Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Conversion Rate Optimisation › Landing Pages

Forms That Convert

How to design forms that convert: field count and qualification, layout and labels, input types and defaults, error handling, and multi-step forms.
Guidebeginner14 min readUpdated

Every field has a cost

A form is the moment a landing page asks for something back. Up to that point the visitor has only been reading; now they have to type, choose, remember and decide whether the thing on offer is worth the effort and the information. Each field adds to the effort, and each piece of information adds to the question of why it is wanted. Conversion rate (CVR) on the form, measured as completions over starts, is the direct read of how that bargain feels. The form is the last section of the page the anatomy guide describes, and it is where everything the page has built either pays off or is wasted.

The cost is not only length. A field that asks for something the visitor does not have to hand (an account number, a company registration number), that forces a format (a phone number with the country code in a particular style), that asks for something that feels intrusive at this stage (budget, revenue, date of birth) or that is simply ambiguous ("reference") costs more than a plain name-and-email pair. Nielsen Norman Group's form usability recommendations begin with keeping forms short, and the rest of their list is about reducing the other costs: grouping related fields, single-column layout, logical order, visible labels, matched input types, clear required-versus-optional marking, format guidance, and specific, visible error messages.

The way to decide what to ask is to work backwards from the next step. If a sales call follows, the call needs a name, a way to reach the person and perhaps one qualifying question; the rest can be asked on the call. If a quote follows, the quote needs whatever the quote genuinely depends on and nothing else. If a download follows, an email address may be all that is justified. Everything beyond that is a request from someone else in the business (marketing wants the company size, sales wants the budget, finance wants the address) and each such request should have to argue its case against the completions it will cost. What happens once a quote has been sent is a separate design, set out in the tutorial on automating quote follow-ups. The same field-by-field discipline applies to the forms customers meet after they buy, such as booking, returns and claims forms, which the guide to reducing customer effort across the journey treats as one kind of effort among several.

Fields and qualification

The tension on lead forms is between volume and quality. Fewer fields produce more completions and less information about each; more fields produce fewer, better-qualified leads and give the sales team something to prioritise. Neither extreme is right in general. The table sets out the choices.

FieldAsk on the form whenAsk later whenNotes
NameAlmost alwaysRarely worth deferringOne field, not first and last, unless the system genuinely needs both
EmailThe next step is by emailRarelyValidate format, not existence; a typo check ("did you mean…") catches most errors
PhoneThe next step is a call and the visitor expects oneThe visitor may not want a callSay when and why you will call; optional phone fields are often left empty for a reason
CompanyBusiness-to-business, and the offer depends on itIt can be inferred from the email domainFree text; do not force a dropdown of thousands
Company size or roleThe offer or routing differs by segmentThe segment can be found afterwardsA short set of ranges, not an exact figure
Budget or timelineSales needs it to prioritise and the visitor understands whyEarly in the relationshipThe most abandoned question on most forms; a later question or an optional one
Postcode or locationThe service is local or the quote depends on itThe service is nationalUse it to confirm coverage before the rest of the form
Message or requirementsA free-text brief improves the next stepThe next step is a standard callOptional and clearly so
Marketing consentRequired by the law that appliesNever bundled into the submissionUnticked by default where the applicable rules say so; separate from the main action

A reasonable rule is that a field earns its place on the form only if the next step cannot happen without it or if the lead would be routed wrongly without it. The rest belongs in the first conversation, or in a progressive approach where returning visitors are asked new questions rather than the same ones. On a new page, the landing page brief is the place to list the fields kept, the fields removed and the reason for each, before anyone builds the form. Lead scoring, which the lead scoring guide covers, often reduces the need to qualify on the form at all, because behaviour and company data fill in what the form did not ask.

Platforms' own lead forms, such as the native forms on LinkedIn described in the targeting and lead gen forms guide, pre-fill from the person's profile and so change the trade-off: more fields cost less there, and the quality problem moves to how easy the form is to submit by accident.

Layout, labels and input types

Layout decides how much thinking a form takes beyond the questions themselves. The reliable pattern is a single column, fields in the order a person would naturally give the information, related fields grouped under a short heading, and the button at the end of the column where the eye arrives. Multi-column layouts save vertical space and cost completions, because the eye has to work out the reading order and often misses a field.

Labels sit above the field, visible at all times, in plain words. Placeholder text inside the field is not a label: it disappears when the person starts typing, it fails the person who looks back to check what a field was for, and it is often rendered in a contrast that is hard to read. If an example format helps, put it in helper text under the label, not in the placeholder.

Required and optional marking should go one way: mark whichever is the minority. On a short form where everything is required, say so once at the top. On a longer form, mark the optional fields as optional, which is clearer than asterisks that visitors have learned to ignore.

