Checkout abandonment is usually treated as a friction problem to be solved with fewer fields. A large share of it is something else: the buyer deciding the seller is not safe.
Checkout is the moment risk becomes concrete
Browsing costs the buyer nothing. Entering payment details commits money to a party they may not know, on the basis of images and text.
At that point the buyer starts looking for reasons not to proceed, and small signals carry weight they would not have carried a minute earlier.
This is why abandonment concentrates at the payment step rather than distributing evenly through the funnel, and why field-count experiments produce modest results.
Unexpected costs read as dishonesty rather than expense
Shipping charges, taxes or handling fees appearing for the first time at checkout are frequently cited as the leading cause of abandonment.
The damage is not primarily the money. A buyer who has been shown one number and then another concludes that the seller was managing their perception.
Once that conclusion forms, the buyer reassesses everything else on the site, and the sale is usually lost even if the total is still competitive.
Missing information is read as evasion
An absent returns policy, no delivery estimate, no company address and no route to contact a person all leave the buyer without answers to the questions risk raises.
Where the answer is absent, buyers assume the unfavourable version, because assuming otherwise costs them money if they are wrong.
Supplying the information plainly, including the parts that are unflattering, generally performs better than omitting it, because a stated limitation is still a demonstration of candour.
Familiar payment options substitute for trust in the seller
Recognised payment methods let a buyer rely on an intermediary they already trust rather than on their assessment of an unfamiliar shop.
This is why offering them frequently lifts conversion more than any change to the form itself, particularly for sellers without an established name.
The cost is a higher processing fee, which is best compared against the sales that would not have completed rather than against the fee on sales that would.
Recovery messages work on a different segment
Reminder emails recover a portion of abandoned carts, and most of what they recover is people who were interrupted rather than people who doubted.
Treating recovery rate as a measure of checkout quality confuses the two groups, since the doubting segment rarely responds to a reminder at all.
Separating them, by looking at where in the process each cohort left, points at the specific step that needs work instead of at the checkout in general.