Inside the Clubhouse

About a month ago I got a DM inviting me to “a community for casual, drop-in audio conversations.” So I downloaded the app, created a profile, and was greeted by a blank screen.

Uhhhh...

Then, a face appeared! Its name was Paul — and it spoke!

“Hey! Welcome!”

After we exchanged pleasantries, Paul explained how the app works. There’s one global “room,” and when you join you start off on mute, but anyone can unmute themselves. When you open the app, it sends push notifications to everyone on the app, so they can join you and chat if they’re free.

Strange, but delightful!

We said our goodbyes, and the app shrank into its place, like a genie sucked back into a bottle. The icon was a black and white photo of a guy with a funny grin, for some reason. Beneath it was the name:

“Clubhouse”

. . .

If your corner of Twitter is anything like mine, you’ve seen a lot of talk about Clubhouse in the past weeks.

The reactions are, like many things these days, polarized and overly meta:

Some people LOVE IT because it’s the first exciting new social app they’ve seen in years, and they miss the good old days when the internet felt fresh and full of possibility.
Other people HATE IT because it’s become known as an invite-only “elite silicon valley” thing. (To be fair, the founders want to open it to everyone as soon as possible, and anyone in the beta would testify it’s definitely not ready for scale yet.)
Some people love to let you know they’re in it, because they think it makes them look cool. (Oh shit is that me?)
And finally, some people love to let you know that they know the only reason people talk about Clubhouse is because they think it makes them look cool, because they think that makes them look cool. (Get it?)

But honestly, I’m not interested in the chatter. I’m interested in the product, and its potential as a business.

So let’s focus on that.

. . .

Social apps are just like any other product. In order to understand their probability of success, you analyze them the same way you would anything else:

What need does it fulfill?
How does it compare to the alternatives?
Why should it emerge now, of all times?
Who is paying, and why?
What’s the accumulating competitive advantage?

If those add up, you might have a valuable thing on your hands.

In this essay, I’ll quickly tackle each of these questions in order to understand why Clubhouse is doing so well now, and how far out we can reasonably extrapolate their initial traction.

Then, I’ll focus on friction in the current Clubhouse experience, and suggest a few ideas for what I would do next if I were the founders.

(And after that, as a bonus, I’ll pitch Jack Dorsey and Daniel Ek on why they should try to buy this app — but not immediately!)

. . .

Clubhouse is working because it’s halfway between a podcast and a party, and people love both of those things.

The great thing about podcasts is they allow you to learn how your heroes think and see the world. The more you listen, the more you become like them. You start to know the things they know, think the way they think, and talk the way they talk — right down to the verbal tics.

This is more important than it may seem. Research has shown that humans aren’t that much better than other intelligent creatures at first-principles reasoning, but we are absolute killers when it comes to “monkey-see-monkey-do.” (The formal term for this is “social learning.”)

“But most people aren’t interested in learning!” You might say. And it’s true, the popularity of lifestyle, gaming, comedy, and sports vastly exceeds the kind of nerdy stuff that you and I get excited about. But it’s also wrong, because if you’ve spent any time on TikTok or Twitch or YouTube, you’d know that the reason people are there is absolutely because they want to learn. It’s just that instead of learning about business strategy, they’re there to learn far more practical skills, like how to dress, and how to respond when someone pokes fun at you, and what certain cool references mean. Those skills are way more obviously useful than anything you’ll learn on nerdy newsletters like this 😂

Of course, live audio conversations aren’t the only way to socially learn. Text and video are great, too. But unscripted conversations have unique competitive advantages that are unbeatable in certain scenarios.

As is well known, audio is better than writing and video when you need your eyes for something else, like driving or cooking. (Lately I’ve been loving Clubhouse while I tend a fire for the BBQ I’ve been smoking on the weekends.) Another great thing about audio is it’s a much better medium for conversation than any other, because video of talking heads gets boring after a few minutes, and interview transcripts are often hard to follow. Audio also enables people to create without worrying about how they or their environment looks.

But why live audio? Why show a list of everyone who’s listening? Why allow listeners to become speakers? These are the things that make Clubhouse distinct from podcasts — more like a party — and they’re built into the product for several good reasons.

First, live audio enables a uniquely efficient form of social learning. When you can see a person responding to new information on the fly, it’s easier to piece together the deeper patterns that govern their thinking. When they’re in an environment that’s not perfectly controlled and edited — like an essay — it’s easier to mask shallow understanding. “Live” takes the spontaneity of a podcast and cranks it to eleven.

