I've implemented a web form for users to request a transfer or withdrawal of funds they have deposited with our service. Our financial person, who handles most finance-related customer support, is asking me to add a confirmation step -- after they submit the form, it goes to a page that displays all their choices and requires them to submit this to confirm.
I know this is a common practice at many web sites; for instance, my town's website for paying excise taxes and parking tickets has one of these confirmation pages. But I wonder if any studies have actually been done to determine how effective they are at catching and preventing submission errors. I suspect most people just glance at the page and automatically click the submit button, so they find it to be just a nuisance.
My application isn't really amenable to the "shopping cart" and "checkout" metaphor, it's just a one-time transaction. We're discontinuing a service that uses these deposited funds, so the users need to move those funds, and the form asks them where to move them to.
More specifically, the user has the following choices:
- Move some of the funds to another type of account.
- Withdraw the remaining funds; in this case, they have a choice between sending to a PayPal account or receiving a paper check by snail mail.
So the purpose of the confirmation step would be to ensure that they entered the move/withdraw split correctly, and that they specified the correct destination for withdrawals. We provide defaults for the destination information if we have it in our records, but we don't have this information for all customers (or if we do, it's not necessarily verified correct).
The question is: when is it appropriate to add a confirmation step in the userflow where one didn't already exist? Is there research or case studies to show when it is effective to do so, and when it is unnecessary?