I first wrote this essay a few years ago. A founder mentioned it to me over the weekend, and so I decided to re-publish it here. One thing that's bothered me in the time since I wrote it is the way in which toys, which are meant to be harmless and fun can become companies capable of wreaking havoc on the world. I mention Facebook below, which was quite fun when it first showed up on campus. I don't personally find Facebook fun now, and think it has caused a lot of damage in the world. Thinking through the way in which you build things, even toys, to create positive forces in the universe is hard, and maybe more important for that challenge.
Some of the biggest technology companies look like toys in the beginning. From a classical business building perspective, this shouldn't happen. Toys are for fun. Businesses, especially huge ones, are for making money. Toys are small and of limited use. Large companies contain multitudes and perform a huge array of functions.
This trend does not fit with history either. Standard Oil, US Steel, and Boeing were all iconically huge companies that were built as businesses. None of them went through a phase where they looked like toys. Startups can be different, though, because of the expectations of them and the seriousness with which people approach them.
If you give people a tool and tell them it will perfectly solve an important problem, any imperfection in the tool is going to make them angry. If you give someone a toy and say “Look what I made! Isn't it fun? It kinda does this thing.” then you've set yourself up for a positive reaction. It's much easier to beat low expectations than high ones, so you've materially increased your chances at having a happy user.
And “happiness” is an important way to think about early users. People spend more time with something that makes them happy, especially when they don't expect it. Happy users are easy to get feedback from, because they know that you can make the product better and make them happier. They're also likely to tell friends about the cool new product that they're using, which means you start to get users without having to dip into the dark arts of marketing.
When you look at something you build from the perspective of how happy it might make someone vs. how angry it could make them, it also becomes easier to experiment and put things into the wild. This isn’t just about low stakes, it’s about how seriously you take what you’re doing and how seriously other people take it, at least at first.
Business is about making money and working with customers. These are very serious and scary things. Toys are for playing and trying new things. This isn’t serious at all.
Maybe that seems bad if your goal is to make your toy into a startup, and your startup into a big company. That implies you have to be serious right from the start. But if you are serious right from the start, a number of things start to go wrong.
The first thing that goes wrong is you become unwilling to experiment with ideas that aren’t clearly aligned with making a big company. This means that people building serious things focus rapidly on revenue. They become risk averse and innovation averse. Companies built on new technologies have to capitalize on non-obvious ideas, ones that wouldn’t pass muster in large corporations. Otherwise, the large existing companies would do these things themselves.
Facebook is a great example of this. Early on, all users could do was look up people they’d met at parties on campus at Harvard from other dorms and poke them. This seems silly because it was. Very few people saw it as more than a toy, which is why they were willing to give it time. It was something to play with when not working. I don’t think we would have been willing to play with something that felt like a serious business, which would have meant that Facebook wouldn’t have gotten it’s early happy and engaged users.
The second thing that goes wrong when you take your toy too seriously is that you signal to the bigger and better funded companies already in the marketplace that you are onto something important and profitable. This is bad, because those companies will start paying attention to your toy too early and copy/buy/kill it. Airbnb looked like a doofy hipster thing to hotels for a very long time. And then, when it was too late, they realized that it wasn’t a toy at all. By that time, Airbnb had enough customers, revenue, and funding to survive the attacks of the incumbents.
The third thing that goes wrong when you take your toy too seriously is that you immediately start optimizing on the things that you believe serious businesses should - profit and margins. While these things are important in the long run, focusing on them too early injects an impossible set of things for an early startup to do.
Startups only have so much time and ability to focus. At the earliest stages, that focus needs to be on making things that users love and want to play with. A startup’s early and heavily engaged users are its only real base of strength and chance for growth. Focusing on anything else puts them at an immediate disadvantage to better funded, organized, and wide reaching companies.
When toys become companies
Not all big companies start out as toys, just as not all toys eventually become big companies.
This is just as often a question of motivation and goal for the creator as it is a question of whether or not the toy was good or bad. Most people who make things don't want those things to become companies, which is great - it would be unfortunate if every interesting thing made for the world had a commercial purpose behind it.
The founders who do turn toys into companies are generally the ones who relentlessly push what they've made to users and obsessively improve the toy in response to feedback. They then have to master a whole set of skills that are orthogonal to making things - hiring, managing, business building, fundraising, etc.
This is an atypical path, which is why it's so exciting when it happens. When we meet founders who we think are going to combine the toys they've built with the ability to build something lasting, we generally fund them whether or not we're sure the toy is actually a business.
 Two examples: Facebook was a way for people to waste time while Apple helped hackers build home computers before there was any business case for it.