Getting paying customer(s) before we'd written the code.
Understanding the value of what your solution will bring and targeting customers who will pay for that value. For years we sold software at an average ticket of $5K back in 2000 - it was like hunting rabbits.
Then we hired someone to help us interview our first group of clients and really zero in on the value we were providing. Pretty soon we were selling that very same software for an average ticket price of $125K. This made an enormous difference in our company -- not to mention in helping me replace all the generic Cheerios in my cupboards with name brand cereal like Cracklin Oat Bran, HONEY nut Cheerios (omg, the honey on top of the cheerios .... it was just mind blowing) and other favorites. :)
We just didn't know how to accurately describe what we had created and fell into the trap of undervaluing and zipping out feature lists a mile long to prospects. As a result of these interviews and follow on projects with customers and prospects, we were also able to develop ROI calculators so we could help prospects explain the value we provided. Then as we added more features to our solution, our value increased and before we knew it we were selling to larger companies for larger ticket prices.
A lot of times you will want to focus on the product and its many features. I think it is <at least> equally important to focus on your pricing model so you can move your solution into the marketplace in a way that maximizes your ability to fund your venture. If done correctly, you may not need any investors - ever.
We enjoyed a smaller client list in the early days, but much larger revenues. Then we focused on making those first batch of customers evangelistic in every way about our software.