Where E-Commerce Checkouts Lose Money

The typical online store loses roughly 70% of the people who add something to a cart before payment completes. That figure has been stable across two decades of research, which tells you the causes are structural rather than fashionable.

It also means the checkout is the single most expensive surface on the site. A homepage redesign moves a percentage of visitors. A checkout fix moves a percentage of people who have already decided to buy.

Here is where the money actually goes, in the order it goes.

1. Costs that appear at the last step

This is consistently the largest single cause of abandonment. Shipping, tax, duties and fees that were not visible earlier arrive at the final step, and the total is no longer the number the customer agreed to in their head.

It is not that the costs are too high. It is that they are a surprise, and a surprise at the payment step reads as a bait and switch.

The fix is arithmetic, not psychology:

• Show shipping cost on the product page, or at minimum in the cart. A postcode field that returns a real number beats "calculated at checkout."
• If you offer free shipping above a threshold, show the gap: "€18 more for free delivery" converts better than a banner nobody reads.
• For international orders, show duties and import VAT before the payment step. A customer who discovers a €40 customs charge on delivery does not order again, and often refuses the parcel.

2. Forced account creation

Requiring an account before purchase is asking for a commitment before you have earned one. Guest checkout is the fix, and the objection — "but we want the customer data" — has the causality backwards.

Offer the account after the purchase, on the confirmation page, pre-filled with the details they just entered. One click to set a password. You capture more accounts this way than by demanding one up front, because you are asking someone who has already bought rather than someone who is still deciding.

3. Too many fields

Every field is a chance to abandon. Most checkouts collect things they do not need.

• Company name — only if you sell B2B, and then optional.
• Address line 2 — merge it into line 1, or make it clearly optional.
• Title / salutation — remove it. It adds nothing and it forces a gender declaration.
• Phone number — only if the carrier genuinely requires it, and say why: "for delivery notifications."
• Confirm email — remove it. Inline validation catches typos better than a second field people paste into.

A useful audit: for each field, ask what would break if it were gone. If the answer is "nothing, but it would be nice to have", delete it.

Use address autocomplete where it exists for your market. It replaces five fields with one and eliminates the mistyped postcode that produces a failed delivery.

4. Payment methods your market does not use

This is the one that quietly kills international expansion. A store offering only card payments works in the UK and fails in markets where cards are not how people pay.

• Netherlands — iDEAL is the default. Without it you lose most of the market.
• Germany — invoice and direct debit remain dominant. Card penetration is far lower than English-speaking markets assume.
• Poland — BLIK.
• Turkey — instalment payments on local cards are expected for higher-value items.
• Gulf states — cash on delivery is still significant.

You do not need every method everywhere. You need the two or three that dominate each market you actually sell into, shown by default to visitors from that market.

5. Trust gaps at the payment step

The moment before entering card details is the moment doubt arrives. What resolves it is specific, not decorative:

• A visible return policy at the payment step, not a link in the footer. "30-day returns, we pay the postage" is worth more than any badge.
• A delivery date, not a delivery window. "Arrives Thursday 12 September" beats "3–5 business days."
• Contact details that look human. A phone number and a real address say more than a chat widget.
• Familiar payment logos. These do measurably more than security seals, because customers recognise them.

Security badges have weak evidence behind them and some studies find they reduce conversion by introducing the idea of risk. Skip them; make the padlock real instead.

6. Mobile forms that fight the keyboard

Over 70% of e-commerce traffic is mobile and it converts at roughly half the desktop rate. A large slice of that gap is form mechanics.

<input type="email" inputmode="email" autocomplete="email" />
<input type="tel" inputmode="tel" autocomplete="tel" />
<input inputmode="numeric" autocomplete="postal-code" />
<input autocomplete="cc-number" inputmode="numeric" />

Correct autocomplete tokens let the browser fill the whole form in one tap. Correct inputmode brings up the numeric keypad for a postcode instead of a full keyboard. These are two attributes per field and they are missing from most checkouts.

Also: never disable paste on card or email fields. Password managers exist and blocking them costs you customers to gain nothing.

7. Errors that arrive too late

Validating on submit means a customer fills eight fields, presses pay, and is thrown back to a page with a red banner. Validate on blur, inline, next to the field.

And when a payment fails, say what to do next. "Your card was declined" leaves the customer stuck. "Your card was declined — this is usually a bank block on international payments. Try another card, or call your bank" recovers a share of them.

What to fix first

In order of return per hour of work:

1. Show total cost early, shipping included
2. Enable guest checkout
3. Delete every field you cannot justify
4. Fix mobile autocomplete and inputmode
5. Add the payment methods your top markets actually use
6. Put returns and delivery dates on the payment step
7. Move validation inline

Items one to four cost days, not weeks, and they are where most of the recoverable money is.

Measure the right number

Track revenue per session, not conversion rate. A change that raises conversion by lowering average order value can leave you worse off, and conversion rate alone will not tell you.

And instrument the funnel per step — cart, address, delivery, payment, confirmation. Without step-level data you know people are leaving but not from where, and you will optimise the wrong step confidently.

If your store does not have enough traffic for statistically valid A/B testing — and most stores do not — say so honestly and work from research-backed practice, measuring directionally. Pretending a test on 400 sessions is conclusive is worse than not testing.