Education startups do not appear to succeed because it takes too long to measure their success. You can't just look at graduation rates and senior test scores:

Education startups do not appear to succeed because it takes too long to measure their success. You can't just look at graduation rates and senior test scores: you're producing people who will be building and maintaining things for some 30 years, then managing and planning for some 30 more, and you need data points from this entire time span. We're only really just starting to get big pictures from the changes implemented in the 1960s. It's too early to say very much about the things implemented later on, and we essentially have no data at all on today's startups. That's why they don't seem to succeed: we can't even really know if they're succeeding or not.