A customer adds a product to the basket and leaves. That event alone cannot tell you whether the price was wrong, delivery was unavailable, the payment failed or the customer was simply comparing options. Reducing cart abandonment starts by locating the point where a plausible purchase becomes difficult.
Do not begin with a discount. Begin by finding out whether a willing customer can complete an ordinary order.
First separate a basket from an attempted payment
Define the stages in your reporting: product viewed, basket created, checkout started, payment attempted and order accepted. Check whether the figures count sessions, people or events. Repeated visits and automated activity can distort a headline abandonment rate.
Use your platform's current documentation to understand its records. Shopify, for example, provides abandoned checkout and payment-event guidance; the details depend on the checkout experience and sales channel. A missing recovery record is not proof that nobody left a basket.
If customers leave before checkout
Look for missing information on the product page. Can buyers see dimensions, compatibility, available variations and a realistic delivery explanation? A visitor who must enter personal information merely to discover whether an item can be delivered may reasonably stop there.
Try a small set of representative baskets: one inexpensive item, an order containing mixed product types and a delivery to a supported postcode outside your usual area. Record the point where shipping, minimum order rules or surcharges become visible. Use actual merchant rules; do not display a delivery promise that fulfilment cannot honour.
A hypothetical example
A homeware shop offers free delivery above a threshold, but a bulky item carries a separate charge. Showing “free delivery” without that exception may encourage basket additions while creating disappointment later. A precise explanation near the product and basket can be a better first intervention than a blanket coupon.
If they start checkout but cannot progress
- Check whether guest checkout is available and appropriate for the store.
- Test address validation with legitimate formats, including flats and business addresses.
- Check that error messages identify the field and preserve the information already entered.
- Try the journey on a phone with the keyboard open.
- Confirm that stock, shipping zones and discount rules produce consistent totals.
A compulsory account can have a valid operational purpose, particularly in a B2B shop. If it is required, explain that purpose and make account recovery usable. Removing every field is not the goal; asking only for information needed to complete the order is.
If the payment fails
Review the payment provider's event record before assuming the customer changed their mind. Distinguish a declined card from an application error or a payment accepted without an order confirmation. A spinner that never finishes needs an engineering fix, not more persuasive copy.
Use the provider's test environment for controlled failure cases. Confirm the accepted-order state, stock adjustment and confirmation behaviour without making artificial purchases in live reporting. Access to payment logs should be restricted, and customer card details should never be copied into a general troubleshooting document.
Recovery comes after the basics
Where recovery messages are permitted and supported, make them accurate and useful. For UK email or text marketing, check the ICO electronic-mail guidance, including any consent or applicable soft opt-in conditions; a checkout email field alone is not permission. A reminder should not claim that stock is reserved unless it really is. Check communication rules and customer preferences before enabling automation. Do not assume that entering an email address authorises unrelated marketing.
Test one identified cause at a time. Compare completed orders and payment success for similar traffic, while watching margin, refunds and support queries. A higher conversion rate produced entirely by heavier discounts may not be a commercial improvement.
A calculation with a clear denominator
As a fictional example, 200 eligible checkout sessions with 150 completed orders give an incomplete-checkout rate of (200 − 150) ÷ 200 = 25%, if each completed order belongs to one of those sessions and the same observation window. This is not a cart-abandonment benchmark. A cart-based report uses a different starting population, and late completions or cross-device visits can change the result.
Keep the product-page launch checklist beside your findings. If the obstacle sits in catalogue, checkout or integrations, use it to brief an ecommerce development review rather than requesting a general visual refresh.
