The second product I worked on in 2019 was Knok. Knok grew out of believing that how we use our phones to talk to the most important people in our lives is deeply flawed.
It all started when my co-founder and I walked through Midtown to get something to eat. He had a product idea: “We should build a social media network for your own family,” he said. What a stupid idea, I thought. One week later, this product idea made some ex-Facebook engineers very rich. They built Cocoon and raised quite a bit of VC money for it. It was exactly what my friend had proposed: a private social media network just for you and your closest people.
With this newfound belief in our ability to predict the future and generate ideas that reflect the spirit of the times, we built Knok, a text-based messenger that fixes everything wrong with texting.
The idea was simple: Your best friends shouldn’t live in an iMessage inbox with UPS delivery notices and ConEd invoice texts. They should feel as important as those connections are to you personally, giving you a more premium feel when talking to them.¹ In the same breath, we also wanted to fix a problem with the way texting works. It’s asynchronous, which is great for sending messages but not so much for receiving them. Say you’re out in the real world, having fun, and once you get back to your phone, there are a million group chats to catch up on and texts to respond to. Texting is becoming more like email.
Since email isn’t fun, we decided to flip the script entirely. Knok borrowed from some of the earliest messengers like AIM and ICQ and only let you text people who were currently online. “Online?” you say, “what does that mean?” Well, we figured out a way to get access to a level of the iOS system that I didn’t think was possible to tell when someone is on their phone. Not just while in the Knok app but also when doing other things on their phone. They are online while texting someone on Knok, looking at the weather, or scrolling through Instagram.
If you’re on your phone, people can text you. If not, they can knock and send you a temporary 10-second notification. A chat is opened if you see and respond to the notification in time. If not, it disappears and doesn’t give you the feeling of missing something important. We completely reversed the role of notifications. We make it harder to reach you and don’t permit anyone to take up your time and brain capacity at any time of their choosing.
Let’s look at a more detailed example. When Alice knocks on Bob, and Bob doesn’t respond in 10 seconds, Bob has no lingering notifications. On the contrary, Alice gets a notification whenever Bob is online the next time. Since Alice immediately knows when Bob is ready to talk, a chat can still happen quickly. Most importantly, this system makes synchronous, fast-paced chats more likely. Those kinds of moments where you see someone typing and texts fly back and forth. We worked hard to make those moments frequent when we built Knok.
Knok was full of UI and UX that made the messenger feel premium. We wanted to give you an exceptional experience for all the most important people in your life. To facilitate this exclusivity, you could only add five people to your list of chats in Knok—your top five. The idea was that Knok is like a dinner party in a Manhattan apartment—fun, cozy, and a little exclusive.
For technology learnings, there are two big ones. First, don’t build end-to-end encryption into a messenger unless someone pays you to add it. It’s a feature that’s much more difficult to build right than you think, and even if implemented well, the tradeoffs are often not worth it. Reading a story about someone losing their cryptocurrency private keys will give you a million-dollar reason not to trust users with storing their keys.
The second learning is more optimistic: Sometimes, you can hack a very closed-off and constrained system to gain a level of access you never thought possible. I would’ve bet a lot of money against the possibility of building the online detection that made up Knok. I won’t go into too much detail here in case the trick ever comes in handy again. Rest assured: It was fully Apple-approved, and even if they knew exactly how we did it, I’d be surprised if they removed the functionality we used.
The more complicated learnings from Knok aren’t tech-based, though. While I believe in all the principles that led us to build this product, it was a failure. Not more than ten people ever used the product, and the two-week retention was a resounding 0%. It’s challenging to make one definitive statement about the “why.” But maybe it’s this: The messenger wars might be over, and there’s a chance everyone today already uses all the messengers they will ever use.
You could launch at one highly concentrated place, like a high school, and become the new hot thing there before jumping to more places. Maybe you need to find a thing that you can do better than any other messenger and do so for less money, but since most messengers are free, that’s a tough sell.
The moat for any messenger is deep. Switching messengers is a nuisance; as a general rule of thumb, it just doesn’t happen. Ever. I haven’t heard anything about Cocoon since 2019, and I wouldn’t be surprised if they shut down the same year they started. Social media is complex, and apart from a few exceptions, no successful social company launched in the last ten years. All the ones that did succeed did eventually shut down. Even BeReal, which marks the most recent exception to the rule, doesn’t look like it will stick around.
Social products need to do more than just messaging if launched post-2014. Messaging is such a mature space that you can’t find a niche to penetrate. There isn’t a segmentation of the user space when it comes to messaging. 60-year-old grandparents use the same products as 14-year-old teenage girls. It doesn’t matter what profession, interest, or taste you have. Building a new product is always easier if you achieve high penetration of a small group as opposed to low penetration of a big group. And unfortunately, messaging is the poster child example of that rule.
I might be too pessimistic here. I could be wrong. There may be a new flashy messenger coming around soon. Whether it will happen or not—my experience building Knok and the Matrix Messenger in 2023 led me to a strong believe.
Summary
Even the best conceivable messenger product is not a venture I want to build—not until finding new insights into the adaptation and segmentation of the messenger market. Making anyone use a new messenger is very tough. Messaging is an unrewarding space. All social apps have this problem, but text-based messaging, more than any other social interaction, is not a product I want to build. I have spent plenty of time gathering evidence that it’s a game I can’t win. Knok in 2019 and the Matrix messenger I built in 2023 were very convincing case studies. I’ll write more about that second app soon.
¹: I vividly remember a landing page about a year ago featuring this exact proposition: cleaning up your inbox to highlight more critical messages. I can’t remember the name of that product, but if you know what I’m talking about, reach out, and I’ll add it here.