Second, and more importantly, on Clubhouse you can meet great people. And that is one of the deepest desires in our hearts. People need each other. But there are so many things that pull us apart. Creating an environment that makes this happen is hard, but incredibly valuable when it works. It’s fascinating to me how some Clubhouse rooms feel like a bunch of kids joking around in the back of the bus, and others feel like a lecture hall. Both can be valuable depending on the circumstance.

This is also an area where Clubhouse is subtly but hugely different from live video apps like Periscope. By default, the content is co-created with a small group. Listening to a conversation from a handful of people is so much more compelling than watching a video from one or maybe two people. It’s also the key to the “party” dynamic — meeting people.

How do you meet people on Clubhouse? You join a room where you know people, and raise your hand to speak, or are invited to. This simple mechanic makes a huge difference. It’s the core “grind” of Clubhouse. And it solves one of the biggest problems with podcasting:

To explain how valuable it can be to break through the metaphorical wall between speaker and listener, let me tell you a story from my own experience using Clubhouse.

One afternoon I got a push notification saying that Paul was talking to Naval Ravikant. If this was a podcast, I would think “hey that’s interesting, maybe I’ll listen later.” I would probably read the episode title and show notes, and I would consider whether the topics were interesting to me, and maybe I’d listen, or maybe I would save it for later.

But with Clubhouse, the conversation is live, and there are no recordings. And when I join they might notice me. So when I got that push notification, I knew: it’s now or never. I had no clue what they were talking about. I was either in or out.

So, when I got that notification that Paul and Naval were talking, I had a few minutes, and I decided to join. As soon as I did, Paul recognized me and said something like, “Hey Nathan! I saw you wrote a post about Roam the other day, we were just talking about that. We’d love to get your take on it.” And then he invited me to speak.

All of a sudden, I’m explaining my dinky blog post to a guy with almost a million twitter followers!

Beat that, podcasts!

Of course, this exact moment was only possible because Clubhouse is a tiny beta community, I was incredibly lucky to get invited, and Paul Davison was able to convince Naval to join. This is what most people focus on when they express skepticism of Clubhouse.

Some people think it only works because it’s small and exclusive, or that it’s only interesting to tech people, but based on my experience I don’t think either of those things are true. I bet it’ll scale better than the bears think.

First, I don’t think the product will lose its appeal once they let everyone in. If the only value were a velvet rope, you’d still see people bragging that they have access, but you wouldn’t see them opening the app several times a day. The bottom line is, this is a compelling product to use with friends and friends-of-friends. And it’s also compelling even when you’re just listening! Sure, when they open up you’ll hear less stories about people getting retweeted by Justin Beiber, but you’re still going to have a good time.

Second, this isn’t just for tech people. Conversation is universal and people can talk about anything here. Sure, it’s dominated by tech folks now but that’s already starting to change. I can imagine a world where every subculture has several rooms going at any given moment.

I think it’ll feel similar to Twitter, in a lot of ways. Sometimes a famous person retweets you or replies to you. More often, a “famous in your world” person does. You make new friends organically and discover new peers. It becomes a way to learn new things and meet new people. It’s not extreme, but it basically works.

And honestly, even just that is enough to change lives. Twitter changed mine, and I was user number 17,424,947.

. . .

Of course, Clubhouse isn’t the first social audio app. So what makes this work where others have failed?

Thinking this question through was a kind of mind blowing experience for me. Sure, a big part of it has to do with creating a “cool kids club” vibe. But the founders of other social audio apps (Anchor, Bumpers, TTYL, Talk Show) had roughly equal access to Silicon Valley luminaries. In fact, Talk Show is a different app created by the Clubhouse founders, mere months ago!

So the early user community is not the most important thing. The bottom line is always the product. And some product choices that seem like tiny tweaks can turn out to make an enormous difference.

So let’s quickly look at each comparison app and analyze their structures to see what explains the difference in result:

Anchor and Bumpers both started in New York around 2015 with a very similar idea: “twitter for audio.” You’d sign up, listen through a feed of short audio recordings from people you follow, and post your own short snippets or reply to others. There was definitely something compelling about the experience, and they each earned fairly large followings, but both inevitably pivoted. Anchor became a podcast creation tool, and was acquired by Spotify in 2018. Bumpers became a video creation tool, and eventually was acqui-hired (I think?) by Coinbase.

What went wrong? The feed. Short audio clips generated by users just weren’t compelling enough to dislodge podcasts. On Twitter, it’s fine if 1 in 10 tweets are interesting, because you can quickly scroll and skim. But audio doesn’t work that way. Even if 1 in 3 clips is a hit, it still feels like a bad experience. Both Anchor and Bumpers had a way to reply to clips with audio, which enabled a sort of asynchronous audio “conversation” to happen, and it felt really cool when it worked! But if you didn’t like the feed experience you’d never get to the interactive “conversation” part.