Input types matter more on phones than on desktops. An email field that invokes the email keyboard, a phone field that brings up the number pad, a date field that offers the native picker, a postcode field with the right capitalisation and width: each removes a small friction that adds up across the form. Dropdowns are overused; a short list of options is quicker as radio buttons where the choices are visible at once, and a long list (country, for instance) is quicker as a type-ahead than as a scroll through hundreds of entries. Defaults should be the most common answer, never the one that benefits the business.

Autofill is a conversion feature. Browsers fill name, email, phone and address when the fields are named in the standard way, and a form that blocks or breaks autofill forces typing that the visitor expected to avoid.

Errors and validation

Errors are where forms lose people who had already decided to complete them. The rules are few and specific. Validate as early as is helpful and as late as is kind: check a field when the person leaves it, not on every keystroke, so they are not told they are wrong while still typing. On submission, keep everything the person entered; a form that clears itself on error is asking them to pay the whole cost again.

Error messages say what is wrong and how to fix it, next to the field in question, in words rather than codes, with a colour and an icon so that colour alone is not the cue. "Enter a valid value" is not a message; "Enter the postcode of the property so we can check we cover it" is. A summary at the top of the form, listing the errors and linking to each, helps on long forms and is part of accessible design; on a short form the inline messages are enough.

Validation should accept what people actually type. Phone numbers with spaces, brackets and leading plus signs; names with apostrophes, hyphens and accents; email addresses with plus signs and unfamiliar domain endings; postcodes in lower case with or without the space. A validation rule stricter than the real world is an error message waiting to happen.

Never use a reset or clear button. Nobody intends to clear a form, and the button sits where the submit button might be, which is the single most avoidable way to lose a completed form.

Multi-step forms

A long form can be split into steps, and the split changes the psychology: the first step asks for little, the visitor is committed by the time the harder questions arrive, and the progress indicator makes the end visible. Multi-step forms often complete at a higher rate than the same fields on one page, especially on mobile, and they allow the first step to capture a contact before the rest is asked, so that a partial completion is still a lead.

The design rules: start with the easiest and least personal questions, put the contact details early enough that an abandoned later step still leaves something usable, show progress honestly (steps, not a bar that lies), allow going back without losing entries, and keep each step short enough to fit on a phone screen without scrolling to the button. The questions that cause the most abandonment (budget, phone, anything that feels like commitment) go last, and are optional where they can be.

Multi-step forms also create a measurement opportunity: each step is an event, so the drop-off between steps says which questions cost the most. The checkout, which is the longest multi-step form most sites have, has its own set of problems and the checkout optimisation guide covers them; the rules here are for lead and sign-up forms. An application that has to stay long, for an admission, a grant or a job, also needs eligibility questions first, saved progress and documents last, which the tutorial on redesigning a long application form walks through. A form also asks for trust at the moment of commitment, so the proof elements placed beside it matter as much as the fields, as the trust and proof guide explains.

Accessibility

An accessible form is a form that works for everyone, and most of what makes it accessible also makes it convert. The Web Content Accessibility Guidelines set out the requirements: every field has a programmatically associated label, not just visible text nearby; instructions and format requirements are provided; errors are identified in text and suggestions for correction are given where known; focus is visible as the person moves through the fields with a keyboard; colour is not the only means of conveying information; touch targets meet the documented minimum size; and the form can be completed without a mouse, without a time limit the person cannot extend, and without content that flashes or moves unexpectedly.

In practice this means using real label elements bound to their fields, error messages that are announced to screen readers when they appear, a logical tab order that matches the visual order, a focus style that was not removed for aesthetic reasons, and inputs that are large enough and far enough apart to tap. It also means not relying on a challenge that only sighted people can solve; where spam protection is needed, a hidden field, a server-side check or a provider with an accessible alternative is the right choice.

Accessibility problems are found by testing, not by inspection: completing the form with the keyboard alone, with a screen reader, and on a phone held in one hand finds most of them. Course enquiry and application forms are a case where this decides who can apply at all, as the education industry guide explains. Forms are only one place where barriers sit; the guide to accessibility as conversion work follows the same checks through navigation, checkout, consent banners and confirmation pages, and explains why the fixes ship without a test.

Measuring completion

A form has a funnel (Funnel) of its own: the page view, the first interaction with a field, interaction with each subsequent field, the submit attempt, the error, the successful submission. Measuring those steps turns "the form converts poorly" into "people stop at the phone field" or "a large share of submissions fail validation on the email field".

