Blinded by the light,
revved up like a deuce,
another runner in the night
— Bruce Springsteen
Humans are exceptionally bad at calculating odds. We let our limited experience strongly influence our perception of the likelihood of an event. For instance, we tend to vastly overestimate the odds of dying due to a terrorist attack, accidental firearm discharge, or a hurricane, and vastly underestimate causes of deaths like falling, drowning or the flu. The reason for this is that the media constantly reminds us of the dangers of terrorism, hurricanes or flash stories about children who were accidentally shot. Seldom do you find stories about a person drowning, falling, or dying due to a flu being reported on the national news channels. News stories have a tendency to be sensationalized, so as to capture one’s attention and hook in large audiences, and as such they contribute to peoples’ bias when it comes to estimating what is, and is not, likely to occur.
Likewise, overexposure to blissfully happy lottery winners holding up their over-sized checks on TV and in the papers tends to distort peoples’ perception of the likelihood of winning by buying a single ticket. A more mathematical and objective approach to the problem would quickly reveal that the odds are much worse than they appear to be on the surface. 
I can’t help but notice that this is exactly what’s happening to the development/startup world, too. It’s the new gold rush. Far too many developers are trying to build the next big social network, be the next Facebook (or YouTube), gather crowds in the millions, in the hopes of being bought for a ridiculous sum of money by a large company. The media loves these sorts of stories.
As a consequence, developers who are trying to build the next Facebook are akin to lottery ticket buyers. A few of them will succeed and win, but most will fail miserably. How many social networks do we really need? The ad-supported model works for some lucky companies that manage to attract huge crowds while keeping their expenses to a minimum (e.g. PlentyOfFish) or which get acquired (e.g. YouTube, who is otherwise costing Google money). Everyone else is burning cash and wasting the money and good faith of VCs in the process.
I fear that a lot of developers are blinded by the light. Their perception of the actual odds of “making it” are skewed by the media’s continuous coverage of million – if not billion – dollar acquisitions and success stories. And some VCs encourage this behavior in the hope of seeing great returns on their investments. After all these are very wealthy people, and they’re are not interested in small scale success.
Aside from the obvious waste of time and resources, I think that many developers are leaving excellent opportunities on the table in order to pursue a highly unlikely outcome. The ratio of the likelihood of making 10 million with a traditional business plan and the likelihood of making a billion a la YouTube, is not proportional to the different quality of life that those amounts can afford you. If you are broke, have $30K in credit card debt, or are middle class, you’ll find that 10 million dollars could increase the quality of your life much more than going from 10 million to a billion ever could. And it’s important to understand that aiming at a more likely, albeit smaller, outcome does not in any way prevent you from “dreaming big” afterward, once you’ve already achieved success with your first (or first successful) venture.
Would you rather enter a draw for a million dollars with a 1 in 20 chance of winning, or a draw for five hundred million dollars with a 1 in 50,000,000 chance? A rational person would opt for the first, yet most startups today are leaning towards the second draw. They do so because they vastly overestimate their odds of being successful with the second draw.
Create a product and charge people for it. Unless you really have to, don’t take VC money, instead consider bootstrapping your company. One of the main advantages of the software world is the exceptionally small amount of capital needed to get started. If you want to stick to Web applications, use the Software as a Service (SaaS) model and make your users pay for the software and service you provide. You’ll have a much smaller audience, less scalability problems and expenses, and a whole lot more revenue and a greater chance of being profitable. Joel Spolsky (with his gorgeous office spaces) makes millions in revenue thanks to a company that, for most of its existence, has sold a web bug tracker. How many free bug trackers do you know of? How many competitors exist in that market? Many, I’m sure. Yet while Joel’s popularity no doubt helped his company, it still showcases how a business can be successful by building a better mouse trap.
But like David Heinemeier Hansson mentioned, there are countless under the radar companies making money like that.  If you take your eyes off the spotlight, you’ll see that many companies are very successful at what they do, though they’re not famous or making news headlines. Some of them actually strive to not attract too much attention to their success (often measured in millions of dollars), in order to prevent competitors from springing up.
Regardless of whether you’re a household name or not, you don’t even have to create Web applications to be majorly successful. Mobile apps for smartphones, including the iPhone, come to mind. But good old-fashioned desktop applications keep a wide range of software companies in business. That’s why the skewed perception that you can’t make money with commercial desktop software anymore, or that desktop applications are dead, is utterly ridiculous. As a developer/micro ISV/startup, your chances of making money with well designed desktop software are much higher than building any sort of YouTube, Flickr or Facebook clone.
To understand how skewed our perceptions are, you just need to talk with companies who are open to sharing their software sales statistics. You’ll be shocked by the amount of money that’s being made with relatively common software. Balsamiq makes a UI sketching application that sells for $79. The author managed to make $100K in revenue in the first 5 months, mostly by selling the desktop version of his application. And he is certainly far from being one of the biggest winners in this industry. I mention this though because it shows how a decent idea that’s well executed can quickly bring in revenue when you charge your users. And if you think that $100K in five months is small, let me ask you how many free web sites manage to net a comparable monthly income. If you are looking for larger revenues, check out Omni Graffle, which earned The Omni Group millions of dollars, or set your sights on B2B applications (in which market some applications sell for thousand of dollars a piece).
While many developers are blinded by the light, wise ones with a mind for entrepreneurship are building actual software businesses. I invite you to get out there and do the same.
 The concepts I summarized here are much more eloquently illustrated by Dan Gilbert in this TED talk.
 David Heinemeier Hansson makes a similar point in a post of his which inspired this one.