When to send it
Analytics show you where visitors click and where they leave. They don't show you what someone was hoping to find. This survey asks visitors directly, while they're on your site, so the answers describe the visit they're having right now rather than a fuzzy memory of it.
There are two ways to place it. On the Free plan, put it on a dedicated feedback page or in your footer as an inline website embed from the Embed page. On Pro, show it as a popup or slide-in with a trigger, such as after a delay or once a visitor has scrolled halfway down a page. Keep the frequency low: once a month per visitor is plenty, and hide it after someone has answered. Avoid the checkout itself, where any interruption costs you orders.
Why these questions
The first question asks what brought the visitor to your site today: a specific purchase, browsing, comparing, checking on an order or something else. It sets the context for every other answer. A browser who couldn't find anything is a merchandising question; someone who came for one product and couldn't find it is a search or navigation problem.
The second asks whether they found what they were looking for. Anyone who answers partly or no is asked what they were after, which gives you a list of searches your site failed in the visitor's own words, including products you don't stock yet.
Next comes a customer effort statement on a seven-point agreement scale: this site made it easy to find what I wanted. Effort suits a website better than satisfaction does: a site is a tool, and the best ones go unnoticed.
The grid asks visitors to rate five parts of the store, from search and menus to how clearly prices and delivery costs are shown. It shows where to start when everything feels like it needs work. The last question asks for the one thing they would change, which often surfaces the problem none of your options mention.
The coupon
A shared 10% code goes to everyone who finishes. It works for buyers and browsers alike, and for someone halfway through deciding, it can be the push to order. Since the survey runs on your own site, a shared code is simpler than unique ones and the risk of it spreading is small.
What to do with the answers
Read the couldn't-find answers first. Repeated product names are demand you're not meeting; repeated vague answers point to search or menus. Sort the grid by the lowest-rated row and fix that before redesigning anything. Re-run the survey a month after each change to see whether the effort score moves.
If visitors are leaving at checkout specifically, the cart abandonment survey asks them why. For the effort of checkout and returns on its own, use customer effort for checkout and returns.