One interesting thing I've learned from studying the history of several dozen startups and from doing about 20 pivots myself is that oftentimes the startup doesn't pivot into the market, the market (or technology) pivots into the startup. In other words, the startup is non-viable at the time of its founding, but later events change the structure of the market in a way that takes the struggling business and turns it into a growth powerhouse.
Google, Whatsapp, Uber, Twitch, AirBnB, EverNote, Roblox, and to some extent Apple can all be seen as examples of this.
This has a bunch of implications for what it's really like to found a startup: in particular, it validates the wisdom of keeping burn rate extremely low and of being extremely persistent, as well as having a product out there that people can use. It also suggests that it may be better to have a lot of usable-but-unpolished working products that are constantly getting exposure (even if they're not succeeding), rather than putting all your eggs into improving one product, and argues for putting failed startups in maintenance-mode rather than shuttering them entirely.
This is really insightful. It aligns nicely with the point Peter Thiel makes that the key to being a successful startup is believing something that is true that others do not. Usually the thing you believe that is true is about something that is going to change, but you probably don't know exactly when.
So I'd argue your point extends his thesis that the key to success is "believe something that is true that others do not, and survive until it manifests into an opportunity."
I think Apple, Google, Twitch & Uber (I don't know enough of the others) were clearly building the right thing for an early market.. I wouldn't say the market "switched to them", it just became more mainstream.
That's one way of phrasing it, I guess, although I'm not sure it's a particularly useful one (what makes a market "early" if not the mainstream switching later? how would you distinguish it from the many other small niche markets - like bingo card creators or D&D character sheet generation - that fail to become large growth companies?). The specifics I had in mind for those were:
When the Apple I came out, there wasn't much you could do with it. It was only the invention of Visicalc (which itself couldn't have happened until the Apple was out) that gave businesses a reason to buy one. The same dynamic applies to nearly all "platform" companies, which is why it's so hard to make a new platform.
When Google was first developed in 1995, the dominant way to find things on the web was Yahoo. That's because in 1995, the best way to find things on the web was a human-curated directory. The press reported Google's dominance as because of its proprietary algorithms, but what makes algorithms better than humans? I'd argue that there were two main forcing functions, both of which were a result of the web's scale continuing to grow: 1) as the web gets better, you physically cannot hire enough people to review and categorize it and 2) as the number of links gets higher, the effectiveness of counting links as a ranking factor improves. PageRank and PageRank-like algorithms are actually really ineffective on sparse datasets (trust me, I've tried, while at Google). So it was the growth of the web itself that led to algorithmic search engines becoming better than human-curated directories.
Game-streaming wasn't really a thing when Justin.TV came out in 2007. It required a number of other changes in the environment - better access to broadband, faster GPUs, a gaming populace that was thrown out of work by the Great Recession - before it got big, circa 2010/2011. Justin.TV, as a lifestreaming company, was well positioned to realize and capitalize on this trend, even though it didn't exist when they founded the company.
Uber started as a way to call a black cab via app. They tried it multiple times between the company's first prototype (in 2008) and the official product launch (in 2010). Mobile GPS wasn't good enough in 2008 to be on constantly: it would drain the phone's battery faster than the power adaptor could charge it. Additionally, they benefitted a lot from the Great Recession, which meant that a number of drivers were newly unemployed and seeking a way to make easy money. (The same applies to the rest of the sharing economy.)
> how would you distinguish it from the many other small niche markets - like bingo card creators or D&D character sheet generation - that fail to become large growth companies?)
The definition of 'startup' I like best is "a new business that's designed to arbitrage a temporary disequilibrium to produce extraordinary returns."
With bingo card creator or D&D character sheets, there is no temporary economic disequilibrium.
Whereas with Apple, the arbitrage opportunity is that there were personal computer chips but no personal computers. With Facebook it was that there were a billion people who owned computers, but there wasn't anything connecting the people behind the computers. It wasn't obvious that Apple and Facebook would be successful in advance, but they were clearly hitching a ride on mega trends that were already obvious ahead of time. When there is an obvious mega trend, but no dominant company for that mega trend, then that's basically an economic low pressure area that needs to get equilibrated.
With bingo card creator or d&d character sheets, nothing had changed to create any sort of arbitrage opportunity. Maybe there is a demand for those products, maybe there isn't, but a new product with demand does not a startup make.
Both of Stewart Butterfield's companies, Flickr then Slack, are spin-offs of tools the team made while trying to make a game. Maybe that's what your coworker was thinking of?
Google, Whatsapp, Uber, Twitch, AirBnB, EverNote, Roblox, and to some extent Apple can all be seen as examples of this.
This has a bunch of implications for what it's really like to found a startup: in particular, it validates the wisdom of keeping burn rate extremely low and of being extremely persistent, as well as having a product out there that people can use. It also suggests that it may be better to have a lot of usable-but-unpolished working products that are constantly getting exposure (even if they're not succeeding), rather than putting all your eggs into improving one product, and argues for putting failed startups in maintenance-mode rather than shuttering them entirely.