedin, Paypal, Carta, Nextdoor, TikTok, Slack, Figma, etc. Real-time interaction can be a feature of these networks — typically added after they have achieved sustainable liquidity — as long as it is not the primary value proposition, e.g. Facebook live, Slack calls, WhatsApp calls, etc.
On synchronous networks, interactions can only occur when users are active at the same time. Discord is one example — a group chat service originally designed for multiplayer gaming. It allowed players to communicate and discuss tactics in real-time, i.e. while they were playing. This meant that its primary value proposition required users to be active at the same time.
For synchronous networks, liquidity depends on not only the number of connections a user has but also the number of those connections who are active at the same time. This is especially problematic when user identity is core to the network, i.e. people want to connect with or follow specific users. This creates an additional barrier to liquidity that asynchronous networks don’t have to grapple with. As a result, synchronous networks typically need to target users that plan to be or are already online together to reach liquidity. Examples of these use cases include multiplayer gaming (Discord), utilitarian video chat (Skype), or workplace collaboration (Around, CoScreen). In other words, interactions need to be planned or occur during scheduled online times. They cannot be truly spontaneous, at least until they achieve sustainable liquidity.
The ”Spontaneous Togetherness” Problem
Meerkat, Periscope, and Houseparty are great case studies of what happens when synchronous networks also aim to be spontaneous. Meerkat was built around the concept of “spontaneous togetherness”. It was a synchronous video livestreaming network that launched and went viral in 2015, piggybacking on Twitter’s social graph. In response to Meerkat’s sudden popularity, Twitter acquired a competing service called Periscope and cut off Meerkat’s access to its social graph. This certainly knee-capped Meerkat, but did not result in sustained success for Periscope either. After a brief wave of popularity, Periscope rapidly lost traction and its livestreaming capabilities were folded into Twitter’s own app. Livestreaming then went on to become a viable feature within Twitter (and other social networks) because their primary use case was asynchronous. It already had a critical mass of users browsing their feed and asynchronously interacting with other users. The existing liquidity made it easier to produce and consume truly real-time content.
Meanwhile, Meerkat’s founders shut it down and pivoted to Houseparty — a private broadcasting network. Houseparty was meant to be more than just another video calling service. The bet was that “spontaneous togetherness” was a better fit with a private social network as opposed to a public broadcasting network. Houseparty had lower requirements for user density because it focused on one-to-one and group conversations. It even attempted to maximize concurrent usage by notifying users when a connection opened the app. However, the results were still poor — after briefly topping the app store charts, Houseparty lost over half of its users. Epic Games eventually acquired Houseparty for $35M, half of what it had raised in funding. Houseparty did see a brief resurgence during the peak of COVID-triggered lockdowns but has since tumbled down the app store charts again.
These examples show that the implementation of “spontaneous togetherness” was never the problem — Meerkat/Periscope and Houseparty were two different implementations of the same basic concept. The real problem was that “spontaneous togetherness” is an oxymoron — you can have one or the other, but not both. Meerkat, Periscope, and Houseparty required users to be active at the same time in contexts when concurrent usage was not likely. This creates a liquidity and sustainability problem for fledgling networks — their novelty helps them go viral, but their synchronous nature makes it very difficult to retain critical mass. This is also one of the causes behind the failure of more recent VR social networks. The installed base of VR devices is very small, but requiring users to be active together exacerbated that problem even further. Previous iterations, like Second Life, only survived because users treated it like a multiplayer game with standalone exploration value and not as a synchronous network.
The Network Liquidity Map: Overcoming “Spontaneous Togetherness”
Can entrepreneurs overcome the “spontaneous togetherness” problem? As the liquidity map below shows, they can take one of two paths to improve liquidity. The first is to target a different segment of users who are likely to have more planned, real-time interactions, e.g. gamers, collaborating workers, etc. The second is to pivot the focus of the product to emphasize asynchronous interactions over real-time ones. While pivoting, startups need to be careful not to dilute the value of user identity with their new proposition — this would make them vulnerable to copycats. At the same time, pivots should avoid replicating the value proposition of larger competitors with strong network effects.
Clubhouse: Route to sustainable liquidity
This brings us to Clubhouse, the most recent iteration of “spontaneous togetherness” — a (still invite-only) network for real-time, audio-only conversations. Given the nature of its early users, it has managed to host numerous conversations that have captured the imagination of VCs and tech community. This led to an unprecedented $10M funding round from a16z, even though it had less than 5,000 beta users at the time.
It is important to keep in mind Clubhouse is a synchronous network that requires users to be active in real-time — and as it expands beyond the tech community, it will require a critical mass throughout the day and within multiple interest niches. In its current iteration, Clubhouse is likely to face many of the same challenges as Meerkat, Periscope, and Houseparty as it exits beta. However, it may be better positioned for a pivot than its predecessors. By attracting interesting personalities and recording these conversations (with their permission), Clubhouse can morph into a podcast alternative with lower entry barriers for creators. In effect, this would move it downwards on the visual above. This would, of course, limit its addressable market relative to mainstream social networks, but would still give it a viable path to sustainable liquidity. The real-time component can still exist, but it would need to become a feature of an asynchronous network. In other words, it would become a content repository first and live conversation channel second. The other option for Clubhouse is to move towards the right — towards more scheduled engagement. There is some evidence that users are already taking it upon themselves to schedule conversations. However, ad hoc scheduling will be challenging to maintain as Clubhouse moves towards general availability.
To summarize, the tech community is obsessed with real-time engagement. This can be critical to building defensibility in some types of startups, like data networks, but it can be an albatross around the neck of other startups — particularly interaction networks. Within interaction networks, asynchronous networks find it much easier to reach liquidity, compared to synchronous, real-time networks because they don’t require their users to be active at the same time. For synchronous networks to become viable, they need to target a segment of users that are likely to be active at the same time. If not, they need to pivot and make real-time interaction a feature within a broader, more asynchronous value proposition.