The 9-point form conversion checklist

Form optimisation has a reputation for being fiddly, and it does not deserve it. Most live forms lose most of their completions to a handful of predictable problems, and you can find them without a testing tool, a designer or a statistically significant sample.

This is the audit we run first. Work through it in order — the early checks usually recover more than the late ones — and change one thing at a time so you know what worked.

Key takeaways

• Delete before you decorate: unused fields cost more than any visual change. - Sequence questions for momentum, with effort in the middle of the flow. - Test on a real phone, one-handed, on a slow connection. - Errors should be recoverable in place, never destructive. - Judge everything on completed submissions per visit.

1. Count the fields, then justify each one

List every field and name the decision it changes. Route, price, prioritise, reply, comply. Anything that does not change a decision is research, and research belongs after the submission, not before it.

Move each field into one of three buckets: essential now, ask later, delete. On a typical business form a third of the fields leave the essential bucket immediately, and nothing operational breaks.

2. Check the sequence, not just the length

Order decides how many people reach the end.

• First question: easy, relevant, answerable without looking anything up. - Middle: the effort — long text, uploads, budget, dates. - Last: nothing new or surprising. No unannounced payment step, no consent ambush.

Attach a reason to every sensitive field. "Budget range — so we send realistic options" outperforms a bare label because the trade is visible.

3. Open it on a real phone

Not a resized browser window. A real device, held in one hand, outdoors if you can manage it.

• Are tap targets comfortable, with space between them? - Does the email field bring up the email keyboard, and the phone field the numeric one? - Does the keyboard cover the field you are typing into? - Can you finish without pinch-zooming? - Does the submit button stay reachable with a thumb?

Mobile is the majority case for almost every public form. If it works there, the desktop version is rarely the problem.

4. Make errors survivable

Every error should tell someone what to do, next to where they did it, without destroying their work.

• Validate format on blur, completeness on submit. - Write instructions, not diagnoses: "Enter a date after today", not "Invalid". - Never clear a field because it failed validation. - Never scroll people to the top of the page to find a summary box.

5. Time the load

Forms are often the slowest page on a site because they inherit every script the marketing stack ever added. Load the form itself first, defer analytics and chat widgets, and keep the first interaction available in under two seconds on a mid range phone. Slow forms lose people before the first field.

6. Say what happens next

Before the submit button, answer three questions in one line each: who sees this, when will I hear back, and what will you do with my data. This is the cheapest trust improvement available, and it reduces "did it send?" support emails to nearly zero.

7. Design the confirmation, not just the submission

A confirmation screen should repeat what was submitted, give a reference where one exists, and state the next step and its timing. Then send the same thing by email immediately. A redirect to the homepage is a lost opportunity and a support ticket.

8. Check accessibility properly

• Every field has a real, visible label — placeholders are not labels. - The whole form is completable by keyboard, in a logical order. - Focus states are visible against your background. - Errors are announced to screen readers, not only shown in colour. - Contrast passes on labels, help text and disabled states.

Accessibility work reliably raises completion rate for everyone, because the same fixes remove ambiguity for people in a hurry.

9. Instrument it, then decide

Three numbers, tracked before and after every change:

1. Completed submissions per visit — the outcome that pays. 2. Field-level drop-off — where the loss actually happens. 3. Time to first response — because a fast reply converts submissions into customers, and no form design compensates for a three-day silence.

Run each change for two full weeks so weekday and weekend traffic both land in the sample.

Run it in this order

If you only have an hour: delete unused fields, fix the worst-performing field, test on a phone, and write a proper confirmation. Those four cover most of the available gain on a typical form. The rest of the list is for the second pass.

Frequently asked questions

How often should I audit a live form?

Quarterly for anything commercially important, and immediately after any change to the surrounding page, tracking stack or fulfilment process.

Which single change usually helps most?

Removing the field with the highest drop-off. It is nearly always cheaper and more effective than a redesign.

Do I need A/B testing to improve a form?

No. Most forms do not have the traffic for reliable tests. Sequential before-and-after measurement with one change at a time and a two-week window is good enough to make sound decisions.

Should I add a chat widget to my form page?

Only if someone answers it. An unattended widget adds load time and covers your submit button on mobile, which is a net loss.

What is a good completion rate?

It depends entirely on traffic source and intent, so compare against your own history rather than a benchmark. A steady upward trend in completed submissions per visit is the signal that matters.