I think the Agile Alliance does a really good job defining Minimum Viable Product which defines it
as that version of a new product which allows a team to collect the maximum amount of validated learning about customers with the least effort.
The purpose of the MVP is to learn something, usually from potential customers and/or investors. In the early stages, you will want to learn whether your product can work as envisioned, whether customers want what you are offering, and whether investors will give you money to invest in your business. You want to avoid spending time, money and other resources on something for which there is little interest. So the idea is to produce that bear minimum working prototype, learn something from it, and then take those learnings into the next development cycle.
As your business matures, new product features and offerings will be considered. The MVP methodology can be applied to them in turn.
One pitfall to avoid is stretching MVP to include a view of what you think the customer wants. This negates the whole development, learning, feedback, development cycle thereby increasing the time and resources required to be successful in the market.
In my experience advising startups, enthusiasm for this feature or that feature takes over, lengthening the development timeline, without any validation on whether the feature is actually desirable. It is far better to get to market efficiently, see what the customer reaction is, see where the pain points are, and then prioritize and tackle them.
Remember, when you are pre-launch and pre-revenue, you are burning cash every day. Get something out there, see what people have to say about it, and then you are in a much better position to figure out what they really want and how to get it to them.
