Having been involved directly or indirectly in hundreds of early stage software and Internet companies, I believe the optimal number of founders for a software/Internet startup is between two and four.
Starting a company by yourself is incredibly difficult. It can be done, but it's the exception not the rule. It's rare that a startup follows a clean and predictable trajectory -- if you are alone as a founder, you'll need to navigate this yourself. That's a lonely and thankless task. At the minimum, having other co-founders along for the journey will give you peers to talk to when you want to beat your head against the wall with frustration.
Observing over 70 companies that have gone through the various TechStars programs in Boulder, Boston, and Seattle has resulted in another insight for me -- the best founding teams are "product leaning" with at least 50% of the founders focused on the product. This doesn't necessarily mean that they are software developers, but it does mean that they spend all of their time thinking about, working on, and obsessing about the product.
So if you have two founders, at least one should be product-obsessed. If three founders, then at least two should be product-obsessed. For four co-founders, you can split it two and two but I always encourage three product-obsessed founders in this case (two devs and one product person.)
Being "obsessed with the product" doesn't mean doing this in a vacuum. The best software/Internet companies follow a customer facing and rapid iteration model such as the Lean Startup Methodology popularized by Eric Ries and Steve Blank. In this context, being product-obsessed means "in the context of the customer," which requires the product leaning people to interact frequently with prospective customers and users early and continually in the product development process.
In our book, Do More Faster, TechStars CEO David Cohen and I have an entire theme on "People". One of our favorite chapters is by Dharmesh Shah, the co-founder and CTO of Hubspot, titled "Avoid Co-founder Conflict." In this chapter, Dhramesh describes a series of conflicts that often arise between co-founders and gives suggestions for resolving them.
One of the most common questions I hear is "how do I find a co-founder?" Ironically, you probably already know your co-founder -- it's either a friend, someone you work with, or a colleague of a friend. Aggressively use your network as a starting point as the chance of finding a fit through someone you are already connected with is much greater than trying to recruit, learn, and team up with someone you don't yet know.
Selecting co-founders and building a lasting and effective professional relationship is challenging. If you recognize this going in and invest real time up front in the partnership it will
pay off many times over.