Analytics tools can record a form start when the first field receives input and a submission when the form is sent, and custom events can be added per field and per error; the events and parameters guide covers how events of that kind are defined and sent. The useful metrics are: starts over page views (does the page get people to the form?), completions over starts (does the form keep them?), the field where the last interaction happened before abandonment, the share of submit attempts that produce an error and which field caused it, and all of these by device. The conversion rate calculator turns the counts into rates. Building those events in Tag Manager, from a confirmed success push to errors by field, is the subject of the form tracking tutorial.

Completion is not the only key performance indicator (KPI). A form that converts more visitors into leads the sales team cannot use has not improved. Lead quality, measured as the share of submissions that reach the next stage, has to be read alongside completion rate, and a change that raises one and lowers the other needs the business to decide which it wanted. The landing pages hub places the form's numbers alongside the page's other measures.

Illustrative example — not a real result

Finding the field that costs the most

A training company's enquiry form has nine fields and converts 12% of the people who start it. Field-level events show that 68% of starters reach the "course of interest" dropdown, 61% reach the phone field, and only 31% reach the "budget" dropdown that follows it; 14% submit. Of submit attempts, 22% fail validation, almost all on the phone field, which rejects numbers with spaces. The team makes three changes: the phone validation accepts any string with enough digits, the budget question moves to a second, optional step after the contact details are captured, and the dropdown becomes four radio buttons. Over the next month, completions over starts rise to 29%, and the first step alone (name, email, course) captures 41% of starters, so even the people who abandon the budget step are leads the team can follow up.

To test field count properly, the lead form length experiment sets out a design that measures lead volume and lead quality together, with guardrails for each.

Form changes that lift completion can lower lead quality, so the experimentation loop for lead generation explains how to test forms on qualified leads and pipeline instead.

Forms opened from a social feed need more screening than search forms, and the TikTok lead generation setup shows which qualifying questions earn their place on an instant form.

When measuring a form, the guide to using the conversion rate calculator for lead generation explains which visitors belong in the denominator and why the lead rate needs a qualified-rate check.

To see how a form's completion rate drives acquisition cost, the cost per lead walkthrough for the CPA calculator enters the click-to-lead rate and adjusts for qualification.

If enquiries are low across the whole programme, not only on one form, marketing not generating leads separates a leaking form from weak positioning, the wrong channels and contacts that are never counted.

When the field view shows many starts and few submissions, the diagnosis for high form abandonment works through field count, errors, phone inputs, privacy doubts, the size of the ask and silent submission failures.

A payment step is a form too, and the card-required trial test shows how to judge adding one at trial sign-up on paying customers per visitor.

What happens after a demo form is submitted can matter as much as the form itself, and the demo scheduling test compares instant calendar booking with a sales callback on demos held per request.

Common mistakes

  • Asking for everything sales might ever want. Each field has to justify itself against the completions it costs.
  • Placeholder text as the label. It disappears, it is hard to read, and it fails anyone who looks back.
  • Two columns to save space. The reading order breaks and fields get missed.
  • Validation stricter than reality. Names, phones and postcodes come in more forms than the rule allows.
  • Generic error messages far from the field. Say what is wrong and how to fix it, next to the problem.
  • Clearing the form on error. The person pays the whole cost again.
  • Dropdowns for short lists. Radio buttons show the options at once; dropdowns hide them.
  • Breaking autofill. Non-standard field names force typing the browser would have done.
  • Measuring submissions only. Without field-level events the form's problems stay invisible.
  • Judging the form by completions alone. More leads the sales team cannot use is not an improvement.

Frequently asked questions

How many fields should a lead form have?

As many as the next step genuinely needs and no more. Work backwards from what happens after submission: a sales call needs a name, a way to reach the person and perhaps one qualifying question; a quote needs whatever the quote depends on; a download may need only an email address. Every field beyond that should have to justify itself against the completions it will cost, and most of the extra questions belong in the first conversation rather than on the form.

Are multi-step forms better than single-page forms?

Often, for longer forms and on mobile. Splitting a form into steps lets the first step ask for little, captures contact details before the harder questions, shows progress and keeps each screen short enough for a thumb. For a form of three or four fields the split adds effort without benefit. The honest answer for any given page comes from testing the two against each other with completion and lead quality both measured.

Should the phone number be required?

Only when the next step is a call and the visitor expects one. A required phone field on a form that promises an emailed guide costs completions and produces made-up numbers; an optional phone field on a quote form with a line explaining when and why the business will call is usually filled in by the people who want the call. If sales genuinely cannot proceed without it, say so and say what the call will be.

Terms used on this page

Tools to try

Where this page is used

Part of these learning paths

Sources

  1. Website Forms Usability: Top 10 Recommendations — Nielsen Norman Group
  2. Web Content Accessibility Guidelines (WCAG) — W3C

Related