A checkout feature earns its place by answering a doubt before the shopper pays
In October 2021 PYMNTS and Checkout.com published the first edition of Removing Friction at the Checkout, a report built on the Checkout Conversion Index that PYMNTS has run since 2016. The research combined an anonymous shopping simulation at 510 leading merchants, with annual sales from under $50 million to more than $1 billion, with a census-balanced survey of 2,139 US adults carried out in July 2021. The data is five years old and American, so read the percentages as a picture of that market in 2021, not as a benchmark for your shop.
The finding that has aged best is not a consumer statistic. It is what happened to the merchants at the top.
The leaders lost ground while the rest caught up
In the index, the best-scoring shops slipped back on the breadth of their checkout, while the middle and the bottom of the ranking closed in on them. The features the rest of the market added are concrete: a buy button, a guarantee or refund, inventory status and live help on the site. Larger merchants were more likely than their smallest peers to show inventory status and to offer free delivery.
PYMNTS’s published summary does not explain why the leaders slipped. The pattern is familiar from checkout work generally, though. Features rarely disappear because someone decided to remove them. A stock indicator does not survive a platform migration, a help button is switched off when a contract ends, a refund promise moves from the payment step to a page in the footer. Each change looks small on its own. Together they leave a checkout that once answered the shopper’s questions and now leaves them to find the answers elsewhere.
Each feature on that list answers a specific question
Look at those features from the shopper’s side and they read as questions asked just before paying.
Inventory status answers whether the item is actually available, before money changes hands. A shopper who cannot see stock or an expected dispatch date either leaves to check another shop or pays and finds out later, which turns a checkout question into a customer service case. Show availability and the dispatch moment on the product page and again in the basket, per item.
Guarantees and refunds answer whether the shopper can get their money back. PYMNTS’s study page puts the share of medium-sized merchants offering guarantees and refunds at 89%. When nearly nine in ten shops of that size make the promise, a promise the shopper cannot see counts for nothing. Put the refund terms in one sentence on the payment step, near the button, rather than behind a link to the conditions.
Live help answers whether there is someone to ask. At the payment step, help should open without taking the shopper out of the checkout or emptying the basket, and it should not cover the button the shopper is trying to press.
The last question is whether paying here is safe. In the survey, 85.9% of shoppers said a security logo shown in the checkout made them more confident. That figure measures a reported feeling, not completed orders, and PYMNTS’s published summary does not say which logo. A badge on its own proves little to a shopper. What they can check for themselves does more: a shop name that stays the same from basket to payment approval, and an amount on the approval screen that matches the total they agreed to.
Shoppers who browse longer arrive with fewer questions left
PYMNTS also saw shoppers take noticeably longer before they reached the checkout, and the authors took that as a sign that they would put up with less friction once they got there. That is an interpretation, not something the survey measured directly, but it fits the features above. A shopper who has spent time comparing has usually settled on the product and the price. What remains are the practical doubts: availability, refunds, help, safety. The checkout is the last place to answer them, and the first place where an unanswered one costs the order.
Satisfaction was high all the same: 88% of the shoppers said they were very or extremely satisfied with their last online purchase. High satisfaction is not a reason for comfort. It shows how much shoppers had come to expect, which is exactly why a feature that quietly disappears is noticed.
Check whether your checkout still gives the answers it gave last year
Write the four questions in a column: available, refundable, help, safe. Next to each, note where your checkout answers it today, on a phone, at the payment step itself. Then do the same for your checkout as it was a year ago, from screenshots, release notes or whoever built it. Any answer that has moved further from the payment button, or has gone, belongs at the top of the list for the next release. Assign one person to own that list, so the next migration or contract change does not remove an answer without anyone deciding it should.
Sources: pymnts.com, Removing Friction at the Checkout, pymnts.com, report findings, pymnts.com, merchant results and pymnts.com, methodology
