Web to app: moving a mobile web visitor into the app without losing their place

Mobile web gets the traffic; the app gets the conversion. The obvious move is to send web visitors to the app. The obvious failure is that the visitor who was three fields into checkout installs the app and arrives at the home screen with an empty cart, and now has to start over. A lot of them don't.

What problem does this solve?

The prompt to install has to carry the visitor's current state across the install. The item, the cart, the article, the search — whatever they were doing on web is what the app should open on. Then you can measure whether the app actually converts better, because you can see the same person on both sides.

How do you set the links up?

Generate the link on the web page with the current context in the payload. A product page passes the product; a cart passes the cart; an article passes the article. The source tag says which web surface the prompt was on, so you can see which pages are worth prompting from at all.

{
  "deepLinkPath": "/product/sku-4471",
  "customData": {"src": "web-pdp", "cartId": "wc_3310", "variant": "banner-bottom"}
}

Where should the prompt live?

• A dismissible banner on high-intent pages — product, cart, saved items — not on every page. A prompt on the homepage is noise; one on the cart is a fair offer.
• Never on the page a visitor lands on from a campaign. They came for the thing on that page; prompting them to install before showing it is how you lose them.
• The link is the same whether it's a banner, a button or a link in the checkout flow. What changes is the source tag.

Does the app actually convert better?

This is the question the setup exists to answer, and the honest answer varies by product. Compare conversion for visitors who were prompted and installed against those who stayed on web from the same entry page. If the app wins, prompt more. If it doesn't, the prompt is friction and you now have the evidence to remove it.

Which metric should you watch?

Conversion rate within 24 hours of the handoff, for users who arrived in the app with context intact, versus the same cohort on web. That comparison is the business case for the app, made with your own numbers.