We create digital solutions that work for businesses
A customer finds a product, compares its details, clicks “Add to cart,” and disappears before payment. Acquisition money has already been spent and the product has created intent, yet no order is completed.
Cart abandonment cannot be eliminated. Some shoppers use carts as wish lists, compare prices, or plan to return later. However, many exits are created by the store itself: unexpected delivery costs, mandatory registration, long forms, unsuitable payment methods, errors, or weak mobile usability.
Baymard’s aggregation of 50 studies reports an average documented cart-abandonment rate of 70.22%. This is not a universal benchmark for every store. Product category, price, device, traffic source, and calculation method all matter. The useful question is not whether your number matches an average, but where your customers stop.
Separate three behaviours: adding a product without opening the cart, viewing the cart without starting checkout, and starting checkout without purchasing. The final group has stronger intent and is therefore the most valuable for usability diagnosis.
A GA4 ecommerce funnel can use:
view_item → add_to_cart → view_cart → begin_checkout → add_shipping_info → add_payment_info → purchase
Measure loss between adjacent steps. A drop before view_cart can signal weak cart feedback. A drop after shipping information points to delivery cost or availability. A payment-stage loss requires gateway, error, and return-flow testing.
Display products, discounts, delivery, fees, and the final total before the last step. If exact delivery requires a postcode or city, show a clear range or calculation rule. Explain free-delivery thresholds on the product page and in the cart.
An account should help repeat customers, not block a first purchase. Allow guest checkout and offer account creation after order confirmation. If an account is essential, explain the real benefit before requesting a password.
A one-page checkout is not automatically easier than three short steps. Remove information that is unnecessary for fulfilment. Make comments optional and reveal company details only when a business purchase is selected.
Baymard’s checkout research indicates that the number of fields customers must consider matters more than the simple number of steps.
Use one column, sufficient spacing, the correct keyboard type, autofill attributes, clear labels, and a visible primary button. Test small real devices. Sticky controls must not cover errors, delivery choices, or consent notices.
Customers need to know where you deliver, the price, and the estimated date. Show relevant services, pickup locations, courier options, free-delivery rules, and limitations. Handle address-API failures without clearing the whole form.
Offer methods appropriate to the market and order: card, supported wallet, payment on delivery, or business invoice. Explain restrictions before redirecting to a gateway.
After a failed payment, preserve the order and allow the customer to retry or switch method. A temporary bank or network failure should not force product discovery and address entry again.
The cart should survive tab closure and later return. Signed-in customers benefit from cross-device synchronisation; guests can use suitable browser storage.
If price or availability changes, explain it instead of silently removing the item. State whether adding to cart reserves stock.
Place a specific correction message next to the field. Preserve name, contact, delivery, and comments after an error, except where payment-security rules require otherwise. Move focus to the first invalid field instead of displaying only a generic message at the top.
Checkout needs the seller’s identity, contact options, delivery and return information, a clear order summary, and a predictable payment process. Policies should open without losing entered data. Real transparency is more persuasive than a wall of generic security badges.
Prevent duplicate submissions after the “Pay” button, show a processing state, and use server-side idempotency. Reduce third-party scripts and unexpected layout movement. Test under mobile-network conditions, not only office Wi-Fi.
Keep checkout focused. Large navigation, promotional carousels, and multiple pop-ups compete with completion. Cross-selling can appear in the cart, but it should not push the checkout button away. Once checkout begins, avoid sending customers back into product discovery.
A large coupon field prompts shoppers to leave and search for a code. Collapse it behind “Have a coupon?” while keeping it accessible. If a code fails, explain whether the issue is expiry, minimum value, product category, or prior use.
Email, SMS, push, or messaging reminders can recover orders when the store has the required permission and lawful basis for that channel. The link should restore the actual cart rather than open the homepage.
Start with a neutral reminder, follow with useful answers to common objections, and use incentives selectively. Automatically discounting every abandoned cart trains regular customers to wait.
Complex or high-value purchases naturally create questions. Provide a visible phone, chat, or concise delivery-and-payment FAQ without covering the form. Support should never request payment credentials in chat.
Fix critical bugs immediately, but test optimisation hypotheses sequentially. Track add_to_cart → begin_checkout, begin_checkout → purchase, performance by payment and delivery method, field errors, device differences, duplicate submissions, revenue, and margin.
Define guardrail metrics such as cancellations, returns, average order value, and support requests. More completed checkouts are not a win if unprofitable or erroneous orders increase.
There is no single abandonment rate until the denominator is defined:
post-add abandonment = 1 − purchases / add_to_cart
checkout abandonment = 1 − purchases / begin_checkout
Use users or sessions consistently and do not mix event counts with item quantities. Exclude tests, bots, internal traffic, and duplicate purchases. Send a unique transaction_id, value, currency, and item data with purchase.
Segment by device, channel, new versus returning customer, category, order value, delivery, and payment. An overall average can hide a failure limited to one browser or payment method.
Week 1 — measurement: validate ecommerce events, complete test purchases, collect errors, and identify the largest funnel loss.
Week 2 — critical barriers: fix mobile failures and duplicate orders, reveal total cost, allow guest purchase, and retain data after errors.
Week 3 — usability: reduce fields, improve autofill and validation, simplify delivery and payment, remove pop-ups, and improve speed.
Week 4 — recovery and testing: preserve carts, launch one responsible reminder flow, test one priority hypothesis, and compare completion, revenue, cancellations, and support demand.
If catalogues, SEO pages, and product management work and the loss is concentrated in checkout, a full redesign may be unnecessary. Improving the cart, delivery integration, forms, payment handling, and ecommerce analytics can be sufficient.
Consider a new architecture when the platform blocks checkout changes, integrations fail repeatedly, required performance cannot be reached, or order data disagrees across the website, CRM, and inventory system.
The stage before the cart is covered in BB STUDIO’s article about the structure, SEO, and conversion of a high-performing product page. Together, the two pages cover the path from product evaluation to purchase.
Cart recovery starts with a reliable checkout, not a discount email. Show the total, offer guest purchase, reduce fields, preserve information, provide relevant delivery and payment, and only then use responsible reminders.
Measure every transition. “Too many abandoned carts” should become a specific task, such as fixing mobile pickup selection or failed-payment retry. These improvements can recover orders without buying additional traffic.
There is no universal normal. Baymard reports an aggregated average of 70.22%, but category, price, device, source, and methodology matter. Compare against your own history and segments.
Yes. A cart can be abandoned at any point after an item is added. An abandoned checkout means the shopper began the order process, which more strongly suggests a form, delivery, or payment barrier.
Usually not. Start with a useful reminder and an easy return link. Automatic immediate discounts reduce margin and can teach repeat shoppers to wait.
No. The number and clarity of fields matter more than the formal number of screens. Three short logical steps may outperform one overloaded page.
At minimum use add_to_cart, view_cart, begin_checkout, and purchase. Add add_shipping_info and add_payment_info for more precise diagnosis.
Analytics and Conversion
Design & UI/UX
For business
Web development
Let’s create something amazing together Leave your number — we will call you back within 15 minutes during working hours.
We will call you back shortly.
