Why Forms Are the Most Tested Part of Any Site We Build
Forms cause the majority of checkout abandonment and accessibility failures, so we test them harder than anything else on a site.
A form is the one part of a website where a user has to stop reading and start working. That's exactly why it breaks more often, loses more revenue, and gets tested more heavily by us than any other component we build.
This isn't a stylistic preference. It's backed by a decade-plus of usability research showing that form design decisions — how many fields, what order, how errors are shown — directly move conversion rates and, separately, decide whether a site meets legal accessibility requirements. Below is the evidence, and what it means for how we build and test forms.
The numbers behind the problem
Baymard Institute has run large-scale checkout usability research for over 14 years. Its work is based on 25 rounds of qualitative usability testing with 4,400+ test participant/site sessions following the "Think Aloud" protocol, where users narrate their reactions as they try to complete a purchase. The consistent finding: checkout design is often the sole cause for users abandoning their purchase, either because they get frustrated enough to leave in anger, or because they simply don't understand how to fill in a field.
The scale of this is not small. Baymard's aggregated review of 50 separate studies puts the average cart abandonment rate at 70.22%. A secondary summary of Baymard's checkout research adds that optimizing checkout flow and design can increase conversions by 35.26% — one of the highest-impact fixes available on an ecommerce site, driven by reducing fields and removing friction points.
Device matters here too. The same source notes mobile abandonment rates reach 80.2%, about 10 percentage points higher than desktop. That gap matters a lot for Indian businesses, where most commercial traffic arrives on a phone — though this particular figure comes from a general dataset, not an India-specific study.
What actually breaks: field count and layout
Nielsen Norman Group states the business case bluntly: usability problems on forms really hurt business, and are usually the reason a conversion funnel shows a big drop-off right at the forms page.
Baymard's applied research gives a concrete threshold for when this starts happening. Test participants consistently become intimidated when 10–15+ form fields are viewable on a single checkout page. The fix isn't always removing fields — sometimes it's consolidating them. Baymard specifically recommends avoiding splitting single input identities (e.g., replacing "First Name" and "Last Name" with "Full Name") and hiding infrequently used optional fields behind a link.
Google's own web.dev guidance backs this at the layout level, not just the field-count level: research shows it's best to use only a single column, since users don't want to spend time searching for the next form control, and best practice is to position the label above the form control so users can scan a form quickly. Google states plainly that good form design can significantly reduce form abandonment rates.
This is why, before we ever ship a form, we're checking:
| Check | What we're testing for |
|---|---|
| Field count on screen at once | Baymard's 10–15 field intimidation threshold |
| Column layout | Single column vs. multi-column scanning friction |
| Label position | Above field vs. inline vs. floating |
| Field consolidation | Whether split fields (first/last name) can be merged |
| Optional field visibility | Whether rarely-used fields are hidden behind a toggle |
| Mobile vs. desktop completion time | Given the higher mobile abandonment rate |
Errors are a compliance issue, not just a UX one
Testing a form isn't only about pushing up conversion. It's also a legal accessibility requirement in most markets that follow WCAG standards. The W3C's WCAG standard, Success Criterion 3.3.1 (Level A), requires that if an input error is automatically detected, the item that is in error is identified and the error is described to the user in text, with the intent of ensuring users are aware that an error has occurred and can determine what is wrong.
WebAIM turns that requirement into a testable checklist. An accessible form needs a 3-step process for usable and accessible error recovery: alert the user to the presence of the error in an apparent and accessible manner, allow the user to easily access the form controls that need to be modified, and allow resubmission and revalidation of the form.
None of that can be verified by looking at a design mockup. It requires actually testing the form with a keyboard only, and again with a screen reader, to confirm that:
- An error announcement actually reaches assistive technology, not just a red border that a sighted user notices.
- Focus moves to the erroring field, or at least a link to it, so a keyboard-only user isn't left hunting.
- The form can be corrected and resubmitted without losing already-valid data.
This is the layer of form testing that has nothing to do with conversion rate and everything to do with whether the form is usable at all for a meaningful share of visitors — and whether it holds up against an accessibility complaint.
Where the research disagrees, and why we don't chase a single number
It's worth being honest that the secondary literature on "how many fields is ideal" is inconsistent, and we don't treat any single recycled statistic as gospel. One widely-cited figure puts the average checkout containing 23.48 form elements, with an ideal flow functioning with just 12–14 elements. A separate summary of Baymard's more recent tracking gives a lower and different figure — an average of 11.3 fields in 2024, down from 11.8 in 2021 and 12.7 in 2019.
These numbers aren't measuring the same thing — elements versus fields, different years, different site samples — so they shouldn't be quoted interchangeably. The practical takeaway for us is not "hit exactly 11.3 fields." It's that the underlying Baymard research, not a blog restating it, is the reference we test against, and that the direction across all these sources is the same: fewer, better-organized fields consistently beat more fields, regardless of which exact number a given study lands on.
What this means for a site built for the Indian market
The standards behind all of this — WCAG, Google's web.dev guidance, Baymard's checkout research — are not regional. HTML validation and accessibility conformance apply the same way to a site built in Pune as one built in Berlin. Nothing about how a browser parses a form or how a screen reader announces an error changes because the business is Indian.
What we can't do here is point to India-specific data — a Digital Personal Data Protection Act requirement on consent-form design, a Government of India accessibility guideline (GIGW), or a TRAI statistic on Indian mobile form-abandonment rates — because none of that surfaced in this research pass. If your business operates under specific Indian regulatory consent requirements, that needs a dedicated compliance review rather than an assumption drawn from global UX research.
How this changes what we test before launch
Given all of the above, our pre-launch checklist for any form treats it as three separate problems, not one:
- Conversion risk: field count, column layout, label placement, and whether the form can be shortened without losing required data — tested against Baymard's and Google's guidance.
- Abandonment risk: specifically re-testing on mobile, given the higher mobile abandonment figure, not just assuming desktop testing covers it.
- Accessibility risk: keyboard-only and screen-reader passes against the WCAG error-identification criterion and WebAIM's three-step error recovery process.
A form that passes all three isn't guaranteed to convert at some specific rate — nobody can promise that. But it removes the most common, well-documented reasons a form fails: too many visible fields, unclear error messages, and a layout that makes users search for the next input instead of just filling it in.
Questions people also ask
How many form fields is too many?
There's no single universal number, since sources measure this differently. Baymard's applied research found that users become intimidated once 10–15+ fields are visible on one screen, which is the practical threshold we design against rather than any single "ideal field count" statistic.
Does form design really affect conversion rate, or is that overstated?
It's well documented, not overstated. Baymard's checkout research across 4,400+ test sessions found checkout design is often the sole cause of abandonment, and a related summary states that optimizing checkout can lift conversions by 35.26%.
Is accessible error handling on forms actually required, or just best practice?
For sites that need to conform to WCAG, it's a formal requirement under Success Criterion 3.3.1, which mandates that detected errors be identified and described in text. WebAIM's testing checklist is the practical way to verify a form actually meets it, rather than assuming a red border is enough.