The best answer is the one from Christoph Möller, who points out it's really that roughly speaking for every 10 startups, roughly: A) 3 will fail outright B)

The best answer is the one from Christoph Möller, who points out it's really that roughly speaking for every 10 startups, roughly: A) 3 will fail outright B) 3 will return less than the invested capital C) 3 will return the investied capital D) 1 will be highly successful and make investing in all 10 startups highly worth while. I've heard slightly different breakdowns about categories C and D. which are sometimes called the "living dead", and point out of the 1 (or sometimes 2) that actually have a successful exit, only 1 in 100 have an amazingly successful exit such as Google. The bottom line is that success or failure metrics vary greatly whether you are talking about it from the perspective of the VC, the founder, or the employees who just want to have food with their meals. The employees might not consider a "living dead" outcome to be a failure, and depending on the founder's motivations if the "living dead" startup is providing some great service and he or she is more passionate about providing that service as opposed to becoming rich beyond the dreams of all avarice, a founder might not mind all that much either. A good example of this is Cygnus Support, which was technically in the "living dead" category for a decade before the Red Hat acquisition.