What this guide covers
- Measuring conversion properly
- What a realistic completion rate looks like
- Diagnosing your form in 20 minutes
- Tactics 1–5: structure
- Tactics 6–11: the fields themselves
- Tactics 12–15: copy and trust
- Tactics 16–19: technical and mobile
- All 19 tactics ranked by impact
- How to test without fooling yourself
- Frequently asked questions
Measuring conversion properly
Most teams track one number — submissions — and that number cannot tell you where the problem is. Form conversion is three separate rates, and each one fails for different reasons.
| Metric | Definition | When it's low, the problem is |
|---|---|---|
| View → start | People who interact with the first field, divided by people who saw the form | The form looks long, the headline doesn't match the promise that brought them, or it's below the fold |
| Start → complete | Submissions divided by starts | A specific field. This is the rate worth optimising |
| Field drop-off | Per-field abandonment | Points at exactly which field to fix |
| Time to complete | Median seconds from start to submit | Unexpected complexity — a lookup, a slow validation, an unclear question |
| Error rate | Validation failures per submission attempt | Over-strict validation, usually on phone numbers and postcodes |
What a realistic completion rate looks like
These are the ranges we see across the forms we review. Treat them as sanity checks, not targets — the variation within any category is far wider than the variation between them.
| Form type | Typical start → complete | Main drag on the rate |
|---|---|---|
| Newsletter signup (email only) | 60–80% | Unclear value of subscribing |
| Contact / enquiry form | 40–60% | Phone number, message field length |
| Event registration | 35–55% | Dietary and accessibility fields asked too early |
| B2B demo request | 25–45% | Company size, budget, job title |
| Quote request | 20–40% | Fields requiring lookup or measurement |
| Job application | 10–30% | CV upload, duplicate manual entry of CV contents |
| Insurance / financial application | 5–20% | Length, document requirements, trust |
Diagnosing your form in 20 minutes
Open field-level analytics and find the biggest drop
One field almost always accounts for a disproportionate share of abandonment. Start there rather than rewriting the whole form.
Fill in your own form on a phone, on mobile data
Not on your desktop, not on office wifi. Every keyboard-type mistake, every slow-loading element and every cramped tap target becomes obvious within thirty seconds.
Count the fields, then justify each one out loud
For each field, name the decision that changes based on the answer. Fields that fail this test are the cheapest conversion gain available to you.
Check what happens when you get it wrong
Submit an invalid email, a phone number with spaces, a postcode in lower case. Over-strict validation rejecting valid input is one of the most common and least visible causes of abandonment.
Watch someone else do it
Anyone who has not seen the form. Say nothing. Where they hesitate is your list, in priority order.
Tactics 1–5: structure
1. Remove every field you do not use
The highest-leverage change available, and the one most often blocked by internal politics rather than evidence. Every field is a request for effort and a small increase in perceived risk. Audit ruthlessly: if no decision, routing rule or follow-up message changes based on a field's value, it does not belong on the form. "It's useful to have" is not a decision.
Expected impact: large. Removing two or three genuinely unused fields from a mid-length form is frequently the biggest single gain in an optimisation programme.
2. Split long forms into steps — with a progress indicator
Above roughly five fields, a multi-step form usually outperforms the same fields on one page. The mechanism is that the first screen looks achievable, and once someone has invested effort they tend to continue. The progress indicator is not optional: a multi-step form without one replaces a visible long form with an unknowable long form, which is worse.
Expected impact: moderate to large on long forms, negative on short ones.
3. Put the easy questions first
Name and email before budget and company size. Early questions should cost nothing to answer — their job is to start the commitment, not to qualify. Qualification questions belong on the last screen, where the respondent has already invested and is far more likely to push through.
Expected impact: moderate, and free to implement.
4. Make the form's length honest
"Takes about 2 minutes" or "6 questions" above the first field reduces the uncertainty that causes people not to start. This helps the view-to-start rate specifically. Do not lie — an underestimate that becomes obvious halfway through costs you more than no estimate at all.
Expected impact: small to moderate on view-to-start.
5. Use conditional logic to hide irrelevant questions
A 20-question form where any individual respondent sees eight feels like an eight-question form. Branching is the only way to be thorough without being long. Every major alternative supports it — our comparison of 14 form builders covers which ones include it on free plans.
Expected impact: large on forms serving multiple audiences.
Tactics 6–11: the fields themselves
6. Make the phone number optional
Phone number depresses completion more reliably than any other common field, because it signals a sales call. Making it optional typically recovers a meaningful share of abandoners, and the people who do provide it are self-selected as open to being called — which usually improves lead quality rather than harming it.
Expected impact: large where phone is currently required.
7. Replace open text with choices wherever possible
Typing is expensive; tapping is cheap. "What is your budget?" as a free-text field asks someone to decide on a number and commit to it in writing. The same question as four ranges is a single tap, and the data is more consistent to analyse. Reserve free text for the one question where the answer's specificity is genuinely the point.
Expected impact: moderate.
8. Use the right input type on mobile
An email field should raise the email keyboard with the @ visible; a phone field should raise the number pad. When a text field raises a full QWERTY keyboard for a phone number, you have added friction to every mobile submission. This is one line of configuration and it is wrong on a surprising number of live forms.
Expected impact: small individually, meaningful across a mobile-heavy audience.
9. Loosen your validation
Validation exists to prevent unusable data, not to enforce formatting preferences. Accept phone numbers with spaces, brackets and country codes. Accept postcodes in any case. Accept names with apostrophes, hyphens and non-Latin characters — rejecting them is both a conversion problem and an exclusion problem. Normalise on your side, silently.
Expected impact: moderate, and often invisible until you look at error rates.
