The one I never really quite understood was Slack. Who builds an internal chat client, while building an online game, and then decides to build a b2b business around that chat client?
The rest at least have some logic to them. You build something, people use it for something else, so you generalize. Or you pursue a neighboring market or use-case. But the Slack one just seems so random.
Early people at Slack were nerds and unhappy with the communication tools that were available at the time.
So they went the nerd way and hacked together something they would not spend most of the time complaining about. "Hacked together" is important. It was not dedication, it was "scratching your own itch ".
The hack was satisfactorily working. For them.
And when Glitch failed, they were left with (among other things) their communication tool. "Someone" thought it was worth trying to market it, following the proverbial dogfooding strategy, because there was nothing else left to do anyway with Glitch.
And boom. Slack.
Source: memories of an article read many years ago that I can't find traces of, but was quite fascinating.
I've heard similar stories many times, where someone discovers a gap in the market through an actual need that they have while trying to do something else.
Like the Kitopi Cloud Kitchen. The story I heard is that the guy already had a successful sweet business, and got the idea to have little mini distribution kitchens for it, instead of opening a new business everytime he wanted to extend his reach. Then decided to make that idea a whole business model.
Wow, I didn't know Kitopi Cloud Kitchen. This is truly a disruptive idea (I never use this word), as it breaks with the traditional idea that restaurant food cannot be industrialized and has a connection with the chef but also the place.
Even after reading this, it is hard for me to think "I could order food from this restaurant but the actual meal will come from a partner kitchen".
Brilliant from a business perspective, yet somewhat questionable culturally.
One cool part about it is that because the portions are so industrially controlled there a whole slew of reliably calorie counted restaurants available.
I believe the story is: their game failed, and they went back to Sequoia to return $5M in outstanding investment. Sequoia told them to 'keep it, and build something'. So they tried the Slack thing.
What's funny is that he's the same dude who founded Flickr, which also started out as some kind of game. Ha ha.
Also, $5M is a small amount for Sequoia - and it would have been a hassle to deal with, and frankly, probably already written off.
If you have a good founder and a team already moving ... $5M is actually enough to get a small team moving in a new direction so it's a smart choice.
It's a smart choice in modern Valley language though, instinctively, most other investors would have not done that.
While we can give credit to the investors, that only works in a system of plentiful bounty. You need to have large acquirers, tons of talent, etc. etc. in order for that investment logic to make sense.
If I recall correctly, Slack was built utilizing IRC on the backend.
I remember that resonating with me as I was sole developer on an in-house invoicing system at a Fortune 500, and I added a whole corporate chat functionality using IRC libs. It only took a few days to implement. It made me aware how easy it can be to build a billion dollar business by mistake. Too bad not my billion dollar business!
A similar case for a company I previously worked for. They started out making high quality photo albums, I believe targeted primarily at pets. They built in the deep-linking experience, so that you could send links to friends, etc. They eventually realized that every app ends up building in their own linking, and in turn became branch.io
Segment is similarly weird. They were building EdTech and then decided that this boilerplate file called analytics.js was the coolest thing they had built and they pivoted the company around that.
I don't know Segment, but EdTech is all about analytics. Everyone wants hyperspecific data, e.g. do students who attend the first three lectures do better than students who didn't do that, but were there the last lecture?
So not so surprising, "from a certain point of view" [0].
Ok, good point. I think their analytics.js was just tracking website usage, but you're probably right that they were planning on deep-inspecting that data in the way you suggest.
The same people who pivot an MMORPG's screenshot-sharing feature into the top photo site of its era, I guess! Butterfield pulling that off twice just completely floors me.
There's a fantastic episode of Reid Hoffman's "Masters of Scale" podcast[1] that goes into great detail with one of the founders, Stewart Butterfield. He also did basically the same thing to form Flickr.
It's not that surprising when you know the history. This is the founders' second or third try at live chat with media.
At Ludicorp, in the early oughts, they were building a game called Game Neverending. They built a chat feature into that game. Then they added the ability to drop photos into chat. Digital cameras and cameraphones had just become affordable, so suddenly that was the main feature of the game. Flickr had to drop the single live chat window when they became too popular, and then it became a web-based photo sharing community. But they always had plans to bring it back, they just never got around to it. Once Flickr was acquired by Yahoo those plans became even more difficult to realize.
Like many companies in the middle-oughts, Flickr did everything over IRC. When they were acquired by Yahoo, most of them moved to San Francisco, but some employees never left Canada. So it was a distributed, remote workplace the entire time. It was natural to do everything over IRC and add bots and such to help you do things.
When the Flickr founders left Yahoo, they founded Glitch, and it was also quite distributed, half in Vancouver, BC and half in the Bay Area. I'm not sure how they came to build their own sharing-media-in-IRC solution again, but they had the tools to hand.
A fun aside: as Glitch was failing, but before they pivoted to Slack, they downsized. And a lot of those downsized employees reformed as "Tomfoolery" and created a product called "Anchor", which was basically Slack! There aren't a lot of traces that this thing ever existed but here's an article from Fast Company:
I assume that those employees realized that their internal tools were actually the best thing that they had made. Anchor was led by a former Yahoo executive who had I think been COO at Glitch. Tomfoolery/Anchor didn't get much traction and was acquihired by Yahoo just a few months later - most of those employees were ex-Yahoo anyway.
A few months later Glitch pivoted to Slack and the rest is history.
It's unclear to me why Tomfoolery failed when they had all the knowledge about how Glitch's Slack worked and a head start of many months. I remember a period in 2013 when a group I was involved in was choosing between Flowdock (yet another thing that was basically Slack) and Anchor and the recently-launched Slack. Anchor didn't have the rich integrations of Flowdock. Slack was very new and immature and was worse than both of them. But Slack improved faster and people like Stewart Butterfield had way more goodwill.
The rest at least have some logic to them. You build something, people use it for something else, so you generalize. Or you pursue a neighboring market or use-case. But the Slack one just seems so random.