Why online stores lose sales, and how to fix it

Online stores lose sales in the same few places, whatever country they sell in, and most of these problems can be fixed without any testing. A/B testing is useful when you have a clear idea and enough traffic. Before that, there is usually a list of things that are simply broken.

Surprise costs at checkout

Unexpected delivery fees, taxes or payment charges that appear at the last step are the most common reason people abandon a basket. Show the full price, including delivery and any tax, on the product page or at least in the basket, not at the final step.

Payment options that do not match the market

People pay differently in different countries. Cash on delivery is still popular in parts of the Middle East and Asia. Buy now, pay later matters in the UK, Europe and the US. Bank transfer and local wallets matter in much of Europe. A checkout that only takes cards loses the buyers who wanted something else.

Offer the two or three methods your buyers actually use, and show any fee for them early. Where cash on delivery is common, add a light check for first or high-value orders to keep returns under control.

For cards, the problem is often the extra security step, 3-D Secure. Buy something from your own store with a real card on a real phone. Many sales are lost at this step, and analytics often blames the payment provider instead of your setup.

Missing delivery details

"Fast delivery" is not enough. Buyers want to know if it arrives before the weekend, what it costs to their city or country and what happens if they are not home.

Show a real delivery date range on the product page, by region if you can, with the fee next to it. Moving this information from checkout to the product page reliably reduces drop-offs.

Checkouts that switch language

A common problem for stores that sell in more than one language: the store is translated, but the checkout switches back to English at the payment step. A customer who has read Arabic, German or Spanish for six screens suddenly sees an English card form, and many of them stop.

Go through the whole purchase in every language on a phone, including error messages. Errors are almost always the part that was missed, and an error in the wrong language at the worst moment looks careless.

Mobile forms

Most shoppers use phones, but most checkout forms were designed on a computer. The usual problems:

• The number keypad does not appear for phone and card fields.
• The form demands a postcode in a country where many addresses do not have one.
• Autofill is switched off.
• A required "company name" field on a store for consumers.

Each of these is a quick fix, not an experiment. Go through your form field by field on a real phone.

Trust in the right place

People decide to trust you on the product page and at payment, not on your About page. Show your returns policy, a real phone number or chat, and your registered company name in the footer. In markets where WhatsApp is the normal way to ask a question, a visible WhatsApp button on product pages often turns a hesitant visitor into a buyer, and people tell you exactly what stopped them.

Measure first, then test

Most stores cannot see their own checkout clearly. Before you change anything, track each checkout step by payment method, device, country and language, and log every error. Then watch a week of session recordings of people who left at checkout. You will see the same few problems again and again.

When testing makes sense

Once the obvious problems are fixed and tracking is in place, testing is worth it for things like price display, product page layout and checkout order. With fewer than a few hundred sales a month, most tests take too long to give a clear answer.

So: fix what is broken, measure what is left, then test.

A worked example

A store sees plenty of people add items to the basket, but few finish checking out. Before testing new designs, the team walks through the checkout on a phone. Delivery costs appear only on the last step. The local payment method people prefer is missing. The phone number field opens a letter keyboard. And the page switches from Arabic to English halfway through.

None of these need a test. They are simply fixed: delivery costs shown on the product page, the missing payment option added, the right keyboard set for each field and the language kept consistent. Only after that does testing make sense, on questions where the answer is not obvious.

A checkout checklist

• Total cost, including delivery and tax, visible before the last step.
• Payment options people in your market actually use.
• Guest checkout, with an account offered after the purchase.
• Delivery time and returns explained next to the buy button.
• Short forms, autofill turned on and the right keyboard for each field.
• One language from basket to confirmation.
• Trust signals at the moment of payment: secure payment marks, returns and contact details.

Common mistakes

• Testing button colours while costs are hidden. Fix the big problems first.
• Forcing an account before purchase. Many people leave rather than make another password.
• Only checking on desktop. Most shoppers buy on their phone.

Questions we hear

How do we find where people leave?

Set up funnel tracking from product page to confirmation, then watch session recordings of people who left. The pattern is usually clear within a few days.

When is A/B testing worth it?

When you have enough orders to get a clear answer and the obvious problems are already fixed. Below that, careful changes and watching the numbers work better.

Should we offer free delivery?

Often, above a set order value. Test the threshold rather than guessing, and show it clearly before checkout.