The 70% Problem: What a Decade of Research Found About the Cart Abandonment Rate

The 70 percent figure is real, but it is not one number

The headline statistic that circulates in conversion writing is a genuine aggregate, not a marketing invention. The Baymard Institute maintains an ongoing meta-analysis that pools fifty independently conducted cart-abandonment studies; the mean across that corpus is approximately 70.22 percent, meaning that of every ten shoppers who add an item and begin a purchase, roughly seven leave before completing it.

Two cautions belong next to that number before it is used. First, an average pooled across fifty studies with different populations, devices, and definitions is a central tendency, not a benchmark any single store should expect to match. A business with a returning, high-intent audience will sit well below it; a store driven by cold discovery traffic may sit above. Second, the figure conflates shoppers who never intended to buy in this session, the price-comparers and the list-savers, with shoppers who intended to buy and were stopped by the funnel itself. The research that matters is not the aggregate rate but the second group, because only that group is recoverable through better design.

What the checkout abandonment reasons actually are

The reframe comes from the reason taxonomy, not the rate. When Baymard surveys shoppers who abandoned a purchase they intended to make, the causes cluster, and they cluster away from price. The single most-cited reason is that extra costs, shipping, tax, and fees, appeared too high or too late at checkout, named by roughly 39 percent of respondents. The next structural cause is that the site required the shopper to create an account before buying, named by roughly 24 percent.

The precise wording of the leading reason matters. It is not "the price was too high." It is that the total was unexpected, disclosed after the shopper had already committed attention and effort. That is a statement about sequencing and honesty, not about willingness to pay. A shopper who would have paid the full amount had it been shown on the product page abandons when the same amount is revealed on the final screen, because the late reveal reads as a bait, and because it forces a re-decision at the moment of highest friction.

Checkout optimization is mostly the removal of steps

Baymard's second research asset, its Checkout Usability program, gives the constructive half of the picture. Built on more than 300,000 hours of large-scale usability testing since 2011, the program's central quantitative finding is that the average large ecommerce site can gain roughly a 35.26 percent increase in conversion rate from checkout-usability improvements alone, before a single change to traffic, price, or product.

That gain is a corpus-level estimate from one research organization, not a promise for any individual store, and it should be read as directional evidence that the ceiling on checkout improvement is high, not as a figure to forecast against. What gives it credibility is the mechanism the same program identifies: most of the recoverable loss traces to funnels that ask for far more than the transaction requires. Baymard finds that a well-designed checkout can be built from as few as 12 to 14 form elements, of which only 7 to 8 are true input fields, against the 20-plus elements typical on sites that have never been audited. Checkout optimization, in this evidence, is less about adding conviction and more about deleting steps.

How to reduce cart abandonment: sequence and disclose

If the leading causes are late cost and forced accounts, the highest-impact corrections follow directly from the taxonomy rather than from tactics. To reduce cart abandonment in a way the research supports, the moves are structural.

• Disclose the full landed cost early. Show shipping, tax, and fees on the product or cart page, or make them estimable before the final step, so the total the buyer commits to is the total they see at payment. This directly addresses the roughly 39 percent who abandon over unexpected cost.
• Offer a genuine guest checkout. Let a first-time buyer pay without creating an account, and offer account creation after the purchase, not before it. This addresses the roughly 24 percent who abandon over forced registration.
• Cut the form to the fields the transaction needs. Move from a 20-plus-element checkout toward the 12-to-14-element shape the usability corpus describes, removing optional fields and inferring what can be inferred.
• Reduce the mobile penalty. A checkout that is longer or more hostile on a phone abandons more, because that is where a growing share of the traffic decides. Parity between mobile and desktop friction is part of the fix, not an afterthought.

Checkout usability is now a compliance surface too

The disclosure finding has acquired a second dimension since Baymard first quantified it. In its 2022 staff report "Bringing Dark Patterns to Light," the Federal Trade Commission named buried or hidden fees, sometimes called junk fees, as one of four recurring patterns it treats as potentially actionable under Section 5 of the FTC Act. The tactic the research identifies as the top abandonment cause is the same tactic a regulator has flagged as a legal exposure.

The practical consequence is that early, honest cost disclosure is no longer only a conversion tactic; it is the low-risk position on a live compliance question. A store that reveals the full price early both recovers the largest named source of abandonment and stays clear of the pattern the FTC has singled out. Honest sequencing and defensible design point in the same direction here, which is rarely the case and worth stating plainly.

Reading the evidence carefully

The Baymard figures are the strongest public dataset on this question, and they still deserve careful handling. Baymard is a commercially interested party: it sells checkout research and auditing, so its point estimates should be treated as directional rather than as universal constants. The 70.22 percent aggregate pools studies with differing methods. The 35.26 percent conversion-gain figure is an average across the sites in its corpus, not a lift any specific site is guaranteed. The reason percentages come from Baymard's own survey instrument, with the sampling and question wording that implies.

None of that undermines the core claim, because the core claim does not depend on any single decimal. Across independent restatements, the direction is stable: abandonment is high, and the largest recoverable share of it is structural friction, cost surprise, forced accounts, and excess form length, rather than price or persuasion. The right way to use this research is as a diagnostic lens that tells you where to look on your own funnel, followed by measurement on your own store, not as a set of numbers to quote as if they were your results.

From a statistic to a diagnosis

The gap between the research and a decision is specificity. Knowing that the average cart abandonment rate is near 70 percent, and that cost surprise and forced accounts lead the reasons, tells you what to suspect. It does not tell you which of those causes is active on your store, how much each is costing, or which one to fix first. That requires walking your real purchase path, on the devices your buyers actually use, and locating where intent-to-buy turns into abandonment.

This is the difference between a benchmark and a diagnosis. A benchmark is a number you compare yourself to. A diagnosis is a ranked map of the specific leaks on your specific funnel, ordered by revenue at risk against effort to fix, with the reasoning behind each priority. The research in this article is the map legend. The territory is your checkout, and the only way to read it is to measure it.