10. Validate inline, not on submit
Showing an error next to the field as soon as someone leaves it is far less punishing than rejecting the whole form after they press submit and scrolling them back to the top. Error messages should say what to do — "Please include an @ in your email address" — not what went wrong.
Expected impact: moderate on longer forms.
11. Pre-fill everything you already know
If the respondent arrived from an email you sent, you know their name and address. Pre-filling those fields converts a five-field form into a three-field form. Most builders support URL parameters or hidden fields for this; it takes twenty minutes to set up and applies to every campaign afterwards.
Expected impact: large for email-driven traffic.
Tactics 12–15: copy and trust
12. Make the button say what happens next
"Submit" describes your database operation. "Get my quote", "Book the demo", "Send my application" describes the respondent's outcome. This is the cheapest change in this guide and it reliably produces a small lift.
Expected impact: small but consistent.
13. Explain why you need the uncomfortable fields
A one-line note under a phone number field — "so we can call you back within one working day" — converts an unexplained demand into a reasonable request. Do this for any field a respondent might resent: phone, company size, budget, date of birth.
Expected impact: moderate on the specific field.
14. Say what happens after they submit
"We'll reply within one working day" reduces the uncertainty that causes hesitation at the final step. It also reduces duplicate submissions from people who were not sure it worked.
Expected impact: small to moderate.
15. Put trust signals next to the form, not in the footer
A privacy note, a security badge, or a one-line customer proof point placed adjacent to the submit button does work that the same content in the footer does not, because it is visible at the moment of hesitation. Keep it to one line — a wall of badges reads as compensation for something.
Expected impact: small to moderate, larger on forms asking for sensitive data. If you handle health or EU personal data, the underlying compliance matters more than the badge — see our HIPAA and GDPR form builder guide.
Tactics 16–19: technical and mobile
16. Fix load time before anything else
An embedded form that takes four seconds to appear has already lost a share of visitors before the first field renders — and none of them appear in your start-to-complete rate, which is why slow forms often look deceptively healthy. Measure the time from page load to the form being interactive on a mid-range phone on mobile data, not on your laptop.
Expected impact: large where load time is currently poor.
17. Enable save and resume on long forms
Any form that takes more than about five minutes, or requires a document the respondent may not have to hand, needs a way to come back. Without it, every interruption is a lost submission. Cognito Forms and Zoho Forms both include this; see our Cognito Forms comparison for how it works in practice.
Expected impact: large on applications and long registrations.
18. Capture partial submissions
Some builders store answers as they are entered rather than only on submit, so an abandoned form still leaves you a name and email to follow up. This does not improve your completion rate — it recovers value from the people the rate is failing to count. Treat those contacts with care: someone who abandoned a form has not opted into a sales sequence.
Expected impact: recovers a share of abandonment; check your privacy obligations first.
19. Give every field enough room to tap
Cramped tap targets, fields that sit under a mobile keyboard, and dropdowns that require precision scrolling all cost submissions on phones. Since most consumer form traffic is now mobile, an optimisation that only works on desktop is optimising the minority case.
Expected impact: moderate on mobile-heavy audiences.
All 19 tactics, ranked by impact
Ranked by the size of the lift we typically see, weighted by how often the problem is present in the first place. Work top down.
| # | Tactic | Impact | Effort | Fixes |
|---|---|---|---|---|
| 1 | Remove unused fields | High | Low | Start → complete |
| 2 | Make phone optional | High | Low | Field drop-off |
| 3 | Fix form load time | High | Medium | View → start |
| 4 | Pre-fill known data | High | Medium | Start → complete |
| 5 | Conditional logic to shorten paths | High | Medium | Start → complete |
| 6 | Save and resume on long forms | High | Low | Abandonment |
| 7 | Multi-step with progress indicator | Medium | Medium | View → start |
| 8 | Easy questions first | Medium | Low | Early drop-off |
| 9 | Choices instead of open text | Medium | Low | Field drop-off |
| 10 | Loosen validation rules | Medium | Low | Error rate |
| 11 | Inline validation | Medium | Low | Submit-step drop-off |
| 12 | Explain uncomfortable fields | Medium | Low | Field drop-off |
| 13 | Mobile tap targets and layout | Medium | Medium | Mobile completion |
| 14 | Capture partial submissions | Medium | Low | Recovers abandonment |
| 15 | Trust signal beside the button | Medium | Low | Final-step hesitation |
| 16 | Correct mobile input types | Small | Low | Mobile completion |
| 17 | Outcome-based button text | Small | Low | Final-step conversion |
| 18 | State the form's length | Small | Low | View → start |
| 19 | Say what happens next | Small | Low | Final-step hesitation |
How to test without fooling yourself
Know whether you have the traffic
Detecting a modest improvement reliably takes a few hundred submissions per variant. If your form receives fifty a month, an A/B test will take longer than the form will exist in its current form. Below that threshold, apply known-good practice and measure the before-and-after over a longer window rather than running a split test you cannot conclude.
Change one thing at a time
Bundling five improvements tells you the bundle worked. It does not tell you that three of them helped, one did nothing, and one hurt — which is the usual outcome, and the one worth knowing.
Run whole weeks
Form traffic is strongly day-of-week dependent. Always compare complete weeks against complete weeks, and never end a test on a Friday because the numbers looked good.
Watch what happens downstream
A change that increases submissions but halves the proportion that turn into customers is not an improvement. Track the metric at the end of your funnel, not the one at the top of it. This is the specific failure mode of removing qualification fields: more leads, worse leads, more sales time wasted.
Build forms that get finished
Conditional logic, field-level analytics and mobile-first layouts — on forms.app's free plan.
Try forms.app Free → Compare all 14 alternatives