More sales from the traffic you already have
Merchants keep asking which button label converts better, but the published cart abandonment numbers settle a different question first, which is that the checkout flow around the button matters more than the two or three words printed on it.
Merchants searching for add to cart button optimization usually expect a verdict on which words sell more: Add to Cart, Buy Now, Add to Bag, something shorter, something bolder. That question is not what the published research actually answers. The numbers that exist and are checkable describe the abandonment rate around the entire cart and checkout process, not the isolated effect of button text, and the gap between those two things is the reason so many button-copy claims online cannot be traced back to a real study.
Baymard Institute’s aggregated dataset, pulled together from studies it tracks over time, puts the average cart abandonment rate at 70.22 percent. That is the number a merchant should hold in mind before touching button copy, because it means roughly seven out of every ten carts never turn into an order, for reasons that span shipping costs, account creation friction, payment options, and dozens of other steps between the button and the confirmation page. A button label is one small piece of that path. Changing two or three words on it is not going to move a rate built from that much friction upstream and downstream.
Baymard’s list is not one study, it is a collection, and the spread between the studies is wide. Uptain’s 2025 figure, retrieved in August 2025, came in at 71.72 percent. SalesCycle’s 2023 figure, retrieved in July 2023, was 79.53 percent. Fresh Relevance’s 2022 figure, retrieved in August 2022, was 68.70 percent. Here is that spread laid out against the aggregate.
| Source | Year retrieved | Abandonment rate |
|---|---|---|
| Baymard aggregate | ongoing collection | 70.22% |
| Uptain | 2025 | 71.72% |
| SalesCycle | 2023 | 79.53% |
| Fresh Relevance | 2022 | 68.70% |
The spread between 68.70 percent and 79.53 percent is over eleven points, and none of that spread is explained by button wording. It reflects different sample populations, different device mixes, different points in the checkout funnel where each vendor started counting. A merchant running an internal A/B test on button copy for a few weeks can easily see swings inside that same range purely from traffic mix, and mistake it for a button effect.
Baymard frames its $260 billion figure as orders recoverable by focusing on checkout usability issues alone, and ties that to a 35.26 percent increase in conversion rate. That figure describes fixing the checkout flow broadly: form fields, shipping cost visibility, account requirements, error handling. It says nothing about button wording specifically, and no source here breaks that $260 billion figure down by which UI element contributed what share. A merchant chasing conversion gains should treat the button label as the smallest lever in a much larger machine, and the $260 billion estimate as the argument for auditing the whole checkout rather than one button.
Shopify’s own checkout documentation, read as loaded on September 5, 2026, records two specific capability changes: support for product, order, and shipping savings from a single function extension shipped in April 2025, and support for cart metafields on function input queries shipped in July 2025. These are platform features that let merchants build more sophisticated checkout logic, not conversion results. They matter for a different article, one about what a store can technically configure at checkout, and they do not tell a merchant anything about which two words to print on a button.
Run the audit Baymard’s own findings point toward, on shipping cost disclosure, account creation, and form length, before spending time on button copy, because the 70.22 percent aggregate and the $260 billion usability estimate both say the bigger loss is upstream of the button. If a button copy test is still worth running afterward, expect any result to sit inside noise as wide as the 68.70 percent to 79.53 percent range already documented across published studies, and size the test accordingly.
The published facts in this article do not contain a controlled test isolating button label wording as the only variable, so no number here answers that directly. What the facts do show is the scale of the problem the button sits inside: an average cart abandonment rate of 70.22 percent across the studies Baymard tracks. Any claim about a specific percentage lift from swapping button text is not something these sources support, and a merchant should treat such claims from other sites with the same skepticism this article applies to its own numbers.
Baymard's aggregated figure across the studies it tracks is 70.22 percent. Individual studies inside that aggregate range from 68.70 percent from Fresh Relevance in 2022 up to 79.53 percent from SalesCycle in 2023, with Uptain's 2025 figure at 71.72 percent. A merchant should use the 70.22 percent aggregate as the planning baseline rather than any single study, because the aggregate smooths over methodology differences between the individual reports.
Baymard's figure is $260 billion in orders it considers recoverable by focusing on checkout usability issues alone, which the institute ties to a 35.26 percent increase in conversion rate. That number describes the checkout flow broadly, not the add to cart button specifically, so it should not be read as a promise tied to button copy. It does explain why usability testing on the whole checkout path is a bigger lever than word choice on one button.
Shopify's checkout documentation describes platform capability rather than conversion outcomes. It notes that support for product, order, and shipping savings from a single function extension shipped in April 2025, and that support for cart metafields on function input queries shipped in July 2025. Neither entry measures how button microcopy affects conversion, so they belong in a discussion of what merchants can technically build into checkout, not in a discussion of what words to put on a button.
Given that the published aggregate abandonment rate sits at 70.22 percent and individual studies swing between 68.70 percent and 79.53 percent, a merchant's own store is likely to sit somewhere in a similarly wide band for reasons that have nothing to do with button text. That variance means a small button copy test run for a short period can easily be reading noise rather than a real effect. A merchant who wants to act on button wording should first check whether their checkout flow has the kind of usability issues Baymard's $260 billion estimate is describing, since fixing those is the larger and better documented lever.