Form conversion rate: how to measure it, and how to increase it

A form sits at the end of a lot of work. Someone wrote the ad, the email or the page, and a visitor clicked through. Form conversion rate is how much of that work turns into something you can use: a lead, a registration, an application, a response.

It is also one of the most misquoted numbers in marketing. Two people can look at the same form, report 18% and 40%, and both be right, because they divided by different things. One ground rule for this guide: published benchmarks rarely say how they were counted or what kind of form they measured, so none are quoted here. Every figure on this page is arithmetic or a labeled illustration.

What form conversion rate means, and the two ways it gets calculated

Form conversion rate is the percentage of people who viewed a form and went on to submit it:

That is the definition most marketers mean, and the one this guide uses. Many form tools, Upform included, also show submissions divided by starts, where a start is someone who got past the first question. People call that conversion too, but it is better named the completion rate, and it answers a narrower question.

The three rates fit together: conversion rate is start rate multiplied by completion rate. An illustrative example: a demo request form gets 2,400 views in a month, 1,080 visitors answer the first question, and 432 submit. The start rate is 45%, the completion rate is 40%, and the form conversion rate is 18%, which is 40% of 45%.

That split is the most useful idea on this page. A low conversion rate is either a starting problem or a finishing problem, rarely both to the same degree, and each has different fixes. Starting problems live outside the questions: the promise, the page, the placement, the first screen. Finishing problems live inside them: length, effort, sensitive asks. The form abandonment guide goes deep on the finishing side.

How to measure form conversion rate in Upform

Every Upform form has a Results tab, part of the built-in form analytics. It records three events per form: a view when the published form loads, a start when someone moves past the first question, and a submit when they send it. Every card on the tab is built from those three events.

Here is how each card is counted, which matters once you start comparing periods:

• Form Views counts loads of the published form, at most once every 30 minutes in the same browser tab, so a visitor who refreshes does not inflate it. Previews inside the builder are not counted.
• Unique Views counts distinct visitors, recognized by browser. Use it as the denominator if you would rather count people than visits.
• Form Starts counts people who moved past the first screen. If your form opens with a Statement screen, clicking past it counts as the start.
• Form Submissions counts sent forms, and Unique Respondents the distinct visitors who sent them.
• Incomplete Submissions counts sessions that started and never submitted.
• Form Completion Rate is submissions divided by starts: the finishing half of the equation.
• Average Completion Time runs from start to submit, averaged over submitted responses.

Two filters sit above the cards. The device filter (desktop, tablet or mobile) recalculates the cards for one kind of device, which is the fastest way to spot a mobile problem. The time filter (today, yesterday, the last 7, 30 or 90 days, or all time) is what you use to compare one period with another.

One thing the Results tab cannot tell you is which question an unfinished visitor stopped on. It knows how many people started and did not submit, not where they left. The Question Completion Funnel further down the tab is built from submitted responses, so a drop there means a question was optional or skipped by logic, not that people quit on it. Read the start and completion rates first. They tell you which half of the form to work on.

What is a good form conversion rate?

There is no honest universal number. A one-field newsletter signup on a page that already sold the idea and a twelve-question job application reached from a job board should not share a target. The rate also depends on things the form does not control: how warm the traffic is, how well the link described what was behind it, and whether the visitor arrived on purpose.

Three comparisons are worth more than any benchmark:

• The same form, period over period. The last 30 days against the 30 before, with the same traffic sources.
• The same form, by device. If mobile converts at half the desktop rate, that gap is yours to close, whatever the industry average says.
• The same form, by source. A form linked from your newsletter will convert better than one reached from a cold ad. Upform does not split results by traffic source, so if sources matter, give each one its own copy of the form (see testing below).

When the rate drops, check the traffic mix before you touch the form. A new campaign that sends curious but uncommitted visitors can lower the rate while bringing in more submissions in total, and nothing about the form has changed.

How to increase form conversion rate: seven levers

Each lever below is labeled with the rate it mostly moves, start or completion, so you can match it to what your own numbers say.

