An abandonment survey tells you what to look for, not how often it happens to you
Baymard Institute’s cart abandonment page is quoted mostly for its average rate. Further down sits a second set of figures that travels into more slide decks: the reasons US online shoppers give for leaving an order. Leaving aside the 42% who were only browsing, the list runs from extra costs that were too high, named by 40%, down to too few payment methods at 9% and a final 7% who do not know why they left. It reads like a ranking of what goes wrong in a checkout, and it is often used as one: start at the top, work down.
That use asks more of the list than it can give. The page itself says little about the study behind it. It calls it Baymard’s latest quantitative study of reasons for abandonment, links to the full study in its Premium research section, and gives no sample size, survey date or question wording. The page was last updated in September 2025. None of this makes the figures wrong. It does decide what they are good for.
The percentages count shoppers, not lost orders
Add up the eleven reasons and the total comes to 182%. That only works if respondents could name more than one reason, so each figure is the share of shoppers who mentioned a reason, not the share of abandoned orders it explains. A shopper who once left over a declined card and often leaves over delivery charges appears under both. The list therefore cannot be turned into a split of your lost revenue, and the gap between two neighbouring reasons says little about which one costs you more.
The population matters as well. These are US shoppers, and several answers carry the US payment mix with them. Not trusting the site with card details, at 19%, and a declined card, at 10%, are card questions. In the Netherlands the Betaalvereniging reports that 71% of online purchases are paid with iDEAL and 8% by credit card, so in a Dutch checkout those two answers concern a minority of orders, while a bank approval that fails or never returns the shopper to the shop does not appear on the list at all.
Shoppers report the reasons they can name
A survey records what a shopper remembers and can put into words afterwards. A delivery charge that felt too high is easy to name. A field that rejected a postcode without saying why, a button that did not respond, or a label that made the shopper unsure is harder to recall as a reason and more likely to end up as a too complicated checkout, an error, or the 7% who do not know. The list is strongest on visible, memorable causes and weakest on the small frictions that a usability test picks up. Baymard’s own remark that many of these issues can be fixed through design alone applies most to that second group, which the survey sees least clearly.
Each kind of reason needs its own evidence in your shop
The useful step is to sort the reasons by how you can check them, not by their percentage.
Some are already in your own systems. Errors and crashes, declined cards and failed payments leave a trace: error logs, payment statuses per method, the number of shoppers who return to the payment step after a failure. Count them instead of asking about them.
Some you can find by going through the checkout yourself. Whether an account is required, whether the total including delivery is visible before the first field, and how many methods are on offer can be established in ten minutes on a phone. Your analytics then show how many shoppers leave on the step where each one occurs.
The rest only the shopper can tell you: costs that felt too high, delivery that was too slow, a returns policy that did not convince, a site they did not trust. For those, ask your own shoppers. A single question shown to people who leave a checkout step, with Baymard’s reasons as answer options, more than one answer allowed and a free-text field, gives you a list in the same format with your own shoppers behind it. Show it on the checkout steps only, so that visitors who were only browsing do not fill it, and add the reasons Baymard’s US list cannot contain, such as a bank approval that did not open or a payment method the shopper did not recognise.
Rank the reasons on your own figures this quarter
Make a table with Baymard’s reasons in the rows. Next to each, write where your evidence comes from (systems, walk-through or shopper question) and your own figure for the last quarter. Then rank the rows on orders lost in your shop, not on Baymard’s percentages. Keep the published list for one purpose: checking that no row in your table has been left without evidence. A reason you have not measured is the one most likely to be costing you orders you do not see.
Sources: baymard.com and betaalvereniging.nl
