The Checkout Is Where Sales Go to Die: A Field Guide to Fixing It

You paid for the ad, won the click, survived the comparison shopping, and earned the "add to cart." Then roughly seven out of ten of those carts died in your checkout — the highest-intent, lowest-cost traffic a business ever gets, leaking at the last step. Some abandonment is human nature; a large share is design: costs revealed too late, accounts demanded too early, forms grown too long, and payment methods missing at the exact moment of decision. This is the field guide to the fixable share — the friction audit, payment coverage, trust signals, mobile physics, and the recovery stack for the carts you still lose.

Why carts actually die

The abandonment literature is remarkably consistent across years of buyer surveys, and the reasons sort into three families. Unfixable-ish: browsing, price research, saving for later — the window-shopping share you optimize around, not against (and partially recover later). Economic: extra costs revealed late — shipping, taxes, fees — the perennial number-one fixable reason; slow delivery; weak return terms (the conversion role of the refund policy we covered in the refunds guide). Frictional and trust: forced account creation, long or confusing flows, totals you can't see until the end, security doubts, missing payment methods, and technical failures — including the declined payment that ends an intent that was never in question. The organizing principle for everything that follows: a buyer at "add to cart" has budgeted a certain amount of money, time, and trust for the transaction — and abandons the moment the checkout demands more of any of the three than they budgeted. Optimization is the discipline of not exceeding the budget.

The friction audit

• Surprise costs → total transparency, early. Shipping estimates on the product page or in the cart, taxes computed before the final screen, no fee ambushes. If your economics require a shipping charge, show it before the buyer has typed their address to discover it.
• Forced accounts → guest checkout, always. Account creation is a relationship request at a transaction moment — sequence it after payment ("save your info for next time?"), when saying yes is one click instead of a toll.
• Form length → ruthless field triage. Every field is a micro-toll: kill optional fields, merge name fields, use address autocomplete and browser autofill (correctly tagged inputs), and never make the buyer type anything twice. The best checkout form is mostly typed by the browser.
• Flow opacity → progress and editability. Show where the buyer is, keep the cart visible and editable, and make the final review screen an actual review — total, items, delivery, one button.
• Performance → speed is conversion. Every second of load time sheds buyers; the checkout should be the fastest pages on the site, not the heaviest.

Payment coverage as conversion infrastructure

A missing preferred payment method is an invisible decline: the buyer doesn't file a complaint, they just leave. Coverage priorities: cards, cleanly implemented (correct field formatting, camera card-scan on mobile, immediate validation — per the acceptance guide); the major digital wallets — phone and browser wallets that collapse the entire form into a biometric tap, offered as express checkout at the top rather than an afterthought below the form they'd have replaced; and pay-by-bank where the ticket size or margin profile fits, per the account-to-account report. Two second-order effects merchants underweight. Wallets don't just remove typing — they improve authorization quality: network-tokenized, device-authenticated transactions get approved by issuers at measurably higher rates than hand-keyed card numbers, so the wallet button fights both abandonment and the false-decline problem from the declines guide simultaneously. And stored credentials (with proper consent) make the second purchase one click — checkout optimization compounds through repeat rates, which is where its ROI actually lives for most businesses.

Trust at the moment of doubt

Somewhere around the card-number field, every first-time buyer runs a private risk assessment: is this site real, and will this go wrong? The checkout's job is to answer before the question fully forms. The signals with evidence behind them: recognizable payment logos and wallet buttons (borrowed trust — the strongest single signal a small brand can display); security indicators where the card is typed (the padlock, the badge, a one-line "your card details are encrypted" — backed by real PCI-compliant handling underneath); the refund policy, visible at checkout — a one-line guarantee near the buy button converts doubt into margin-of-error; human reachability (a support email or number near the payment step reads as accountability); and coherence — professional, consistent design through the payment pages, because buyers read sloppiness as risk precisely where money changes hands. Trust signals are cheap, additive, and — unlike discounts — cost nothing per conversion they save.

Mobile's special physics

Mobile is where most traffic lives and where abandonment runs highest — because every desktop friction is worse on a phone: forms are harder, typos likelier, tabs more disposable, sessions more interruptible. The mobile checklist is the desktop one, intensified: wallet-first (on mobile, the wallet button isn't an option among options — it's the difference between a biometric tap and a nine-field form typed with thumbs); numeric keyboards for card and ZIP fields, camera card capture, aggressive autofill; single-column layout with buttons sized for thumbs and the primary action always in reach; and session persistence — the interrupted buyer who returns should find the cart exactly as left, because on mobile, interruption isn't abandonment until you make it so. Test it the honest way: buy from your own store on your own phone, on cellular, in a hurry. Every sigh is a percentage point.

The recovery stack

1. Capture early. Email requested before the payment step (with honest framing) is what makes recovery possible at all.
2. The sequence. A reminder within a few hours ("your cart's saved"), a follow-up next day — with the cart preserved behind a single click, on any device. Most recovered revenue comes from removing re-entry friction, not from persuasion.
3. Discount discipline. Hold coupons for the final touch, if at all — a store that discounts every abandonment trains its buyers to abandon on purpose.
4. Failed payments are their own category. A decline is an intent interrupted, not withdrawn: smart retry timing, instant "try another method" paths, and card-updater logic recover a large share — the full machinery in the involuntary-churn playbook.
5. Instrument the funnel. Track abandonment by step — cart → info → shipping → payment → done — because the step where buyers leave is the diagnosis: payment-step exits point at methods, declines, and trust; shipping-step exits point at costs.

Frequently asked questions

This guide is educational. Abandonment benchmarks vary by industry, ticket size, and traffic mix; test changes against your own funnel data.