1. Fewer, easier questions (completion)

Every question is a small decision to keep going. Length rarely shows up as one dramatic drop. It shows up as a lower completion rate and a longer average completion time. Run the delete test on every question: if its answers vanished from every response, would anyone do anything differently? If not, remove it. Questions you only need later, such as company size for a sales call, can often be asked on the call.

Then make the survivors cheaper to answer. "Tell us about your team" asks for a paragraph. "How many people are on your team?" with four ranges asks for one tap. Use the question type that does the work: Single Choice or Dropdown where the answers are predictable, Date (separate month, day and year boxes that check the entry) instead of a free text box, Phone Number with its country flag and dial code instead of a plain text box. Upform has 21 question types, and picking the right one is often a bigger win than rewording.

Showing one question per screen helps here too. A conversational form never shows the whole list at once, so nobody counts the fields and decides against it before the first answer. The multi-step forms guide covers when splitting a form into steps pays off and when it only adds clicks.

2. A clear promise (start)

People start a form when they can see what they get and roughly what it will cost them. Say both before the first question: what happens after they submit, and how long it takes. "Five questions, about two minutes, and we reply within a day" does more for a start rate than any button color.

Then keep the promise visible. Upform's design settings let you show progress as a bar, dots or a percentage, and "about two minutes" stays believable when the indicator moves one step per question. If the form is longer than the promise suggests, change the form or change the promise. A mismatch shows up as a healthy start rate and a completion rate that sags.

3. Placement (start)

The same form converts differently depending on where it lives. A form one click away from the moment of intent, such as the pricing page, the end of a case study or a confirmation email, catches people when they want to act. A form linked from a footer catches whoever is left. The Share panel gives you three ways to put it in front of people:

<iframe src="https://app.upform.com/preview/demo-request" width="100%" height="600" frameborder="0" style="border:none;border-radius:8px;"></iframe>Copy

Embedding usually beats linking out, because the visitor never leaves the page that persuaded them. Put the embed where the page's argument ends, not above it, and make sure the words right before the form match what the form asks. The how to embed a form guide covers the details, and share and embed lists every option.

4. Mobile (both)

A form that works at a desk can be a chore on a phone. Upform's published forms use full-width answer targets and a sticky back and next bar on mobile, but your choices still matter. A plain text box for a phone number opens the wrong keyboard, a dropdown of two hundred options is a long scroll, and a long text question asks someone to type a paragraph with their thumbs.

Check it with numbers rather than a hunch. Set the Results device filter to Mobile and note the start and completion rates, then switch to Desktop. A large gap in completion means the questions are probably fine and the inputs are not. Then fill out the form yourself on a real phone, one-handed, over cellular data.

5. Trust signals (start and completion)

A visitor deciding whether to hand over an email address is really asking whether you will use it well. Small signals answer that. Your logo, colors and font on the form, so it clearly belongs to the site they came from. A line of help text under the email question saying what you will send and how often. A short privacy note on the first or last screen. Upform's design settings cover the logo, colors, font and background on every plan, and on Pro you can remove the Upform badge. Spam protection costs real people nothing: public forms use invisible reCAPTCHA v3, so there is no puzzle to solve.

6. Sensitive questions later (completion)

Phone number, budget, address, date of birth. People rarely quit on these because they are hard. They quit because the form has not yet earned the answer. The same question at question five, after the easy ones and with a line saying why you need it, costs far less than at question two. Better still, ask whether you need it at all: a phone number collected "in case we need to call" is a number nobody dials.

7. Conditional logic (completion)

The most reliable way to shorten a form without losing information is to stop asking people questions that do not apply to them. With conditional logic, each rule joins conditions with AND or OR, using 20 operators, and takes one of four actions: show a question, hide it, jump to another question, or end the form. A visitor who says they are a student skips the company questions. Someone who picks "Just browsing" gets a shorter path. Every question a person never sees is one they cannot quit on.

Logic also settles a worry sales teams have about shorter forms: fewer questions can mean less qualified leads. Instead of asking everyone for budget and timeline, ask one qualifying question and show the detailed follow-ups only to the people they apply to.

