I should have said, "traditional market research" instead, including trend and competitive analysis, segmentation, primary research such as focus groups and surveys, prototyping in front of a representative sample of users, etc.
An iterative approach is fine, once the above basics have been done, but if you need to deploy code changes in an app every day or even every week, the product is probably not ready for V1 yet.
Well of course all that is done before all of the developing of the app. The idea is to reduce the amount of uncertainty by not creating a full blown app that has had a lot of time and money put into it, when some of the features aren't helping in terms of getting new customers, keeping old customers and bringing any value to the product/service.
An iterative approach is fine, once the above basics have been done, but if you need to deploy code changes in an app every day or even every week, the product is probably not ready for V1 yet.