I believe he's quoting numbers from Shikhar Ghosh at HBS. Here's one article that google brought up with a little more detail about how "failure" is defined:
The statistics are disheartening no matter how an entrepreneur defines failure. If failure means liquidating all assets, with investors losing most or all the money they put into the company, then the failure rate for start-ups is 30 to 40 percent, according to Shikhar Ghosh, a senior lecturer at Harvard Business School who has held top executive positions at some eight technology-based start-ups. If failure refers to failing to see the projected return on investment, then the failure rate is 70 to 80 percent. And if failure is defined as declaring a projection and then falling short of meeting it, then the failure rate is a whopping 90 to 95 percent.
I founded a startup that's profitable and doing $5mm in revenue but would be considered a "failure" because at the beginning we played the hockeystick revenue projection dance (partly to appease potential investors, who discount whatever number you give them anyway, and partly because we were wildly optimistic).
1. The 90..95% definition is stupid. Projections are irrelevant other than for purposes of planning, appeasing investors and roughly determining whether the business model is viable. Depending on the entrepreneur's goals, actual long-term ROI is all that matters.
2. In that respect, 70 to 80% failure to generate projected ROI is not bad. I've never been in the 80-percentile of anything of I've done, therefore these statistics are actually pretty encouraging to me.
These terms invalidate the given statistics, rendering them essentially meaningless. These figures are geared toward being consumed by finance professionals, not by people actually running small businesses or who found startups.
The argument Phil Libin is making very silly and very typical kind of advice from people who are not really trying to help you at all.
If you're working for a startup hopefully you are doing it for better reasons.
Excuse me. It is a way to make money. That figure assumes the definition of entrepreneur is only startups as discussed in places like HN, Techcrunch etc.
Not only is the figure of 95% or 99% wrong with respect to "entrepreneurs" (unless of course you increase the time to, say, 100 years at which case it may be true across the board) but its like saying "don't apply to Harvard or YC because the chance...". It doesn't take into account any particular persons idea or qualifications to be a success.
But most importantly, the definition of being an entrepreneur also includes running a small restaurant or opening a physical warehouse and selling exercise equipment or starting a website to sell new and used office desks or a million other ideas that will never make it to HN or TC that can make you a good living.