Most checkout work starts with addition: a reassurance line, a progress bar, a consent checkbox, a tooltip about the phone number. Each addition looks harmless. Together they make the form slower to read and easier to abandon.
The better question is not whether a field could ever be useful, but whether the order can be fulfilled without it. If not, delete it, move it after payment, or make it optional and say so on the label.
A shorter form is not automatically a better one. Remove the wrong field and the cost moves to your operations: failed deliveries, support tickets, weaker fraud checks.
Which checkout fields actually earn their place?
Sort every field into three groups. Needed to fulfil the order: payment details, a delivery address, an email for the receipt. Needed by law, tax rules, your processor or your courier: a billing postcode for verification, a VAT number on a business invoice. Collected for reporting: company name, job title, "how did you hear about us".
Only the first two justify a required field, and phone numbers are the classic offender: most stores collect one for delivery, then never call it. Count the times your team dialled a customer last month; if the answer is near zero, this is a marketing field wearing a logistics costume.
| Field | Why it is there | Consequence of removing it |
|---|---|---|
| Receipt and delivery updates | No receipt or tracking link | |
| Payment details | The transaction, via your processor | No sale |
| Shipping address | Fulfilment and courier routing | Order cannot ship |
| Billing address | Fraud checks and tax jurisdiction | Weaker fraud checks in some regions |
| Phone number | Courier contact in theory, marketing in practice | Usually nothing, real for bulky deliveries |
| Account password | A login nobody asked for | Slower checkout, account offered after payment instead |
Should checkout be single column or multi-step?
Single column, one field per row, wins for most checkouts: the eye moves in one direction and the mobile keyboard never jumps sideways. Two columns cause skipped fields, usually the second address line or the postcode.
Multi-step pays off when the form is genuinely long or the path branches by delivery type. Each step is a small commitment, but also a page load and a fresh chance to leave. One page with clear groups, Contact, Delivery, Payment, also keeps the total visible. The same rules apply to shorter captures in our guide to high-converting web forms.
Is guest checkout still the right default?
Yes, with account creation offered after payment. A password field above the pay button turns a purchase into a signup and asks for a commitment the shopper has not agreed to yet. Move it to the thank-you page, where saved addresses mean something. Logged-in customers should never retype a stored address, and the guest option should not hide in small grey text.
Does address autocomplete beat typing it out?
Autocomplete does not reduce fields, it reduces keystrokes and postcode errors: one search box beats five handmade ones and produces fewer undeliverable parcels. Keep it optional, since new and rural addresses are sometimes missing from the index, and show the parsed fields for confirmation. Silently accepting a wrong match fails worse than typing.
When should validation fire, and what should errors say?
Validate on blur, when the field loses focus, and again on submit. Never on every keystroke: someone typing an email passes through states that are invalid on their own, and a red message mid-word tells them they are failing while they type.
Do not flag a field just because it was tabbed through empty, do not clear what was typed, and do not move the layout when an error appears. On submit, focus the first problem field and keep an error summary at the top.
Error copy should state the fix, not the fault: "Enter a 5 digit postcode, for example 10117" beats "Invalid input". A declined card is "Your bank declined this card, try another payment method".
Where do the order summary, trust signals and keyboards fit?
The order summary belongs where the decision is made: beside the form on desktop, so total, shipping and tax stay visible, and collapsed at the top on mobile with the total always shown.
Trust signals sit next to the pay button, and one credible line beats four badges: the returns policy in a sentence, a support contact, a note on who processes the payment. See website trust signals.
On mobile, correct input types and autocomplete tokens bring the right keyboard and let the browser fill card and address, removing most error-prone typing. On tap targets and thumb reach, see mobile conversion optimization.
How do you measure whether removing a field helped?
Instrument before you cut. Track field level completion, so you can see which field people reach, fill and abandon, plus time spent on each. Track form abandonment as started versus submitted, and record validation errors by field and message, because high error rates usually mean a bad label rather than an unnecessary field.
Then watch the downstream metrics: checkout completion, payment failures, delivery failures, support contacts per hundred orders, refund reasons, fraud and chargeback rate. Removing the phone field can lift completion while making delivery exceptions harder to resolve; removing the coupon field can push the same requests into your inbox.
Change one thing at a time and decide in advance what result would make you keep it. A shorter form is a hypothesis, not a conclusion, and the same logic runs through cart abandonment recovery. Most teams win first by deleting three fields nobody has defended in a year. That measurement loop, plus the surveys that collect the answers you dropped from checkout, is what heycustomer.byako.dev is built for.
FAQ
Q: How many fields should a checkout form have? A: As few as fulfilment allows, which for a typical physical product is roughly six to nine. There is no correct number, only fields the order justifies.
Q: Should I remove the phone number field entirely? A: Remove it if your team never calls customers, and keep it optional if couriers need a contact for bulky or signed-for deliveries, labelled "delivery updates only".
Q: Is a single page checkout better than a multi-step one? A: For most stores yes: the total stays visible and there are no extra page loads. Multi-step wins when the form is long or the path branches by delivery type. Test both on a phone.
Q: Does address autocomplete actually improve conversion? A: It helps by cutting keystrokes and postcode errors rather than by impressing anyone. Keep manual entry available and show the parsed address for confirmation, so a bad match cannot become a delivery failure.