Most teams wait till people stop using their app before asking them why they left. This makes sense. 

You can always improve your chances of getting responses if you speak to people before they leave. You just have to ask them why they’re using your product less.

This involves investing in analytics that helps you understand how often people use your app. The ranges of engagement will be different for every product. When someone drops in activity several weeks in a row,  that’s when you should check in to see why they’ve tuned out. 

They’re still using your app so they’re more likely to respond, but they're heading out the door so it’s important to understand why.

In most cases, you already have a decent sense of why people leave. Feedback is more confirmatory than anything else. Gather your teams best guesses and send them to people as a multiple-choice question. 

It typically boils down to your thing not working, you sending them too much spam, it being too expensive or them finding something better. 

You want to know if there is a predominant reason people leave so that you can do something about it.