Around 70% of all e-commerce carts are abandoned. That doesn't mean every one of those shoppers was ready to buy. Comparing prices, browsing, and saving products for later are all normal parts of online shopping. But a meaningful share of abandonment is caused by avoidable friction in the checkout.
That's where the opportunity lies. Baymard estimates that large e-commerce sites could increase conversion rates by an average of up to 35% through improvements to checkout design.
Why button color isn't the real question
“Red converts better” is one of the most familiar rules in CRO. An often-cited A/B test by Performable seems to support it at first: Across more than 2,000 visits, a red CTA button produced 21% more clicks than the green version.
But this is exactly why A/B tests need to be interpreted carefully. The test doesn't prove that red converts better than green in general. It shows that in this particular page design, the red version performed better.
Color never works in isolation. Contrast, visual hierarchy, brand identity, copy, and placement all affect whether a CTA stands out and is understood.
So the better CRO question isn't “Which color converts best?” It's “Which version works better in our context?”
This matters especially for premium brands. An aggressive red purchase button may attract attention while clashing with a restrained brand aesthetic. That's why the decision should be tested rather than derived from generic rules about color.
A cart isn't a purchase yet
A product in the cart signals interest, but it doesn't mean the purchase decision has been made. Baymard's research shows that a substantial share of cart abandonment is simply part of normal online shopping behavior: People compare prices, save products, or aren't ready to buy yet.
A good cart should therefore do one thing particularly well: make it easy to pick up where the customer left off.
Cart contents should remain available for returning users, products should be easy to recognize, and customers should be able to change quantities, variants, or delivery options without unnecessary friction.
Abandoned-cart messaging can also help customers return to an unfinished purchase. But open and conversion rates vary widely by industry, audience, and timing. Instead of relying on a supposedly universal benchmark, retailers should measure open rate, click-through rate, and actual recovered revenue in their own store.
What checkout research reveals about the biggest reasons for abandonment
1. Forced account creation
An account can be valuable to the retailer. Before the purchase, it's often just more work for the customer.
Baymard currently reports that 19% of users abandon checkout because they are required to create an account.
The takeaway is simple: Make guest checkout clearly visible and move account creation to a point after the purchase has been completed.
2. A checkout that's too long or complicated
Not every additional field automatically costs conversions. But unnecessary fields increase perceived complexity.
According to Baymard, 18% of users say they abandon a purchase because the checkout process is too long or complicated.
At the same time, Baymard's benchmark shows that most checkouts could get by with around 8 form fields, while the 2024 average was still 11.3 fields.
The takeaway isn't “remove every field.” It's: Every field has to earn its place in the checkout.
Company name, address line 2, phone number, or additional registration information should only be prominent when they're actually needed.
3. Additional costs appear too late
The biggest avoidable reason for abandonment isn't button design. It's price.
Current Baymard data shows that 39% of users abandon because of additional costs such as shipping, taxes, or fees.
The strategic takeaway: Show the actual total cost as early as possible. When shipping costs appear only in the final checkout step, the customer isn't just facing an additional expense. They have to reevaluate a purchase decision they thought they had already made.
4. Lack of trust
Checkout is the moment when customers hand over sensitive information and money. Trust matters accordingly.
Baymard's data shows that 19% of users abandon a purchase because they don't trust the site with their credit card information.
Trust isn't created by adding as many logos or “trust badges” as possible. It's the overall experience that matters: a professional interface, familiar payment providers, transparent costs, clear return and delivery policies, and a technically sound checkout.
5. Too many steps, or simply too much work?
“Fewer checkout steps are always better” sounds logical. It's also too simplistic.
Baymard's research shows that the number of form fields has a greater impact on perceived checkout complexity than the number of steps alone. A well-structured three- or four-step checkout can therefore feel easier than an overloaded one-page checkout.
A progress indicator can help users stay oriented in a multi-step process. But it isn't a guaranteed conversion lever.
The better rule is: Don't minimize steps. Minimize effort.
6. Mobile is the biggest opportunity and the biggest challenge
Mobile dominates e-commerce traffic but still converts at a lower rate than desktop.
Current Contentsquare benchmarks summarized by Shopify show:
That doesn't mean poor mobile UX alone explains the difference. Customers switch devices, research on mobile, and buy later on desktop. Product category and usage context also play a role.
Still, the scale of the gap makes one thing clear: Mobile checkout can't be treated as a scaled-down desktop experience. Input fields, keyboards, autofill, address lookup, payment, and error messages all need to work for small screens.
7. Missing payment options
Not every customer wants to pay by credit card.
Baymard's current data shows that 9% of checkout abandoners leave because their preferred payment method isn't available.
Which payment methods matter depends on the market. Retailers shouldn't integrate as many payment providers as possible. They should prioritize the options that matter to their customers and markets.
A/B testing: how to do it properly
The most important lesson from CRO isn't that red buttons work better or that a particular form field always costs conversions.
The lesson is: Context beats best practice.
An A/B test that's stopped after 200 visitors because variant B happens to be ahead hardly provides a reliable basis for a decision.
Four principles matter:
- Determine the sample size in advance: It depends on baseline conversion, the minimum detectable effect, statistical power, and significance level. There's no universal minimum such as “15,000 visitors per variant.”
- Formulate a clear hypothesis: Not “We're testing a red button,” but, for example: “Stronger visual contrast on the primary CTA increases the checkout completion rate.”
- Define the primary metric in advance: Checkout completion, revenue per visitor, or another business-relevant KPI. Different metrics can lead to different conclusions.
- Don't stop because of an early lead: Day-of-week, weekly, and seasonal effects can distort short-term results. Test duration and statistical analysis should be set before you look at the results.
A good A/B test therefore doesn't just answer the question “Which variant wins?” It also answers “What have we learned about our customers' behavior?”
What this means for B2B
B2B e-commerce, quote request flows, and configurators differ significantly from a classic B2C checkout. The underlying UX question is similar, though: Where does the process create unnecessary friction?
In a B2B process, that friction can look different: 18 required fields for a quote request, registration before downloading technical data, unclear next steps, or a configurator that doesn't show users how far along they are.
That's why the specific B2C percentages can't be transferred one-to-one to B2B. The methodology behind them can: Measure where users drop off. Form hypotheses about the cause. Reduce friction in a targeted way. Test changes under controlled conditions.
Because the biggest checkout killer isn't necessarily a red button, an extra form field, or a missing progress bar.
It's the friction you're not measuring.
Sources: Baymard Institute (2025), Cart Abandonment Rate Statistics · MacDonald (2026), How to Reduce Cart Abandonment, Baymard Institute · Scott (2024), Checkout Optimization: 5 Ways to Minimize Form Fields, Baymard Institute · HubSpot, A/B Test Performable (Red vs. Green Button) · Contentsquare (2026), Digital Experience Benchmark, via Shopify.
Published June 2026