Fail Fast, Fail Often

If you consider only serious start-up companies, and not the vast majority of business attempts that get less than optimal efforts due to day jobs, family commitments, etc, about 1 in 6 start-ups make it.  That is to say that they don't go bankrupt in the first few years (reference).  Of start-ups that make it completely through funding and have an IPO... well, the stock typically enjoys a "run-up" of an average of 105 days, and then within ten years about 63% of them are gone. (source).  Further, according to Martin Lindstrom, about 90% of all new product launches fail. 

Based on my 14 years as both a private equity investor and a software architect before that, I believe that there is very little you can do to improve those statistics.  

However, what you can improve-- and improve dramatically-- is the amount of time, effort, and money that you spend on each attempt.  In every case of a failed start-up that I know about, the proximate cause of death was running out of money, and you run out of money when you do things out of order.  For example, it does not make sense to complete an expensive, functional prototype until you know that somebody will pay you for it.  And if the prototype was not created to scale, you will have to rebuild it if you do find interest.  Of course, the idea may be bad to begin with, but the term 'start-up' is really just a euphemism for 'we don't know what we're doing... yet.'  Ideas can be changed-- and almost always are.  

A big risk related to doing things out of order for start-ups is that you can hire a consultant or employee to do anything.  Paid workers will happily complete whatever task is assigned to them.  Workers like this really have no incentive to work with you in solving the core issues as they are effectively paid by the hour.  Established companies can adjust to this problem, but start-ups have a much harder time recovering from assigning paid workers to ineffective tasks. 

One of the more pernicious problems for innovators is what I call 'Software that does X.' (see X, The Innovator's Software Dilemma)  X is software that you cannot buy as a product or service and that is not templated for a custom-built solution.  It is a completely novel piece of software, often seemingly simple, that appears relatively unimportant at first, but solving X out of order becomes an almost impossible dilemma.  X is one important ingredient in your 'special sauce,' and almost always becomes intricately tied to the core components of your business.  You have to identify X before you build your prototype and treat it with secrecy and care.

Finally, and perhaps most importantly, many innovators do not position their company to be funded or bought.  You have experienced this when you were trying to find buyers or capital if the other party referred to your business as a 'lifestyle company.'  I have never actually heard this vocalized or seen it in print anywhere else, but in my experience, most potential buyers and funders of your company are also a potential threat.  If you have a great, cash-flowing business and have crossed all the hurdles in your path, your business is still only worth what it would cost for the other party to reproduce your effort now.  If you are carrying the debt of past mistakes, have an efficient system in place, or a purchasable X component, what you are likely to find is that not only will you not get funded or bought, but you'll have a brand new competitor in about 6 months who knows everything you know.  NDA or not.