This is during the slow start then laccélération this company that he brings out his theory of Lean for startups.
It recommends creating what it called a Minimum Viable Product (MVP), this is to say the simplest possible product with just the added value able to attract the "early adopter" and start immediately on the market.
His hypothesis is that although it is far from perfect MVP, even stuffed with bugs, it does not matter anyway since the "early adopter" are not looking at the quality of the time that it are enough to interest of innovations.
Then we need to improve it by adding back MVP (or removing) functionality and see if that brings value to users (by studying their behavior).
The best to be completely objective is to have different populations of "early adopters" (each new modification MVP) nayant not been in contact with the ancient version of the product. This can be done for example by distinguishing at linfrastructure uap former users of all new.
When the paying customer of use or conversion rate increases as we continue improving incrementally MVP.
By cons when the rate of use or conversion into paying customers of these new populations stagnate, it should denvisager a "pivot". This is to say use one or more of the following:
Changing the focus of the product on other people,
Highlight a product feature that will become the product itself
Add other features to enrich the product
Eric also warns against certain metrics such as total number of users (including all populations) giving the impression that the product is increasingly used as new populations "early adopter" are less and less numerous and therefore laccélération slows.