I spoke on a tech careers panel last night in Oxford and one of the questions was whether students should start startups when they graduate. More and more people wonder this. For most of them, the quick (and true) answer is no. But I think there are some people who should try and that for those people deciding to start something is one of the most valuable decisions they can make.

It’s a very bad idea to try to convince people to start startups. They’re treacherously seductive enough in their own right and anyone who needs convincing is probably not going to have a great experience. This seems particularly true if you grew up in any sort of physical or cultural proximity to Silicon Valley: if that’s the case, you certainly know that startups are an option, perhaps even an attractive one; by the time you reach the point of deciding what you want to do with your life, it’s quite unlikely that you’ve not at least implicitly considered whether you might want to start one.

Europe is not like that. One of the most important ways in which Europe is not like Silicon Valley is that for smart, ambitious young people in Europe, starting or even working in a tech company is not close to being the most desired option, let alone the default. In fact, it’s quite possible that you could be an exceptional engineer or computer scientist in Europe, reaching the end of your degree or your first job, and never have considered that you could start a company.

I wrote this – and, indeed, we started EF – for those people. I’m assuming that if you’re reading this, you have a lot of options – perhaps you’re considering (or already are) working in a bank or hedge fund, in a big technology company or in professional services. These are very much the European “default” options for ambitious people – and for most people they may be good options. My goal is not to try to convince you to start a startup, but to walk through what kind of person starting a startup might be a good idea for. To keep it as simple as possible, it boils down to two questions:

Do you have a good reason to want to do it?
Is it worth trying (for you)?

There are lots of bad reasons to start a startup, particularly any that remotely conceive as startups as a get-rich-quick scheme. Startups are hugely difficult, stressful and time-consuming. They are harder than you expect, even when you take that into account. Worse, the median financial outcome is zero, even though lots of the people who start companies are smart and hard working.

The best reason to try to start a startup is because there is something you want to exist and you think you can build that thing. The startup-as-a-company is almost incidental; it’s something that emerges out of having built a product that you wanted to exist.

Our experience at EF over the last three years suggests that it’s not actually very important to have a particularly clear or refined idea about what that thing might be. You certainly don’t need a blueprint or a “perfect idea”.[1] The best thing to have is an obsession – either with a particular (often quite niche, often quite unbusiness-like) area or with a particular technology. It’s amazing how effectively smart and determined people can turn obsessions into startups.

Perhaps that’s too abstract. A good, concrete test is to ask yourself, “What would I do with (say) six months of time and space which I could use to build anything I wanted?” If you can’t think of anything at all, you shouldn’t start a startup. But if there’s something that comes to mind – particularly if the thought fills you with excitement – then even if it’s more an area you’d want to explore than a product you’d want to build, you probably pass this test. We have many very promising startups at EF that started exactly this way and have gone on attract lots of customers [1b] and lots of investment.

Is it worth trying (for you)?

If you believe you are no more likely than average to be successful, starting a startup is a very bad idea. As I’ve said, the median outcome is failure. Fortunately, though, as Peter Thiel has memorably phrased it, “You are not a lottery ticket”. Everyone’s chance of succeeding is emphatically not equal and, for exceptional people, the expected value is probably very high indeed – not just financially but professionally and intellectually.[2]

The interesting thing is that there appear to be no really reliable predictors of who will make an exceptional founder. Obviously, most successful founders are smart, determined and industrious – but so are many unsuccessful founders. The reason is that your abilities as a founder are not only unknown before you start a startup, but they are unknowable. One metaphor might be that your abilities as a founder are to some extent procedurally generated by your experiences once you get started. When I look at the 120 or so founders we have worked with at EF, it is striking how little their ability, experience and skills at the start of the programme predict their success. I’m not saying that those things are not important, but they’re just a foundation — necessary but not sufficient. The really good founders discover both that they are really good and how to be really good as a consequence of getting started. Working out how good a founder you are is a discovery process, not an analytical one.

Given this, there’s no way to find out if you have what it takes except to try. It’s like a lottery in which some tickets are orders of magnitude more likely to win, but the only way to find out whether you have one of those is to play.[3]

Assuming that you think there’s some chance you could be a good founder (and — to repeat — if you don’t, you definitely shouldn’t try!), I think you should see the decision to start a startup as much more like buying an option, in the financial instrument sense, than buying a lottery ticket. You’re buying the opportunity to find out whether you could build something big. The potential upside is enormous, but you can’t lose more than you pay for the option.

An important question, then, is how much the option costs – that is, what you give up when you decide to try to start a startup. The two kinds of cost most people worry about are “career cost” – by which I mean the opportunity cost of giving up a great job to do it – and financial cost. The answer is that both these costs are now much lower than you might think.

I actually think the “career cost” is close to zero. I can’t imagine any job that you’d be qualified to do at time t, but unqualified to do at t + 7 months (which is how long you have, funded, at EF to find out if you’re onto something). The evidence at EF is that people who haven’t succeeded after seven months easily transition into jobs at startups, bigger tech companies and venture capital funds. The risk – if you can call it that – is that starting a startup may change your preferences: almost everyone who fails at EF wants to join a startup as their next step, whereas I think before they joined us they might have thought they’d join a bank or a consultancy.

The financial cost is also low. This is true generally – the costs of starting up have fallen dramatically because of the rise of AWS and open source software – but particularly at EF, where we literally fund you to discover whether you have what it takes. It’s certainly not enough to fund a champagne lifestyle, but it’s enough to give you time and space to build something and learn whether you could be a great founder.[4]

You are an underpriced option

If I’m right, this option is wildly underpriced for people towards the start of their careers who think they might have what it takes to succeed. For relatively little time and money, you can gain some extremely valuable information: whether you could build something big. In general, options are good and cheap options are even better. I’d go so far as to say that if I had to distil life advice for a student down to three words it would be, “harvest cheap options”.

A little while ago, the EF team was trying to come up with taglines we could use in our promotional material. One (which we eventually rejected as too brutal) was, “Don’t worry: Google will still be there if it turns out you’re not good enough”. I don’t necessarily endorse the tone, but I do endorse the idea. If you’re a talented and ambitious computer scientist, engineer or hacker in Europe, you have an option that didn’t really exist ten years ago.

And we believe it’s an option worth harvesting.

***

[1] There are no perfect ideas on day one, perhaps ever. Good ideas emerge from efforts to build something useful. If you intend to wait to start a startup until you have the perfect idea, you will have a very long wait.

[1b] This originally said “users” (not “customers”) but as Andrew Jude pointed out, this can be a vanity metric. At EF, we strongly agree, so this was a silly thing to write. Edited 27/10/2014

[2] Paul Graham’s discussion of Bill Gates’ ex ante probability of building a big business is interesting. He suggests Gates might have had as high as a 50% chance of creating a $100m+ company.

[3] One important note is that if you believe you have a losing ticket, you’re probably right. What great founders seem to have in common is that they understand how low the odds are in the population, but still believe they will succeed. Michael Dearing calls this personal exceptionalism and it’s something we’ve seen very strongly in our best founders at EF.

[4] One question I’m asked a lot is whether you wouldn’t be better to wait, to get a couple more years experience before starting up. Maybe – but it worries me when the person asking gives the impression of “playing the odds”. A couple of years’ experience doesn’t shift the unconditional expected value of starting a startup enough to make it a rational decision for most people. What definitely seems true is that the number of people who say they will start something after a couple of years in, say, finance is an order of magnitude greater than the number who do. The longer you wait, the harder it is to quit the (rising) monthly salary and comfortable lifestyle.