TTYL took a totally different approach, and like Clubhouse is/was focused on spontaneous audio conversations, but only amongst friends. There’s no global view of conversations you can join, and no speaker/listener distinction. This created an extreme liquidity challenge. In order for the app to be useful, you had to personally know maybe 50 other people that are on it. Because when you open it, it sends a push notification to your friends, and what percent of them can join at any random moment? Maybe 5%. So if you have 50 friends, maybe 2 or 3 can join you, and it’ll be fun. But it’s incredibly hard to get 50 of your friends to join. Also, because these are your existing friends, you can always just talk to them other ways. So the “private / friend” focus ended up limiting TTYL’s growth.

Talk Show is/was a way to livestream podcasts. You could open the app, send an invite to a friend to record a talk show, and they could join you on the app or over the phone. Then, once you start recording, you can tweet the link and people can listen as you talk. The idea was to reduce the friction to recording conversations, but it was just a marginal improvement over regular podcasting for creators, and not different or better in any substantial way for listeners.

Out of Talk Show, Clubhouse was born. You don’t have to schedule a conversation or reach out to anyone and ask if they want to chat, you just open the app and start talking. Anyone can join to speak or listen. In this way, it nails the interactive part, like TTYL. But because the conversations are public, it solves the liquidity problem that slowed their growth. Being able to join as a listener isn’t just good for liquidity, it’s also less demanding: you can open the app just to listen, when you’re not ready to speak. And it solves the “feed” problem that plagued Anchor and Bumpers, because listening to interesting people have a conversation for 40 minutes is a compelling and familiar experience in a way that a stream of short snippets is not.

Based on my experience using the product over the past few weeks, I can say it’s displaced a lot of time I spent listening to podcasts before, and works in a way that none of the above products did for me. If I were the team, the number one metric I would look at is retention. This is what separates a genuine banger from a flash in the pan.

I haven’t seen their metrics, but from what I can gather by looking at the app every day and noticing who is hanging out, it seems like retention is pretty good.

If they can keep it up, Clubhouse will become an important company very quickly.

INTERMISSION

Ok, we’ve covered a lot! Feel free to take a break. Stretch, breathe, get some water, etc. Then come back.

There are only two sections left in this essay: what I’d do if I was running Clubhouse, and why I think Twitter or Spotify should buy it (but not immediately, if possible).

You good? Ok! Let’s dive back in 😅

The main strategic goal of Clubhouse is to make conversation liquidity even better.

What does that mean? In an ideal world, you’d be able to join at any moment you like, and enter a room of the coolest people you can imagine, who are talking about exactly what you’re interested in, and when you join they notice you and invite you to speak, and you say something amazing that makes everybody spontaneously and unanimously burst into cheers of “hip hip, hooray!” and “for [she/they/he]’s a jolly good fellow!”

Of course, scarcity is a fact of life, and liquidity will never be perfect. But we can get closer by removing the obstacles that prevent the above scenario from happening.

One obstacle is choosing the best room to join. If there are twenty, how do I know which is most likely to yield the ideal outcome?

Right now all I have to go by is a list of names, with a “💬” emoji next to speakers:

This is pretty good. I know the size of the room, and I know who is speaking and who is listening.

But if I don’t know the people by name, I’m out of luck. And this is a pretty big problem. For example, not many people know Chris Lambert by name, but he is the CTO of Lyft. He was chatting with Mike Krieger (the co-founder of Instagram, another name many people may not recognize) and it was awesome, but I bet a lot of Clubhouse users missed it because they didn’t realize who these people were.

I think it’s worth it to pack in more signals here that will help people choose the right room, even if they’re not familiar with all the speakers ahead of time.

So here’s a stab at a redesign:

By consolidating information about the listeners, and expanding the information about the speakers, I think they could make a lot of rooms more enticing to join. When I tap a room to join, I’d love to see this information there for everyone — speakers and listeners.

One thing that’s notably absent here as in the original design: names for rooms. I don’t think they should add them. Conversations on Clubhouse meander, and that’s a beautiful thing. Any name that tries to describe what’s being talked about will go out of date quickly unless someone manages it, and no one wants to do that. One thing that might be kind of interesting is a real-time transcript of what’s being said, like closed captioning. But the cost to run that process for hundreds of rooms simultaneously would likely get into “ridiculous” territory pretty quickly.

But this is just a marginal redesign. What are some bigger changes Clubhouse could make to improve liquidity, and to increase the odds of an ideal experience?