Which lever to reach for first

How to test changes one at a time

Change three things at once and the rate goes up, and you have learned nothing you can reuse. Change one thing and you know what worked. There are two practical ways to do it in Upform.

Before and after on the same form

The simplest test: make one change, then use the time filter to compare equal periods, such as the 30 days before the edit against the 30 days after. It works when traffic is steady and comes from the same places in both periods. It fails when something else changes at the same time, such as a new campaign or a holiday, so write down the date of every edit.

Two versions side by side

For a cleaner comparison, run both versions at once. In the Forms list, the menu on any form has Duplicate. The copy arrives as an unpublished draft with "(Copy)" in its name, and once published it has its own link and its own Results tab, starting from zero. Make your one change on the copy, check its Connect panel so responses reach the same places as the original, publish it, and send part of your traffic to each version.

Upform does not split traffic for you, so the split is yours to arrange: half your email list gets one link and half gets the other, two ad variations point at the two forms, or the embed on your page alternates week by week. Whatever the method, keep the two audiences as alike as possible. When the test ends, make the winning change on the original form, so the link and embeds you already use keep working, and unpublish the copy.

• Change one thing. One question, one line of copy, one placement. Not a redesign.
• Keep everything else equal. Same traffic source, same days of the week, same length of time.
• Read the right rate. A change to the promise or placement shows up in the start rate. A change to the questions shows up in the completion rate. Conversion rate shows both.
• Wait for enough data. With a few dozen views per version, a handful of submissions swings the rate by several points. Keep going until one or two extra submissions would not change your conclusion.
• Stay out of your own data. Builder previews are not counted, but visits to the live link are.

A worked example: one change, two versions

Here is the loop on the illustrative demo request form from earlier. The numbers said finishing, not starting, was the problem: a 45% start rate was reasonable for a form reached from the pricing page, and a 40% completion rate was not. The form had nine questions, with the phone number second and an open "Tell us about your company" in the middle.

The team duplicated the form and made one change to the copy: the phone number moved from question two to question five, with help text saying it was only used to confirm the demo time. The emails promoting demos linked to each version on alternate sends for four weeks.

Illustrative numbers from a sample demo request form over four weeks. They show the method, not a result to expect.

The start rate did not move, which is what you would expect: the first screen was identical in both versions. The whole gain came from inside the form, and because only one thing changed, the team knew which thing it was. The next test, replacing the open company question with a Single Choice of team sizes, started from the winning version.

Not every test ends with a winner. If the two versions finish within a point or two of each other, the honest reading is that the change did not matter much, and the next idea deserves the traffic.

Realistic expectations

Form changes have a ceiling. They can stop you losing people who already wanted what the form offers. They cannot create demand for an offer nobody wants, and they cannot turn visitors who arrived by accident into visitors who arrived on purpose. If your start rate is very low, the fix is more often in the ad, the email or the page than in the form.

Expect the biggest improvements from the first few changes, the obvious ones: the question nobody reads, the phone number at question two, the open text box that should have been a choice. After that, gains get smaller and tests take longer to call. That is normal, and usually the moment to turn back to traffic.

And watch what happens after the submit. A shorter form that converts at 30% and fills your calendar with calls that go nowhere is not better than a longer one at 20% that books the right people. Conversion rate is a means, not the result.

A checklist to increase form conversion rate

Run this against any live form. Each line maps to a section above.

• You know the start rate and the completion rate, not just the conversion rate they multiply into.
• Every question passes the delete test, and predictable answers use choices rather than open text.
• Before the first question, the visitor knows what happens after submitting and roughly how long it takes.
• The form sits where intent is highest, and the words just before it match what it asks.
• Phone, budget and address come after the easy questions, each with a line saying why you need it.
• Questions that apply to some people sit behind logic, so everyone else never sees them.
• Mobile completion rate is close to desktop, and someone has filled out the form on a real phone.
• Each change is tested alone, on equal periods or on a duplicate, and its date is written down.

The form design best practices guide covers the layout habits behind several of these, and the lead capture forms guide applies them to the forms that feed a sales pipeline. Start with the split, though: find out whether your form has a starting problem or a finishing problem, and fix that one first.