A customer adds an item to the cart, reaches checkout, sees a delivery charge, and leaves. The store owner sees an abandoned cart and thinks, “We need a better checkout.” Maybe. But if the real objection is the cost or timing of shipping, a cleaner form may make the same bad surprise arrive in a more attractive layout.
Checkout design and shipping policy meet at the moment of commitment, but they are different levers. One reduces interaction effort and uncertainty. The other changes the economics and promise of fulfillment. Decide which to investigate by learning where people leave and what they expected.
Separate ordinary browsing from preventable friction
Cart abandonment is not a single diagnosis. Some shoppers are comparing, saving items, checking total cost, or simply not ready. Baymard’s checkout research reports an average cart abandonment rate near 70% across studies, while noting that many abandonments reflect ordinary browsing rather than a fixable usability defect. The number is a broad reference, not a target for every store and not evidence that 70% of a particular shop’s carts can be recovered.
Start with your own funnel. Track product view, add to cart, checkout start, shipping step, payment step, and completion. Confirm the events fire correctly and segment by device, traffic source, geography, and order value where sample size allows. A big drop at the shipping step points to a different question than a drop at payment entry.
Ask what kind of friction you see
A surprise at shipping selection: Investigate when cost and delivery timing first become visible. If shoppers only learn the likely total late, show estimates earlier on product and cart pages. If the shipping charge itself is not competitive, presentation alone will not remove the economic objection. Consider whether thresholds, local pickup, or alternative shipping options fit your margins and fulfillment capacity.
A long or confusing checkout: Review required fields, account creation, address validation, error messages, payment availability, and mobile keyboard behavior. Remove steps that do not support fulfillment, security, or a legal requirement. Explain why sensitive information is needed.
A payment failure or unfamiliar handoff: Test payment methods, errors, currency, and redirect flow on real devices. A buyer who cannot recognize the payment page may hesitate even when the transaction is legitimate. Make the merchant identity and support route clear.
A delivery or returns concern: Show realistic delivery estimates and a usable returns policy before the final step. Do not promise speed or free returns unless operations can consistently support them. A clear policy can expose an unattractive offer; then the business must decide whether to change the offer or accept the cost.
Use a policy-versus-design decision table
| Observed pattern | First investigation | Possible response |
|---|---|---|
| Exit when shipping appears | Surprise or cost | Show estimates sooner; review shipping economics |
| Exit while entering details | Effort, errors, privacy | Simplify fields and explain requirements |
| Exit at payment | Method availability, trust, failure | Test payments and clarify the secure handoff |
| Exit after delivery date appears | Timing mismatch | Offer accurate alternatives or adjust promise |
| Many carts never reach checkout | Price, comparison, readiness, product fit | Inspect product information and traffic intent |
Patterns point to questions, not guaranteed causes. Use support tickets, customer interviews, and usability observation to understand why. If you cannot interview abandoners, ask recent purchasers what nearly stopped them and compare that with the funnel.
Test the smallest relevant change
Set a baseline for completed orders per checkout start, average order value, refunds, customer-support contacts, and fulfillment cost. A “free shipping” test may increase conversion while reducing contribution margin. A more prominent shipping estimate may reduce surprises without changing average order economics. Keep an eye on repeat purchase and delivery satisfaction too.
Where volume permits, test one change at a time. If traffic is low, use a staged rollout and look for consistent evidence across several weeks, while accounting for promotions and seasonality. A short experiment with a handful of sales is not a reliable estimate. Record the hypothesis: “Showing the delivery range on the product page will reduce late exits without raising support contacts.”
Avoid optimizing solely for checkout completion. If the policy attracts orders that are unprofitable or creates unfulfillable delivery expectations, the business has not improved overall. Include operations in the decision. Marketing and design cannot promise what the warehouse, carrier, or finance model cannot support.
The evidence is a reason to diagnose, not to redesign automatically
Baymard’s research gives a broad view of checkout usability issues and the fact that abandonment has multiple causes. It supports inspecting the flow, not assuming that a redesign recovers every lost cart. A first-person ecommerce discussion in the research set described a shopper encountering an unexpected shipping cost at checkout. That is a concrete example of price surprise, not a measure of its prevalence.
Baymard’s product-page research also shows why product pages need to answer decision questions. For shipping-sensitive products, delivery costs and timing may belong earlier in the journey—if the business has accurate information to show.
Make the promise sustainable
Before changing policy, calculate the effect by order size, destination, product weight, and return rate. Decide whether shipping is a visible buyer cost, included in price, subsidized above a threshold, or offered through choices. Each approach changes customer expectations and business economics. Explain the terms in simple language and show them before the buyer has invested time in checkout.
The right fix may be a clearer estimate, a shorter form, another payment method, a different carrier, or no change at all if most abandoned carts are ordinary comparison behavior. Find the specific moment and reason first. If the store changes the policy, verify it can deliver. If it changes the interface, verify the policy becomes easier to understand.
A checkout should make a fair deal easy to complete. It cannot turn an unacceptable shipping price into an acceptable one by rearranging the buttons.