Here are a few ideas:

Sometimes the conversation hits a lull. What if the speakers could press a button and the app would play a “conversation prompt” that gives people something to talk about? This was an idea Scott Belsky shared in one room, and called it “kindling.”
Conversation prompts are one thing, but what about other kinds of social games? With a live group of people, there’s a lot that can be done. Clubhouse could create a SDK and open up a marketplace for anyone to develop social audio games for people to play. I can imagine trivia, cards against humanity, and werewolf/mafia all working really well.
In any given room there are a handful of speakers, and sometimes dozens or even hundreds of listeners. Some percentage of listeners would like to speak. How do the speakers choose the best people to bring up? One way is to allow people to type in some short text when they raise their hand, in order to let the speakers know what they want to say. These could even be displayed publicly and voted on by the listeners. It also would help a ton to have social graph information a bit more handy, so you can see who you know that follows this person or what their bio is.
I’ve found myself sending backchannel DMs and text messages to friends that I see are listening to the same Clubhouse conversation with me. This is an obvious thing to build into the product. One idea someone shared in a chat was to even enable side audio conversations, where the volume of the speakers go down and you and your friend can chat on the side, privately. It’s like sitting in the audience of a panel discussion and whispering to your friend sitting next to you.

Of course, it’s easy to rattle off a list of ideas. But the most important thing is not to over-complicate the product.

I bet Clubhouse can sustain 20% week-over-week growth for awhile once they launch. That sounds like a lot but I think they have a good chance to do even better than that if they play their cards right. The core mechanic works, the early community is fantastic, and they can ride those two things a long way as long as they keep the experience simple.

Once it’s clear that product-market-fit is there, and Clubhouse has tens of thousands of daily active users, I’d start to think about monetization. This is much sooner than most consumer social apps would have thought about it in the past, but I think in this case it’s a good idea. The default assumption in the past was that you reach massive scale and then slap ads in there. That’s much harder in a world dominated by Facebook and Google, and people are much more sensitive about privacy now.

For these reasons, Paul (the Clubhouse CEO) has said that he would rather focus on monetizing through creators and users on the app. That makes me incredibly excited about their potential as a business. There’s a ton of stuff they can do: tipping, a subscription fee for advanced creators, Fortnite-style cosmetic upgrades for rooms and users, social games that require someone to pay to unlock them — the list goes on.

The thing is, these monetization ideas have to flow somewhat naturally from the core user experience. It’s a little harder to just “tack it on” in the way you’d typically do with ads. So that’s why I think it makes sense for Clubhouse to focus on it sooner (in the 10-100k DAU range) rather than later.

And there’s one other thing revenue buys you: negotiating leverage with potential acquirers.

. . .

As I’ve mentioned several times already, if I were Twitter or Spotify, I’d try to buy Clubhouse. But ideally I would wait to do it until a little later.

Periscope and Vine suffered from being acquired too soon, I think. Had they had a chance to develop more as independent platforms, the audience and product might have been stronger. I’m sure the creators of those apps still worked hard after acquisition, but it can’t be the same. I think there’s a reason Instagram and Snapchat are as strong as they are today. If you buy a company too soon you might accidentally destroy a lot of value.

But at the right time? This company could be magic for both Twitter and Spotify.

Twitter is obvious. They already bought a livestream video company, why not an audio one? Their core business is to maximize time spent in the app so they can display more ads, and based on my usage of Clubhouse over the past month, it will be a great net-add. It couldn’t fit more perfectly into Twitter’s mission, “to give everyone the power to create and share ideas and information instantly without barriers.”

Also, Clubhouse might be better if it were based on your Twitter profiles and follower graph. But I can argue the other side of this question, too. It takes work to maintain a follower graph, so a fresh start may not be a bad thing for people who feel like their Twitter timeline is getting stale.

Spotify is not quite as obvious, but is still pretty clear. Their strategy to avoid paying all of their revenue to the record labels is to get people consuming other kinds of audio content in their app, namely podcasts. I have spent a lot of time on Clubhouse instead of listening to podcasts in the past four weeks. Therefore acquiring Clubhouse is not only an offensive move, it’s a defensive one as well.

But the best part about it for Spotify is that it would solve their biggest problem with podcasts: they’re available everywhere. For Spotify, Clubhouse would represent a unique form of content you couldn’t get anywhere else. They could also make Clubhouse a lot better by enabling background music in rooms.

Of course, it’s ridiculously early to predict outcomes like this for Clubhouse. By my best guess, they have less than a thousand people in the app.

But it’s fun to think about :)