An app is for customers who come back.

Only if customers buy repeatedly. Apps earn their place through saved details, fast reordering and notifications — all benefits that accrue on the second and subsequent purchase. For occasional-purchase categories, a fast mobile store is the better investment.

Among customers who install it, typically yes — app users tend to purchase more frequently. The difficulty is that installing is a large barrier, so the app reaches a fraction of your customers. Judge it on total sales, not on the behaviour of the minority who installed.

For physical goods and real-world services, generally no — those are exempt from in-app purchase requirements. Digital goods and services consumed within the app are subject to commission. Most Indian retail falls into the exempt category.

It is a genuine marketing cost, usually through your website, order confirmations, packaging and paid install campaigns. It is worth budgeting explicitly, because a well-built app nobody installs is a common and expensive outcome.

The catalogue, pricing and stock should come from one source that feeds both the website and the app. That is a technical requirement worth insisting on, because manually keeping two storefronts consistent fails quickly.

Frequently the right middle path for retail. Added to the home screen, works offline to a degree, sends notifications on Android, and requires no store or install. It covers a good share of the benefit at a fraction of the cost.