Your first idea will very likely not survive contact with a customer. You’ll need to learn and adapt, fast! How can you learn as much as you can about your customer, the problem to be solved, and the potential solution early on, when changing course is not so costly? This is what’s meant by failing early.
In a sense, failing in this way is not really failing. Sure, you’ll need to kiss your original idea goodbye and change direction. By doing so, you take another step on the road to success.
In validation, experiments are the tools you use to try to learn faster. Experiments allow you to “fail” in a controlled way.
When an experiment tells you that a fundamental assumption behind your idea is flawed, you’ll need to change direction: you’ll need to pivot.
A pivot can be relatively simple, like changing the price of a product, or it can be more complex. For instance, your findings might indicate that you need to approach a totally different customer segment, solve a different problem for your customer, or that the customers you’re targeting have a completely different need.