I am about two thirds of the way finished reading "The Lean Startup" by Eric Ries. I was already coming to some of the same conclusions as part of my thinking and reading on entrepreneurship; his book helped me solidify things and fleshed in more of the details of implementation.
Essentially, starting a business (especially a high tech one) is a crapshoot. You come up with a brilliant idea. Then you make a business plan or strategy that incorporates some large (leaps of faith) and small assumptions about your product, your customers, the market uptake, revenue and profits. When you are just starting almost nothing is a given.
Even if you are an established company, when launching something new that part of your business is just like a startup.
So the first step is to clearly define what your strategic/business assumptions are; right down to the details. Then you build what he calls a Minimum Viable Product (MVP). This is a product that is good enough to start testing your assumptions but not the fully featured, high quality product you envision.
One by one you test your critical assumptions by validating with potential customers. Don't worry about your reputation at this point. Honestly if you are just starting, you don't have one and any missteps will only affect a few customers.
Then you measure, learn and adjust as you tune your product and revenue engine. Avoid the vanity metrics and find ones that validate real success based on the changes you are making.
If you find a key part of your strategy is not working, you have to decide whether to persevere or pivot. A pivot is really changing your strategy while keeping what you've learned to date. Your overall vision "may" stay the same, but the details may need to change.
Eric gives lots example of companies that waited too long to pivot. I suspect that you can pivot too early and too often. After all, the measurements or data could lead to false interpretation or it could be more of an issue with critical mass or not enough of the right type of marketing.
In the development of our Manifast product (soon to be released) we pivoted at least 3 times; even though we didn't call it that. I wish I could say we pivoted because of excellent data driven feedback, but we didn't know enough then to it as scientifically. I would guess that some entrepreneurs have a little bit of an instinct for adapting their business to connect with customers.
After what I have read so far in his book, even a lean startup is a crapshoot.
The difference is that you acknowledge this and then build a system that can help you determine the winning strategies from the crap.
P.S. Can't wait to finish the book and implement some of the ideas.